That is, unless they complained about it.

But do you really want to have to rely on customer dissatisfaction and complaints before you can improve service? Wouldn’t you rather be proactive?

Today, we released a new tool to help you do just that. DéjàClick Test On Demand is our new free service that provides immediate insight into the responsiveness and availability of your websites and applications from dozens of locations across the globe.

It’s easy. Simply use DéjàClick to record any multi-step Web transaction, click the Test On Demand icon, and select your desired location from the drop-down menu. (There are over 60 locations for AlertSite customers to choose from. Or, as a member of the DéjàClick community, you can select either Chicago or London.)

Then, click play to watch your recording play back like a movie. As you do, observe the performance metrics that are revealed at each step of the transaction.

You’ll see:

• Action details for each page of the transaction, including HTTP status; response times for DNS lookup, connect, redirect, time to first byte, HTML content and full page content; and waterfall charts.
• Event details that show the performance of every mouse-over, form fill-in, navigation and click, so users can make sure all the interactive and Flash components of their websites are performing as they should.
• Screen captures of errors that appear instantaneously to show users what the page looked like when something went wrong and at what step of the transaction.

By the end of the play back, you’ll have a deep and comprehensive understanding of how that transaction performs for users in your selected location.

Don’t take our word for it, though. Try it yourself. DéjàClick is available as a free download to Firefox users. Run some tests using the Test On Demand tool, and let us know what you think!

For over a year now, ReadingGlasses.com has been using DéjàClick to record and test customer click streams so that customers could browse products, add items to their shopping carts, and check out without a hitch. For Jason Kyle, the company’s e-commerce director, one of the solution’s most valuable features is the instant error notifications. Every time an error occurs on his site, or response times fall below predetermined thresholds, Kyle receives an instant phone alert, enabling him to quickly address the issue. As a result, ReadingGlasses.com has been able to avoid nearly all outages or major slow-downs.

In the full-length case study, which is now available, Kyle describes one particular instance in which real-time notifications alerted him to a situation on his website that could have had devastating results.  Thanks to the alert, Kyle discovered the problem and took action to address it before it could further cripple his site or lead to customer complaints and potential lost revenue.

For the full story, click on the case study below. Also, visit our InfoCenter to find mores stories of how leading Web companies improved their Web performance with AlertSite.

Thanks to Sierra Trading Post, Shoebuy.com, Crutchfield, and all the other retailers who listed us as their Web performance monitoring provider of choice!

Here’s one we came across today, implemented by the “deal of the day” company, Groupon. Groupon launched its first nationwide deal today, $25 for $50 worth of merchandise at Gap, and was quickly overwhelmed by the response. TechCrunch reported that “as of 11 a.m. PST (the e-mail blast went out at roughly 6 a.m. in each time zone), Groupon has sold 200,000 Groupons and is currently selling roughly 10 per second.”

The result has been inconsistent availability and shaky performance. But kudos to Groupon for keeping their customers informed. And for offering to contact them when the site was back up and running.

Here’s the message we got when we tried to visit the site this morning.

As these type of sites gain popularity, providers will need to take extra steps to ensure their websites can hold up to the big deals of the day.

But don’t take our word for it. Join us in two weeks, on Aug. 17, for a live discussion with Bojan Simic, Founder and Principal Analyst at TRAC Research, a B2B market research and analyst company that specializes in IT performance management. Bojan will share some of the most recent research on Web performance metrics and highlight new trends in website performance management.

Plus, we’ll also have Steve Lubahn, General Manager of Supreme School Supply on the line. We first introduced you to Steve earlier this summer following our trip to IRCE in Chicago. He’ll be sharing some of the specific benefits his company has experienced since implementing Web performance monitoring solutions.

If you’re interested in learning more about the latest trends in Web performance monitoring and how it can improve your business, be sure to register today. We’re offering two chances to participate: 10 a.m. or 6 p.m. EST. Hope to see you there!

And the popularity of online banking continues to grow. comScore’s 2010 State of Online Banking Report found that 64 percent of Americans now pay bills online, up 19 percentage points over 2009.

For those customers entrusting the management of their finances to the Web, and for the banks that want to maintain these customer relationships, performance is paramount.

Remember last fall when TD Bank experienced complications integrating the computer systems and websites of the old TD Banknorth and Commerce Bank? A glitch in the system prevented deposits and withdrawals from being posted to customers’ accounts. Customers couldn’t see their current balances and began to fear their bank was going out of business.

The slightest abnormality or delay in a bank’s website performance could put consumers on edge. Availability and quick response times are a must for instilling confidence and creating happy customers.

So how did the websites of some of the industry’s top financial institutions measure up last quarter? AlertSite’s Q2 Web Performance Index for Financial Services shows they performed particularly well. Seven of the sites being monitored delivered consistent or improved availability over the first quarter, and nine improved response times. (SunTrust Bank improved response time by over 56 percent!)

Two sites – Fifth Third Bank and BB&T Corp. — experienced no outages and were available 100 percent of the second quarter. Five banks gave their customers near-instant access to their sites, delivering response times under two seconds.

The graphs below reveal the average response times, availability, and quarterly improvement each site experienced in Q2.

According to Forrester, 59 percent of all U.S. Internet users now use social networks, and 70 percent consume content on social media and social networking sites. And Nielsen data indicates that people are spending nearly five-and-a-half hours per month on social networking sites.

Overnight these tools have become our lifeline to the world, delivering real-time updates on news and events and connecting us with former classmates, colleagues, celebrities, and total strangers. As more people come to rely on social networking websites, performance can suffer.

Today, we’re releasing the Q2 Web Performance Index for Social Networks. For the past 90 days, we monitored the availability and response times of Facebook, LinkedIn, YouTube, MySpace, and Twitter. Which performed the best?

MySpace bounced back from a terrible first quarter and improved its availability by over 13 percent. The result was a nearly impeccable 99.98 percent availability. Only four errors were recorded in 20,450 attempts to access the site during the three-month period, sending the network straight to the top of the leaderboard.

Based on response times, Facebook (which, any day now will reach the 500 million user mark) takes top honors for the second consecutive quarter. On average, users had to wait only 1.02 seconds for the site to load. LinkedIn came in a close second, with an average response time of 1.28 seconds. Both times represented improvements over the first quarter.

Twitter found itself on the bottom of both lists, delivering the longest average response time and the lowest availability throughout the quarter. Performance dropped significantly compared to Q1. The time it took the website to load nearly doubled, and availability slipped nearly half a percent.

During Q2 we witnessed a worldwide Internet event — the World Cup — which began on June 11 and carried through into the current month. In preparation for the event, Twitter created a special section on its site to highlight the key matches, promote special hash tags to show team allegiance, and offer special new backgrounds to decorate profiles with. However, the site was ill-equipped for the volume of traffic it would receive.

The World Cup was responsible for the site’s largest period of sustained activity in its history. During the final 15 minutes of the championship game, there were more than 2,000 tweets per second, and 3,051 tweets per second when Spain scored the winning goal.

The result of this high traffic was outages and a lot of fail whale sightings.

Twitter’s experience demonstrates the effect worldwide events such as the World Cup can have on a website, particularly when it has not prepared in advance. As demand for real-time information increases, consumer expectations for the time it should take a website to load follow suit. The performance of social sites must scale to meet these demands.

Based on websites’ uptime and average response times during the three-month period beginning April 1 and ending June 30, the three retailers with the overall best website performance were:

1. Avon, which delivered 99.99 percent uptime and an average response time of 1.57 seconds.
2. QVC, with 99.97 percent availability and an average response time of 2.03 seconds.
3. Cabela’s, with with 99.96 percent uptime and an average response time of 1.88 seconds.

In addition to being the overall best performing site, Avon was also the fastest retail site in Q2. Its average response time of 1.57 seconds was the fastest average response time among the Top 50 Retailers. L.L. Bean, with an average 1.60-second load time, and Peapod, with 1.75-second load times, came in second and third.

In terms of availability, one site – Williams-Sonoma — experienced no downtime at all, delivering 100 percent availability (a 33 percent improvement over Q1) to its customers throughout the spring and early summer. Close on its heels were Amazon and Avon, with 99.99 percent availability each.

We also performed a quarter-over-quarter comparison to determine which sites had improved their performance over the first quarter of 2010, and the results were overwhelmingly positive. Nearly all of the companies being monitored (43) improved the response times of their homepages, and 31 of them improved both their overall response times and percent availability.

The most dramatic improvements in response times were:

• Blockbuster, which improved response times 66.77 percent, from 9.60 seconds to 3.19 seconds.
• Musician’s Friend, which improved response times 58.81 percent, from 5.73 seconds to 2.36 seconds.
• Cabela’s, which improved response times 46.59 percent, from 3.52 seconds to 1.88 seconds.

The websites that improved availability over Q1 were:

• L.L. Bean, which increased uptime 3.86 percent, for a total 99.34 percent availability.
• The Home Depot, which increased uptime 3.72 percent, for a total 99.19 percent availability.
• Sony Style, which increased uptime 2.14 percent, for a total 99.92 percent availability.

We’re encouraged by the improvements so many of these retailers have demonstrated throughout the first half of 2010. The impact a quick, reliable site can have on business is becoming hard to ignore, and downtime is becoming more and more inexcusable.

Retailers have two very important shopping seasons upon them, back-to-school and the holidays — perhaps the ultimate test of a retailers’ Web performance savvy. For the sites on the bottom of these performance rankings, now is the time to act. Review your site’s performance, analyze its components, determine its limitations, and make the necessary adjustments. You don’t want to be caught with your store closed when the holiday crowds come charging.

In case you’ve missed any of our weekly recaps, AlertSite has been monitoring 11 World Cup websites over the past four weeks to track performance, response times and availability.

Spain claimed the title for World Cup winner on Sunday, and now we are ready to announce our website winner for the best World Cup performance and availability.

Drum roll please, or in this case, cue the vuvuzelas…

The winner is … CBSSports.com. Over the last four weeks, this website had the lowest average response time–3.96 seconds– and delivered 99.93 percent availability.

Second place goes to ESPN360.com, with an average response time of 4.39 seconds and 99.83 percent availability.

And our third place recognition goes to ESPN.com which had a response time average of 5.04 seconds and 99.83 percent availability.

Here’s the full breakdown of how all the sites ranked for availability and response times.

Noteworthy observations:

Fanhouse.com saw a rather consistent increase in their response times as the World Cup progressed indicating that as more people hit the site, it started to feel the strain.

YahooSports.com had the most trending and inconsistent performance throughout the tournament. You can see all the peaks and valleys of their performance graph.

FIFA.com showed the most consistent response times across all of AlertSite’s monitoring locations.

The World Cup brought a record amount of traffic to websites. Some were able to handle it better than others.

These worldwide events show how important it is for websites to test and prepare so they can maintain uptime and provide reliable service to their customers. Website performance DOES matter!

Our Week 4 performance update saw a number of sites with 100 percent availability. A few more jumped on the bandwagon this weekend.

Eight of the 11 sites delivered 100 percent availability including BBC.com, CBSSports.com, ESPN.com, ESPN360.com, Fanhouse.com, Fifa.com, FoxSports.com, and YahooSports.com. Seven out of the 11 websites also showed strong response times over the weekend.

There is no question that people were going online for updates and to view game footage this weekend. Just take a look at the increased response times that FoxSports.com was experiencing.

ESPN.com also had big spikes in response time during the hours that the final game was being played.

The World Cup drew record numbers of online visitors to websites over the last four weeks. Next we’ll have our re-cap of website performance throughout the entire tournament and reveal which website will be crowned as delivering the best World Cup performance.

So what happened in Week 4?

CBSSports.com, ESPN360.com, Fanhouse.com, FIFA.com and FoxSports.com experienced the most impressive victories. These five sites were available for their users 100 percent of the time.

Some of the biggest performance blips were experienced on Wednesday following Germany’s loss to Spain. ESPN.com reported its slowest response times immediately following this match. You can also see the spikes and increased response times experienced by WorldCupBlog.org on this date.

These examples again demonstrate the effect increased traffic and activity can have on a site’s performance. It’s no surprise that as the excitement grows, websites need to deliver optimal results to retain visitors.

Check back next week when we’ll review performance throughout the entire month-long tournament and crown one website winner for the best and most consistent World Cup coverage and site availability.

Last week, many of the sites saw their performance drop slightly over Week 1. And for just over half of the sites being monitored, Week 3 was even worse. Six of the 11 sites being watched had their worst availability and slowest response times of the whole tournament – BBC.com, ESPN.com, ESPN360.com, Fanhouse.com, FIFA.com, and Ustream.tv.

That said, all of the sites except one (BBC.com) demonstrated availability in the 99th percentile.

And in many cases, the outages that did occur did not take place during game time, meaning other issues might have caused the disruptions. Examples?

BBC.com had zero percent availability from 9:15 p.m. to 11:15 p.m. EST on June 30. Similarly, ESPN360.com had zero percent availability from 3:15 p.m. to 3:45 p.m. EST on July 1.

The chart below shows how we have tracked the response times and availability of sites throughout the tournament. Only a week to go until we can declare an overall winner!

At the core of the Red Cross disaster relief efforts is its internal SharePoint system, made up of 2,500 websites that Red Cross employees and volunteers rely on to transfer information between volunteers on the ground and officials at the Washington, D.C. Disaster Operations Center. These sites receive about 20 million hits a month. Back in October, following a number of devastating natural disasters in Indonesia, the Philippines, American Samoa, and Samoa, the Red Cross came to AlertSite because they needed to ensure these internal websites were operating at full capacity without any downtime.

We quickly deployed DéjàClick so the Red Cross could monitor the end user experience and receive immediate notifications if a site’s response time or availability fell below a determined threshold. The Red Cross technical SharePoint team quickly realized that many of the sites were being overloaded with traffic. They were able to respond by scaling the load to improve performance during this critical time and ensure users from anywhere in the world could access the websites and valuable resources.

If you’d like to read more about how our AlertSite solutions are being used as part of the Red Cross’s disaster relief efforts, we’ve posted the full case study.

Almost all the sites, with the exception of FoxSports.com and Justin.tv saw a decrease in availability this week. Though, in most cases, the drops were less than half of a percentage point.

Response times showed more fluctuation. However, like last week, there continues to be a noticeable correlation between game times and response times. When games are in progress, it takes users significantly longer to access one of the sites.

Noteworthy observations from week two:

• FoxSports.com: Even though the site had a high availability last week, at some points on June 20, June 21, and then again June 27, response times reached nearly 100 seconds. These long load times could have driven some visitors away. (See the chart: FoxSports.com World Cup Week 2 performance)

• WorldCupBlog.org: The site experienced drastic increases in response times last week when the U.S. defeated Algeria in group play on June 23. For some users, the blog took over a minute to load. Some errors were also experienced during that time. (See the chart: WorldCupBlog.org World Cup Week 2 performance)

• Justin.tv: It looks like fans may have renewed faith in Justin.tv. Load times for the site soared on the first day of the tournament, but stabilized last week around 11.38 seconds, with no major spikes. This week, average response time dropped to 10.93 seconds, but there were a few minor spikes during game time. (See the chart: Justin.tv World Cup Week 2 performance)

• Fanhouse.com: Response times for the site jumped starting on June 21, when they reached almost 10 seconds. Week-over-week, the average load time jumped 35.15 percent. (See the chart: Fanhouse.com World Cup Week 2 performance)

Check back next week for additional updates. It’ll be interesting to watch how site performance changes as the stakes rise and more fans head online for live games and updates.

A few months ago, Steve noticed the bounce rate on his e-commerce site was rising. The home page was taking between five and six seconds to download, and he suspected this was likely driving visitors away before they had a chance to shop. With approximately 20 percent of the company’s revenue coming from the online channel, Steve wanted to improve the site’s responsiveness and provide a better user experience for shoppers.

After reviewing the Web performance monitoring solutions from AlertSite and several competitors, Steve deemed AlertSite’s the most cost-effective and best suited to meet his company’s needs. With the ability to overlay performance metrics with Google Analytic’s traffic data, Steve could easily correlate how traffic was affecting performance and vice versa.

We don’t want to spoil the whole story though! See below for our video interview with Steve. He explains how he’s using AlertSite’s services to improve Supreme School Supply’s website performance.

The combination of these elements has been negatively impacting the user experience. June has been the worst month on record for Twitter in terms of performance.

As you can see in the chart below, Twitter is still having serious problems with their site. From 7 a.m. to 1 p.m. EST today, Twitter only had 73.74 percent availability.

We caught the fail whale 115 times out of the 438 samples we collected across six U.S. monitoring locations during that six hour time frame. The login process took on average more than 10 seconds to fully render the page. This normally takes between three and four seconds.

We’ll continue to watch this for any changes.

All the sites we’re monitoring have been nearly perfect when it comes to their availability for fans, with CBSSports.com and FIFA.com scoring a perfect 100 percent availability throughout the tourney so far. Here’s the breakdown:

The increased response times of the sites correlate to when the games were on. Take a look at Yahoo! Sports, for example.  Look at all the ups and downs. When games are being played,  response times are 12-13 seconds. But when there aren’t any games on, response times are around five seconds – less than half as long.

Or consider WorldCupBlog.org. There’s a lot of activity on the site while the games are being played. The increase in response times is evidence of that.

Ustream.tv patrons saw an increase in the time it takes to load the homepage too.

Things have seemed to calm down on Justin.tv since last Friday when record Internet traffic attributed to the World Cup slowed the site down. Response times on the site jumped from 10 seconds to 40 seconds at 10 a.m. when the first World Cup game kicked off, and reached as high as 90 seconds at other points during the day.

We’ll continue to watch and report on the performance of these sites as the tournament progresses. Let us know if you experience any issues.

Graphics associated with this post:

• Chart: World Cup Online Performance, Availability and Response Times, Week 1
• Ustream.tv performance chart: World Cup Week 1
• WorldCupBlog.org performance chart: World Cup Week 1
• Yahoo! Sports.com performance chart: World Cup Week 1

“It’s pretty much inexcusable for a retailer to go down during a critical point in the holiday season, and every holiday season, we see that, and in many cases, it’s because of volume.” — Larry Freed, President and CEO of Foresee Results

A few weeks ago, we shared with you the results of ForeSee Results’ Top 100 Retail Satisfaction Index 2010 and demonstrated how the Web-based retailers with the highest levels of consumer satisfaction also ranked among the best performing. Well, at IRCE last week, we were lucky enough to meet up with Larry Freed, President and CEO of ForeSee and a speaker at this year’s conference.

Last Thursday, Larry and Internet Retailer’s Kurt Peters presented Internet Retailer’s Top 500 Guide, a ranking of the largest online retailers, offered insight into what it takes to create a winning site, and highlighted the trends retailers should follow to create winning online strategies and loyal, satisfied customer in today’s economy.

When we caught up with Larry, we asked him to share with us the Web qualities that are most important to consumers. In the following video, Larry shares the key to online success and gives his perspective on Web performance and the role it plays in creating a satisfying Web user experience.

Thanks again to Larry for sharing his perspective with us. For more of Larry’s thoughts on customer satisfaction, be sure to check out his blog, FreedYourMind. You can also follow @foreseeresults or Larry Freed on Twitter at @larryfreed.

For weekly updates on the performance and availability of the Internet Retailer Top 50, be sure to review our Retail Business Benchmarks.

Today we’d like to show you some of the images we saw in Chicago. Below are a few of our favorites. To view the full album, visit our 2010 Internet Retailer Conference & Exhibition album on our Flickr page. Enjoy!

Gary Beerman at work in the AlertSite booth

Ken Godskind explaining Web performance to show attendees

Overhead view of the exhibit hall floor

Our blogging team traveled out to Chicago with us and walked the show floor, talked to other exhibitors and attendees, and gathered their perspectives of Web performance and how performance monitoring fits into broader e-commerce trends.

We were lucky enough to get a few minutes with Kurt Peters, Editor in Chief for Internet Retailer. In addition to overseeing and contributing to Internet Retailer and InternetRetailer.com, Kurt also organizes IRCE and Internet Retailer’s other conferences.

Watch the following video to learn more about the theme of this year’s show, Time to Reboot. Hear Kurt talk about the challenges retailers face in the new economy, how shopper behavior and expectations have changed, and how retailers are performing to meet those expectations and create happy shoppers.

Thanks again to Kurt for taking the time to meet with us at the show, and we hope you enjoy the interview!

The days when shoppers would wait several seconds for a Web page to load no longer exist. Today, online merchants have one or maybe, if they’re lucky, three seconds to deliver a page to their visitors.

With those statistics in mind, the conversation easily turned toward a discussion of best practices and tips Jake’s readers — SMBs — could employ to speed up their websites.

For smaller businesses that might not have the budgets or brand recognition of some of the market leaders, a website offers the opportunity to be on an equal playing field with their larger competitors, provided the site performs and responds well.

Jake published his notes from our conversation in his most recent blog post, “Speed Up Your Website to Keep Visitors Happy.” Some of the advice you’ve seen here before, but I also added an additional tip, and I recommend you head over to the site to check it out.

Thanks again to Jake for taking the time to talk and hitting on this important topic.

DéjàClick does indeed support Mac OSX natively. In fact, DéjàClick works well on all of the platforms that Firefox supports, including Linux, Solaris, FreeBSD and Mac OSX, as well as Windows. As far as I am aware, DéjàClick is the only Web transaction monitoring technology that has the capability to run on the Mac operating system. With so many businesses now moving to the Mac platform, being able to manage website performance right within the native operating system is definitely an appealing feature for Mac users.

While I must confess to owning a Macbook Pro (that the kids have taken) and an iPhone, as well as Apple’s latest toy, the iPad, I don’t consider myself a fanboy of any particular operating system, and have 4 or 5 Windows machines of various flavors as well.

But back to the point of my blog post. If you are a Mac user, are interested in Web application performance and would prefer not to have to boot into Windows to use “other” tools…check out our DéjàClick product that works natively on your Mac. It’s a free add-on for Firefox that you can download here. We also have a very active DéjàClick users forum if you have any feedback, questions, or observations to share.

Apple has been giving Adobe’s popular Web video software, Flash the cold shoulder. In April, Steve Jobs published a little manifesto offering Apple’s thoughts on Flash. Steve provides a litany of reasons that Apple devices, particularly the new iPad, do not support Flash, including proprietary and reliability issues, to suitability for mobile devices. I don’t necessarily disagree with all of Steve’s points, but I also think there might be something else going on behind the scenes.

One hypothesis I have is advertising-related.  Much of the advertising that appears on big media sites today is Flash-driven. Apple’s decision to not support this format could be their way of creating a large enough crack for them to slip into the ad game big time. (iAds, anyone?)

But, I digress. Adobe’s Flash is reported to power some 75 percent of today’s Web videos and is installed on 95 percent of Internet-connected PCs. While naysayers have suggested that Flash is a CPU hog, some of the benchmarks I’ve seen comparing Flash to HTML5/H.264 show that Flash is not really more resource-sapping than Apple’s suggested alternative.

As an iPhone owner and enthusiastic iPad user, I have regretted the lack of Flash support from time to time, as certain websites don’t function, or function poorly, without their Flash components. Nonetheless, I’m very pleased with the device and wouldn’t not use it because it lacks Flash support.

It will be interesting to see how the debate plays out, and the blow Apple’s positioning might take, if Google is able to get Flash to perform well and function reliably on the Android. Or if the new video format it introduced at its I/O conference, VP8, takes off.

That said, I came across the results of a survey last week that indicates an increase in online customer satisfaction. The survey, conducted by ForeSee Results, found that consumer satisfaction with the top 100 Web-based retailers rose five points this year from 2009. The winners? Netflix came in first with a score of 87 (out of 100), Amazon came in second with 86 points, and Avon.com and Apple were tied for third with 83 points a piece.

According to the survey, several factors contribute to a satisfied online customer: fair and competitive prices, variety and availability of products, usability of the retailer’s website, and accuracy and quality of information on the site. I’d go a step further and say that availability and performance of the site also factor into the satisfaction equation.

We already know that site speed can influence a shopper’s willingness to shop and buy. And if a customer can’t purchase an item they want because the shopping cart won’t load, I can’t imagine they’ll be all too satisfied with the experience.

On a hunch, I looked at the performance metrics we recorded for Internet Retailer’s Top 50 Retailers during the first quarter 2010. Not surprisingly, the most satisfying retailers in ForeSee’s survey all performed above average during the three-month period ending March 2010. Homepage availability of these sites was above the industry average of 99.17 percent: Apple, Netflix, Avon and Amazon all experienced availability upwards of 99.50 percent.

Response times mirrored this trend. On average, it took five seconds for retailers’ homepages to load. The leaders of the survey all fell below this threshold.

I’m curious as to how ForeSee Results determined the criteria for its survey, and would be interested in seeing the outcome of a survey that also took site performance into account. I wonder if the results would change at all.

On Thursday, I had the pleasure of sitting down with Fr. Robert “Padre” Ballacer of the Interop Network, and walking through a demo of DéjàClick, including its new ScriptShare functionality. The video was just posted on InformationWeek.com, so I thought I’d share it with you today.

If you’re interested in step-by-step instructions on how to record and play back a series of Web actions, receive alerts on your websites’ performance, and share your monitoring scripts with others, watch the video below. You can also visit the site to view this or other interviews and demos filmed at Interop.

Technology companies, social networking sites, and shortening tools all delivered quicker response times during this three month period compared to some of the other non-technical sectors in our benchmarks, such as healthcare, financial services, and even cell phone carriers. Those whose business relies on the Web recognize the importance of a fast website.

Let’s take a look at some of the hard numbers. Here are the average response times for 10 of the top healthcare websites.

The benchmark average is a response time of 4.9 seconds.

Now compare that to the response times of some of the most popular social networking websites during the same three month period.

And shortening tool websites during the same timeframe.

The benchmark average for the response times of these sites is between 2.53 seconds and 2.65 seconds. Only one of the 10 healthcare websites reported a response time that low.

The social networking and shortening tool websites operate in real-time. Their business rests entirely on the Web, whereas other industries, such as healthcare, financial services, cell phone carriers, etc. have other outlets for consumers to engage. But when a social network is down, it’s down, and when it’s slow, users don’t have the patience to wait.

When you look at sites such as Facebook, LinkedIn, and Twitter, all with average response times between one and two seconds, it’s clear that they have an understanding of how important the performance of the website is for the overall user experience.

If you are interested in learning more about how we compiled this data, you can read about that here.

In July, we’ll check on the quarterly performance data for these sites from April through June. It’ll be interesting to see if the technology sector continues to outperform or if another industry surprises us.

To use the updated beta dashboard:

1. Navigate to your current dashboard
2. Expand the yellow “Beta AlertSite Dashboard” link
3. Click the “Give it a Test Drive” button

This release includes significant upgrades to the general display and reporting functionality, and absorbs all of the content from the traditional Monitoring Console.

Here’s the list of the most interesting changes. The newest release:

• Consolidates all “Monitoring Console” device information into “Dashboard”
• Includes the ability to navigate directly to device configuration from the Dashboard
• Features new Security, ServerAgent, and Load Testing tabs
• Adds the +/- capability to view results by monitoring location
• Improves default charts display
• Has new mini-charts (4-wide) functionality
• Lets users drill down from 25-hour Availability status directly to Detail Report info

There are many other improvements too. And remember, there is a feedback link in the Dashboard. Please submit your input so we can continue to make this functionality even more meaningful to our users.

While there are many things that can impact website speed and performance and influence the user experience, the tips below represent some tried and true ways to begin to improve.

1. Put those blocking JavaScripts at the bottom of the page. As much as organizations are getting pretty savvy about Web performance and the user experience, it never ceases to amaze me how significant online brands sometimes place their visitor tracking scripts on the page before the significant content loads. In my own experiences, I have noticed two or three second (or longer) delays while waiting for the visitor tracking vendor to complete the tracking request.

2. Improve loading time of your key page content. Today’s websites are delivering more and richer content than ever before. As the operators of our websites and the stewards of performance and the user experience, we must systematically implement methods to deliver that rich content while still delivering a great end-user experience. Here are a few best practices:

• Minimize graphic page assets in general to make them as small as possible while maintaining the required image fidelity.
• Enable content compression in your Web server to minimize data transfer times on the Internet pipes.
• Set very far-in-the-future Expires dates for static content, so repeat visitors are not transferring the same content over and over.
• Consider utilizing a Content Distribution Network (CDN) for large dynamic content.

3. Put style sheets at the top of the page content. The formatting of most of today’s Web pages is applied using styles saved in a CSS (Cascading Style Sheets). Many modern Web pages look very poor or completely unreadable before CSS styles are applied. By loading CSS first, the page can be rendered partially and correctly as the rest of the page resources are loaded.

4. Minimize the number of HTTP requests. A significant portion of the time it takes to load a Web page is the time spent downloading all the resources of the page — the HTML, CSS files, JavaScript, images, etc. By reducing the number of individual resource requests, you can significantly speed up the page load time. One great example of this is to use CSS sprites, which basically are a larger image that contains, say, all of the buttons used on the page. With CSS sprites, all buttons are transferred to the browser with one HTTP request, rather than many.

5. Consider deploying Web acceleration hardware. For large-scale sites, it might make sense to deploy Web acceleration hardware like Crescendo Networks‘ Application Delivery Controller (ADC), which can improve the user experience by compressing content, TCP, SSL offloading, and acceleration, as well as enhancing security by providing better protection against Denial of Service attacks.

Today, with the release of DéjàClick ScriptShare, we’re bringing Web performance monitoring and management into this discussion of collaboration, knowledge sharing, and retention.

Previously, Web performance management was largely an individual responsibility, with one person testing scripts and responding to errors. ScriptShare turns it into a team affair, enabling employees anywhere to contribute to the task. Users can now take any click stream they record and save with DéjàClick, upload it to a private Script Repository, and e-mail it to colleagues who can play and edit the script at any time.

This ensures that each member of your IT team is aware of your organization’s Web performance monitoring testing scripts and strategies. Alternately, it enables management to see first hand the proactive measures you are taking to optimize uptime and responsiveness of your Web properties. Plus, if you ever leave  your company, you’ll be able to transfer management of these properties to a colleague or successor so that monitoring can continue seamlessly.

ScriptShare also increases your mobility. If you’re at home or traveling without the machine you recorded the initial scripts on, you can now access them from any Internet-enabled computer. You’ll be able to respond to any errors you receive immediately, rather than having to wait until you return to you primary work station. And if we’ve said it once, we’ve said it a thousand times: Speed is crucial when it comes to delivering a quality online customer experience.

So I invite you to test it out and let us know what you think. Really, the uses of ScriptShare are endless.

What are the major takeaways from Twitter’s successful web redesign? Consider these two key factors when editing or adding new content to your own sites and blogs.

1. Consumer bandwidth. Adding more interactive content and real-time updates to your site may create a more visual and unique experience for your visitors, but be aware that the people you’re targeting might not have the bandwidth to load all the new attributes and graphics. If they don’t, the page will load slowly or not at all, and users might get a time-out error instead.

One way to ensure that your fans  can still digest your page, is to offer them the option to “turn off” certain graphics. Users that can’t load the new features or are continuously being met with time-outs, can opt to view the web page in a more textually based side (versus the newer graphically based version).  This might require additional resources, however, so restricting the types or frequency of updates to your site could lessen the burden.

2. Optimization. Make sure the content you are adding, whether it be images, Flash video, third-party content, etc., is optimized so as to minimize the response times. This goes beyond making sure the homepage loads quickly when a user types in the URL, however. You also need to optimize all interactive aspects of your site. For example, if you have search fields where users can type in text, and the system is interpreting data more slowly than the user is typing, there can be type-over. That’s not necessarily the kind of experience you want your users (potential customers?) to have.

All of the sites performed well, with TaxAct.com taking the gold medal on performance. From April 9 through this morning, the site was available 100 percent of the time, and was delivering 1.73 second response times.

On a hunch that people would be filing for an extension, and curious to see if the IRS was prepared to handle the anticipated heavy flow of traffic, we used DéjàClick to record a two-step script, simulating actual users going first to IRS.com and then clicking on the link that would take them to the page to request the extension. But there were few errors here, too. The page experienced 99.93 percent availability and 10.7 second response times (fairly typical for a two-step process).

Reducing operating expenses has been a focus of many organizations during the last few years, and the services contract should provide a high level of expense predictability. There are also strategic objectives that outsourcing IT operations can help achieve, like becoming a more efficient global operation, to be considered. And finally, who in management doesn’t like the “one hand to shake, one neck to break” idea to avoid internal finger pointing?

Earlier this week, I read about Microsoft’s outsourcing deal with Infosys Systems in India. Microsoft is not the first or only significant IT vendor whose products are an important part of many enterprise infrastructures to outsource. What I really struggle with is this:

If these are the vendors selling us the software,  hardware, managed services, tools, and best practices we are depending on to run our IT operations, and they have surrendered their own IT operations to another, dare I say, more efficient vendor, well…what does that say?!

What do you think?

Find out:

• How fast is your site?
• How does your site compare to your competitors?

The change is indicative of a common perception among Web users that the search isn’t complete until the data has been received. In other words, it doesn’t matter how quickly Google, Yahoo!, Bing, or other engines retrieve your results. It matters how quickly users can obtain the information they’re looking for. If users click on the first or second links returned in their query, and the site takes a long time to load, the users perceive this as a reflection on the quality of the search engine, not the end destination site.

With this change, Google is trying to use its leverage to force their customers (i.e. online companies) to do better and improve overall experience. In effect, it will enable Google to more quickly connect its users with information, thereby improving perception of its own site, setting itself above the competition, and positioning itself to become THE default search engine.

What does this mean for online businesses? As we’ve mentioned before, organizations are really going to have to evaluate and optimize all of the content on their sites — images, Flash videos, third-party content, etc. — and in the case of this third-party content, demand greater delivery and performance. The tricks companies have used in the past to beat the system and boost their Google rankings will now have less of an impact. Speed, a cut-and-dry factor that can’t be fudged, will be the newest determining factor.

Google hasn’t publicly defined what the speed thresholds will be, but it’s safe to say that quicker is better (for customer satisfaction and conversion rates, too). If you can outload your competitors, you’ll likely beat them in the rankings (and make Google look good while you’re at it).

Check back soon for a few free resources you can use to determine where you fall on the speed spectrum.

The servers that host this data will require constant monitoring to ensure that the data is safe, secure, and continually accessible to patients, health care providers, and other authorized individuals.

Security will be a big concern because many PHRs will probably be stored on commercial Web sites, and some of these host sites might have business relationships with drug companies or other parties with a financial interest in a patient’s personal health data.

It’s a complicated picture that will take years to become clear in all its dimensions. But we can be sure that with the coming changes in our national health care system and the migration of patient data to the Internet, there will be a wide range of issues requiring Web monitoring and security to ensure patients’ privacy is adequately protected, and health care providers have consistent and secure access to the information they need to provide optimal treatment.

“The homepage now features a set of algorithmically-selected top tweets that automatically appear every few seconds. It also highlights a random sampling of suggested sources; hover over any of them to see a profile summary and their latest tweet. Trending topics now scroll across the page, allowing us to present a large set of trends using little page real estate. Hovering over some of these trends will show a description explaining why the keyword is (or has recently been) popular.”

Twitter’s no stranger to performance issues. The micro-blogging site often displays an inability to scale to accommodate heavy levels of traffic, such as what was experienced this year during the Super Bowl and immediately following the devastating earthquake in Haiti.

So I think the major question as Twitter incorporates all of the changes is, how will the new features on Twitter’s homepage affect its ability to respond quickly and stay online for users? Stay tuned …

STR was considering upgrading its hardware and contracting with a Content Delivery service, but needed to justify the expense. With the help of our global performance monitoring, STR was able to determine that their audiences in Asia and Europe were experiencing time deficiencies on the site in question. Our performance data from these locations served as the justification to start exploring CDN options.

Thanks to our friends at STR for sharing their experience and for the unsolicited kind words!

Here are five of the most common Web errors we see, and that you’ll likely want to avoid.

1. HTTP 503 Service Unavailable. This error is seen when larger Web sites have capacity or maintenance issues. It indicates the Web server is up, but just minimally, and it can’t fulfill the request. You might be familiar with the “Fail Whale” on Twitter. This is an example of an HTTP 503 error.

2. DNS (domain name services) error. This means the Internet’s directory system doesn’t know how to find the site you’re looking for. It usually indicates a mistake with DNS entry for that Web site has occurred.

3. No response to Web page request. This error indicates the Web server hardware is down or the HTTP services are not running. The HTTP services could also be overloaded from, say, a denial-of-service (DoS) attack like Twitter experienced last August.

4. HTTP 500 Internal Server Error. This usually indicates a problem with the Web application, as opposed to the Web server. Hardware and Web services are functioning, but there is some sort of problem trying to execute the logic required to produce the page you are requesting.

5. Down for maintenance. This means exactly what it says: The Web site is over capacity or down for maintenance. Web site operators are getting better at protecting users from ugly errors, such as some of the ones described above. Instead, they have a friendly message appear to let them know what’s going on. Twitter’s come up with some pretty unique ones for this error as well.

So how have sports sites performed thus far in the tournament? First, you’ll have to indulge in a little NCAA analogy. You may not realize it, but the roads to NCAA and Web site success are actually quite similar.

GAME PLAN FOR A SUCCESSFUL NCAA TOURNEY

Practice, Practice, Practice.

For months, the 64 men’s basketball teams that made it to the Big Dance (and countless others who didn’t make the cut), have been hitting the court and the weight room daily to make sure they are in tip-top shape and playing at peak performance. So, too, should have been the announcers, TV crews, and Web sites involved in producing the event. For popular sports Web sites such as ESPN.com, NCAA.com, and CBSSports.com, this means testing the capacity of  the sites to ensure it would remain up and running smoothly for the millions of fans using the Net to follow their favorite teams and check their brackets.

On Thursday, the first day of the tournament, 3 million unique visitors logged onto the Web for CBS’s live on-demand coverage of the games. During the first four days of the tournament, more than 6 million fans logged on to stream 8.7 million hours of live video and audio. So did the site hold up to this record-breaking traffic? Stayed tuned to find out.

A well-prepared coach that can handle to the pressure.

Every company Web site needs a strong IT guy to lead team practices, react to adversity, and ultimately, lead the team to victory. Designate one person to receive regular performance reports and make adjustments should your site not be responding as well as you’d like. And don’t panic. One missed shot won’t lose you the game. An effective team leader will stay calm, grab the rebound, and pick your site back up.

The B team.

Ideally, you want to have your best talent on the court at all times, but what happens when they get tired or enter into foul trouble? Time to make a substitution. Have a back-up plan waiting on the sidelines that you can throw in should your initial game plan fall apart. A quick personalized message should do the trick.

THE ROAD TO THE SWEET SIXTEEN

So with the first two rounds of the tournament behind us, how did the Web hold up? Surprisingly well. The majority of the sites we were monitoring have reported relatively consistent response times in the 0-25 second range. Only one site, FoxSports.com, experienced truly irregular response times and sporadic slowdowns throughout the weekend.

OUR PICKS

Based on performance thus far, we’re going to have to go with mmod.ncaa.com as this year’s champion. With the exception of a few blips late Friday night, availability (on average, 97.92 percent) and response times (on average, 9.21 seconds) have been steady throughout the first two rounds. With fewer games to air in each upcoming round, we anticipate the strong performance to continue.

As for the teams on the court, I’m going to pick Kentucky. They’re by far the most athletic group, and they demand performance!

We spoke recently with edo Interactive, a company that develops technologies that use payment card platforms for digital marketing purposes, such as the placement of coupons on debit, credit, or prepaid cards. They began using DéjàClick in May 2008 to monitor their customer facing sites, corporate and internal Web sites, and third-party partners and APIs hosted by the company.

With hourly transaction monitoring and AlertSite™s real-time alerts, edo was able to detect and correct a problem in the backend of a Facebook application before it was launched. It was an error that would have jeopardized the company™s investment and customer experience.

edo Interactive has been able to ensure customers get better than 99.7 percent uptime and the fastest possible performance on its site as well as sites through its partner ecosystem.

We recently posted the full case study to our Web site and invite you to take a look.

And, in case you missed it, edo shared their AlertSite experience recently with Inc.com. It’s a great read about how Web site monitoring can benefit young companies.

If you’re an AlertSite customer with a great story to share about your Web performance experience, contact us so we can feature you in one of our upcoming case studies!

At that time, I explained to Howard the basic principles of Web performance monitoring, and how important it is for any company with an online presence to check the availability and performance of their Web properties.  I then gave Howard a demonstration of DéjàClick 2.0, which records a user’s path and click stream through a site. We recorded a script that included clicking on links, entering user names and passwords, and loading Flash objects. Then I explained to Howard how AlertSite’s Web performance monitors could test a user’s site using this script, and report back on load times and overall availability.

Just this week, Howard connected with one of our customers, edo Interactive. He talked to CTO Jay Graves about how edo uses Web performance monitoring to ensure a quality Web experience, which they define as greater than 99.97 percent uptime, for their customers.

I recommend checking out Howard’s column this week for his take on Web performance monitoring, and how it can benefit young companies. He does a great job of articulating what we here at AlertSite truly believe: Website Uptime Means Happy Customers. Thanks so much for sharing our story, Howard. We couldn’t have said it better ourselves.

The dashboard, which has been in beta for the past six months or so, offers some great new features, including:

• Drag ‘n drop performance graphs that can be arranged to meet individual user needs.
• Autodisplay options that automatically open graphs showing the problems customers’ Web sites/applications are having at the top of their screens.
• Automatic monitoring intervals that are set to refresh user monitors and performance graphs every five minutes or more (based on your monitoring frequency).
• Powerful formatting abilities that allow users to sort data multiple ways, resize charts, and view multiple graphs simultaneously.
• Personalized preferences and settings that are preserved after each login.

For more details, listen to the brief podcast below. And remember, we’re always interested in your feedback, so if you have a suggestion or comment while you’re testing out the dashboard, please let us know!

Our monitoring of some of the top online florists including 1-800Flowers, FTD, Teleflora, and ProFlowers, showed that the floral retailers rose to the occasion and delivered. 1-800Flowers, ProFlowers and Teleflora all demonstrated 100 percent availability from Feb. 11 to Feb. 15. Response times were also good taking between 3.34 and 6.29 seconds for the homepages to load.

Those trying to show their love electronically this Valentine’s Day had a different experience. Someecards.com experienced several intermittent over-capacity issues between the hours of 10 a.m. EST and 4 p.m. EST on Friday, Feb. 12. This was most likely due to the fact that people were sending their cards before the weekend and the recipient had to log back into the Web site to view it.

Here’s a sample of an error message that the site displayed.

 

Overall, online retailers did a good job of making sure everyone could express their love online and off this Valentine’s Day.

Below is a copy of the AlertSite performance chart Oliver Chiang used in his story. Thanks again, Oliver, for including our data and pointing out such an intriguing trend!

The AlertSite team, including myself, will be exhibiting at Booth #1. Come visit us for advice on improving the performance of your Web sites and applications, and for a hands-on look at some of our most popular and effective Web performance solutions, including DéjàClickwith TrueScreenâ„¢ technology and InSite by AlertSite. We’ll also be revealing a new look for your performance data that you won’t want to miss.

Hope to see you there!

The vast majority of ISPs today adhere to the principles of “end-to-end” neutrality, which basically states that networks should confine themselves to transmitting general packets without worrying about its contents. This neutrality principle is one of the reasons for the Internet’s incredible growth and its power to transform existing business models and create innovative new ones.

It means that ISPs must provide equal access to all the content on the Internet being accessed by users. Network neutrality does not mean that ISPs cannot charge users more for faster service, or for using more than the allowed bandwidth of its access plans. It has nothing to do with broadband pricing or metering whatsoever.

Three Perspectives to Network Neutrality.

1. Consumers. Consumers are the ones who pay the ISP bill and generate the revenues for the big ISPs. Consumers want the freedom to access whatever lawful online content they desire. Just for clarification, there is nothing in the proposed Federal Communications Commission (FCC) regulations that suggests that consumers who use lots and lots of bandwidth shouldn’t pay more.
2. ISPs. They have spent billions to build out the networks that deliver all this content to consumers. Some of the ISPs suggest that companies like Skype or Netflix are getting a free ride on their networks. I couldn’t disagree more. The consumer is the customer of the ISP, not the Web site, application, or content producer.
3. Web Content Producers. For this group, network neutrality is essential to enabling innovation and new business ideas. Without network neutrality, ISPs could limit consumer access to content or services that don’t have an agreement with the ISP to deliver its content (i.e., don’t pay for access to consumers or some other sort of corporate affiliation). This would essentially turn the ISPs into the content distribution mafia. Without network neutrality, content creators who are not connected with large media distribution outlets would not be able to reach end-users or only with potentially reduced performance.

Why do we need guidelines?

Except in a few cases, ISPs today have provided mostly neutral access. This neutrality is what has allowed the Internet to be the beautiful, organic, and ever-adapting resource it has become. The purpose of the proposed guidelines and regulations is to protect this organic nature.

The concern is that if the ISPs start charging the Web content companies for access to consumers, then the Internet will become a very corporate-controlled media like TV.

On the flip side, there are also valid concerns that government regulation of anything can create unexpected consequences.

Consumers will probably have to endure a shift from the current low-priced, “all-you-can-eat” broadband to a tiered and metered model. And that’s probably a good thing so that 1 or 2 percent of all users are not using 80 percent of the network resources.

ISPs will need to continue to grow the capacity of its networks, as they have always done. After all, bits — unlike other commodities — are essentially free.

DéjàClick lets you record and play back series’ of Web actions, from opening new pages to entering user names and passwords. After you install the toolbar (available at addons.mozilla.org), all you have to do is hit record, and start performing the activities you’ll want to repeat later. When you’re done, hit stop and name and save the script you just recorded. Next time you want to perform the same actions, all you have to do is click once on the play button.

DéjàClick’s ability to simplify the steps necessary to view your favorite Web sites and perform other repetitive, online tasks earned its recognition as one of 2009′s Recommended Add-ons for Firefox.

If you’re interested in trying it out, you can download DéjàClick for free.

Thanks again to the editors at MakeUseOf and Lifehacker for sharing DéjàClick with their readers!

Twitter Login is holding up with 100 percent availability. It’s slowing down here and there, but no inaccessibility.

Note this second chart for the Apple Store. Look at the recent performance. Think everyone is going to see if they can order an iPad now?

Will update when the Steve show is over…

UPDATE: 3:50 p.m. EST

The big show finished shortly below 3 p.m. EST, so here’s the latest.

There were some pretty significant slowdowns on the Apple store page, but no errors.

And Twitter has seemed to move past any slowdowns.

Final verdict: the Apple iPad event doesn’t give Twitter indigestion.

The addition of these three centers brings the total number of data stations in our network to more than 50. By the end of 2010, we plan to double that number. You can see the full list and a map of our monitoring stations on our Web site.

Our monitoring shows Twitter’s Web site was completely down from 6:43 a.m. – 7:43 a.m. EST yesterday morning. The home page was displaying the fail whale.

The diagnostics show that Twitter’s site was responding with an HTTP 502 bad gateway error. An interpretation of that error is the Web server, while acting as a gateway or proxy, received an invalid response from the upstream server it accessed in attempting to fulfill the request.  Meaning, the back end was unavailable.

Interestingly, the outage during this time correlates closely with the 6.1 aftershock in Haiti at 6:03 a.m. EST.

Twitter is a powerful platform. However, the data seems to show that they do not have the capacity to be the 911 system for the world. It can be an early warning system at times, but as a primary means of communication, perhaps not yet.

Here’s a quick look at yesterday’s results:

After monitoring the performance of some of the top online retailers’ home pages during the 2009 holiday season, we’re excited to release our first E-Commerce Holiday Web Performance Index.

Overall, Web performance remained consistently strong, with few sites experiencing slow responses or outages. On average, the home pages of the sites monitored reported 99.89 percent availability and 4.10 second response times.

Seven retailers–Walmart, Victoria’s Secret, Target, Staples, QVC, Macy’s and Amazon–recorded 100 percent availability of their home pages throughout the season.

Starting Nov. 26, Thanksgiving Day, and ending Dec. 31, we evaluated the home pages of more than 20 retailers. Using DéjàClick
, pages were tested every 15 minutes from six locations across the country.

The absence of any serious issues indicates that online retailers are taking the performance of their Web sites more seriously than ever, and are likely preparing ahead of time for the rush of traffic the holiday season brings. Even on this season’s busiest and highest grossing days–Dec. 15 and Nov. 30 (Cyber Monday)–we observed no issues.

Kudos to those online retailers monitored for delivering a quality Web experience to customers this holiday season!

• It is the only way to generate an interaction with your Web site that exercises your application like one of your actual end-users.
• A browser also measures a more precise view of how all the elements of the site content load for the user and how that impacts the user experience.
• Since a real Web browser is driving the entire interaction by entering URLs, clicking links and buttons, and filling in form fields, the Web application is functionally tested each time a test executes.

2. Configuration is easy and efficient. A Web application performance monitoring solution should be fast and easy to use. There are many complex solutions in the Web application performance monitoring space that require a lot of tools and expertise, and are designed to be used by highly-trained technical resources. Look for a Web transaction capture and playback mechanism that is easy and natural to use and can reliably play back on the desktop exactly as it will play back in the service. This is key to allowing the business to participate in defining performance monitoring scenarios and an important part of the total cost of ownership to be considered.

3. Flexible notification and alerting drives awareness of operational issues. Providing error notifications by traditional methods like e-mail, as well as SMS, telephone call, and SNMP are important. Receiving an actual telephone call can be very useful in raising cognizance of issues during off-hours. Integrating alert notifications about Web site performance or availability issues can be integrated with existing Network Operations Centers.

4. Detailed reporting and diagnostics drive troubleshooting and performance analysis and trending. As mentioned above, using a real Web browser provides a very detailed view of the availability and performance of every aspect of today’s complex Web applications. Capturing useful diagnostics like screen shots and request/response headers can really help identify the cause of errors quickly. Network diagnostics like TCP tracerouting provide a very powerful view and should be included in any connectivity notifications or alerts.

5. Monitoring inside and outside the firewall provides a complete view of service delivery. Supplementing external monitoring with the data center perspective often speeds problem diagnoses. A complete solution can perform Web application performance monitoring for internal applications as well as those connected to the Internet.

6. Superior service and support contribute greatly to customer success with any technology.

These social networking sites are part of AlertSite’s home page benchmarks. The home page of each site is loaded at five-minute intervals from 12 geographically dispersed U.S. cities. We are using DéjàClick, which is AlertSite’s completely built-into-the-browser monitoring solution.

DéjàClick provides a true Web browser perspective of how long the pages is taking to load as well as any timeouts or HTTP errors. A more detailed explanation of the methodology is available.

We are also planning to monitor some transactional benchmarks in the future.

On Saturday night, retail Web traffic peaked at 2.9 million visitors per minute, up from 1.9 million the year before, according to Akamai’s Retail Net Usage Index.

This adds to an already (and perhaps surprisingly) successful season for e-commerce, which has seen a 4 percent increase (compared to 2008) from the beginning of November through Dec. 18. Records were also broken on Dec. 15, when single-day sales reached $913 million.

So did the rush of traffic last week and this past weekend hamper the performance of retail Web sites?

Our monitoring of Internet Retailer’s Top 50 Online Retailers’ response and uptimes shows that for the most part, retailers were prepared. All 50 sites were available more than 90 percent of the time, with 16 sites demonstrating 100 percent availability during the week ending Dec. 20. Even Saks Fifth Avenue, which came in last on the list, was available 91.2 percent of the time.

Response times were decent, too, with the majority of sites taking less than five seconds to load.

There are only a few days left in the holiday shopping season, but so far, we have yet to see any major performance issues or outages. Looks like online retailers are taking their performance seriously, and making the necessary preparations to ensure a profitable season.

For complete response time and availability data for Internet Retailers’ Top 50 Online Retailers, click here. To zoom in, click the image.

Specifically, Fritz notes that while innovation is still occurring, one of the many differences that exists between now and just a few years ago is that there needs to be a very practical nature to technology. For people to invest and spend the money on it, it needs to be able to solve a problem today.

For more of Fritz’s thoughts, click play below.

Last month, I came across a great blog post by Dmitry Fadeyev on Usabilitypost about why the iPhone is so tough to match. Reading it, I saw parallels between Dmitry’s analysis of what the iPhone does right and great Web applications.

Dmitry lays out three specific things that the iPhone gets right

1. Flow
2. Responsiveness
3. Polish

Flow is about usability and workflow design. We know Web applications that are easy to use and intuitive will be more successful. Placement and organization of menus, breadcrumb navigation, good visual feedback, and consistent use of design elements all play a key role in how well a user can interact with a Web application. Today, like with the iPhone, users shouldn’t have to read a manual to be able to interact with the application. Our DéjàClickis a great example of a product that makes hard things easy, and has very high usability.

Responsiveness is all about performance, selecting a menu item, or clicking a button. I have blogged in the past about just how important hard response time metrics are to the user experience.

Polish is about look and feel, making sure fonts are consistent and a tasteful palette is used. This is really the finished carpentry. Many of the more mature online entities  have worked really hard to deliver a little bit of their corporate style and brand experience in their Web presence. Presentation counts for a lot once workflow and performance are addressed.

So when you’re developing your next site, think about these three things that make your iPhone unbeatable by the competition, and deploy them in your design. Doing so, you’re bound to improve the user experience and consequently, help you retain visitors and convert shoppers.

Listen to the brief podcast below for some of the key takeaways from Michael’s presentation. Additional thoughts and advice from Michael about SEO are available on Slideshare. Or you can visit his blog for more on his Web 2.0 experience.

For those of you who couldn’t make Web 2.0 Expo 2009, here are some pictures that can be your eyes into the show. We took some photos from the AlertSite booth, Expo Hall, and around the Jacob Javits Center. You can view our Web 2.0 Expo New York 2009 Flickr slideshow here, or click play above.

And yet, despite surging traffic — an average of 4.3 million North American shoppers visited Web sites every minute yesterday — performance of major online retailers held steady, with only a few minor issues popping up on sites such as TigerDirect, Kohl’s, and BestBuy. Amazon.com experienced a fairly significant slowdown during the 10 p.m. EST hour as well.

Here you can see that for the most part, the response times of online retailers’ home pages remained consistent throughout the day. A few sites experienced some blips, but generally, response times were at or below 14 seconds throughout the day. (Click on the graph below to expand.)

Response Times of Major Online Retailers on Cyber Monday

With the exception of Kohl’s this afternoon, and TigerDirect earlier this morning, most major online retailers are exhibiting close to normal response times. At the higher end is RadioShack, with response times over 10 seconds throughout the morning.

Also, during the noon to 1 p.m. EST hour, there was a slight lift in response times. This is almost surely due to increased traffic and shopping activity during that time.

In terms of availability, the sites we’re monitoring showed 100 percent availability between 6 a.m. and 2 p.m. EST. BestBuy and Kohl’s are the only exceptions. BestBuy experienced a time out during the 10 a.m. hour; and during the noontime hour, Kohl’s displayed a ‘sorry the elves are busy page.’

We’ll continue to monitor and report on the performance of these sites throughout the day and holiday season.

Now that Web 2.0 Expo 2009 is over, and we’re back into our normal routine at the office, I’ll be sharing some of our thoughts on how the show went, and launching our Web 2.0 Expo coverage on the blog. We have podcast interviews with speakers, media, and show attendees; video from around the Javits Center; and photos from the Expo Hall floor.

To start, I’d like to share the slideshow that we had looping on the monitor at our booth. If you were at the show, you might have caught a few slides, but considering how busy the booth and show floor were, it’s unlikely you were able to stick around for the whole presentation. So now you can watch it here, or click on the screen cap above. (For full screen mode, click on the “full” button in the menu bar above.)

The slideshow gives you a brief overview of our company, and provides some information on our products — AlertSite Monitoring Suite, DéjàClick, InSite by AlertSite, On-Demand Load Testing, AlertSite Security Vulnerability Scan, and ServerAgent.

And it’s likely to have a big impact on how stringently organizations measure and manage the impact of third-party content on load times and the user experience.

Here are a few examples for consideration:

• Today, most news publishers monetize their content with advertising. That typically means they are using an ad syndication network like DoubleClick.
• Meanwhile, many online retailers use third-party content from PowerReviews.
• Many business-oriented sites use visitor tracking services from Google Analytics, Omniture or Coremetrics.

The potential changes at Google mean there will be a REAL business impact for poor Web site performance, and conversely, in Google’s words, a bonus for good performance. Online organizations will need to look closely at themselves and the other parties that participate in the Web application delivery supply chain to understand and manage this new development in 2010.

If you are an online retailer, the good news is that there are things you can do to ensure Web performance. Optimizing your Web site and user experience will help you optimize sales.

Tune in to my holiday podcast to hear my Web performance advice for this year. I discuss what online retailers can expect this holiday season and how e-tailers can ensure they are offering a quality user experience. When all else fails, I also offer some Plan B options online retailers can use to maintain their customer relationships.

Hope you enjoy the podcast and Happy Holiday preparations!

Historically, site speed has not played a role in search rankings. However, there are some people within Google who believe the Web should be fast and should be a good experience. Based on that logic, they’re starting to consider giving a bit of a bonus to the faster sites.

Google’s announcement brings the issue of Web page performance front and center. And as Cutts mentions, in 2010, we’ll likely see a lot of companies tinkering with how to make their sites fast and rich without writing a bunch of custom JavaScript.

Check out the video of Cutts’ interview with Mike McDonald of WebProNews for more details. Or, for an opportunity to test the speed of your own site,  click here.

The theme of this year’s show is “Power of Less“–less staff, less budget, less bureaucracy.

On Tuesday afternoon, a few of my colleagues and I will be setting up at Booth 1031 on the tradeshow floor to spotlight our own Web performance tools, including the latest version of DéjàClick. If you’re going to be at the show, be sure to stop by for a free demo and a chance to see how your Web site performs.

Record and replay a script with us, and if your Web site is one of the three fastest, with the least amount of errors, you’ll win a free iPod touch!

Hope to see you there!

In response, I wrote a follow up post for Search Engine Land to offer five suggestions for online retailers to ready their website for the holiday rush, and convert holiday visitors into loyal shoppers and customers.

For these tips on Web site testing, monitoring, and performance measuring during the holidays, check out the post.

TrueScreen is pretty incredible, however, there is another really powerful enhancement released in DéjàClick 2.0 that I think is just as exciting.

DéjàClick now includes the ability to execute custom JavaScript on the web page that is a part of a DéjàClick transaction recording.

At its simplest, custom JavaScript could be used to prompt for user input or read values from a web page into the recording being replayed. Or, it could be used for something much more powerful, like scraping a value from one page for use later as input into a Form field later in the recording.

Here’s a short video demonstration of those two examples that incorporate JavaScript into a DéjàClick recording.

 

Happy DéjàClicking!

So how can you be a good custodian of these assets? Of course, all the usual guidelines about being a good community gardener apply.

It might make sense to do some performance monitoring of your Twitter or Facebook home pages, much as you would your own Web site or applications.

Granted, you won’t have control of any performance or availability issues, but as an important asset in your evolving corporate communications toolbox, you should at least be aware of their availability and be alerted if there is an issue. That way, if these sites do go down or experience other problems, you can react and maybe boost up your other communication efforts during that period.

Think again.

Over the summer, more than half of the Internet’s top Web sites were found to have been setting Flash cookies to track users and store information about them, for example, their clicks and visits around Web sites. And only four of those sites mentioned their use of these cookies in their privacy policies.

What’s alarming about these cookies is that most users don’t even know they exist, and they can’t be controlled with your normal cookie privacy controls, meaning you’re traceable, even if you don’t want to be. Not to mention the cookies can be used to “undelete” traditional cookies that a user already deleted, earning them the nickname “re-spawning’ cookies.

Some sites have changed their policies with regards to Flash cookies, and there are some tools out there that address this clandestine tracking method, but I still can’t help feeling like my privacy is being violated in some way.

This is an interesting addition to the cookie conversation that’s currently in full swing among government, marketing and advertising circles, and I’m curious to see if any new regulations emerge as a result.

If you’re a retailer expecting or even just hoping a lot of shoppers will hit your Web store this Black Friday, Cyber Monday, and what we call Cyber Turkey Night–Thanksgiving night–you better start thinking now about how well your site is prepared to withstand the strain. If you wait until the last minute to start testing the limits and performance of your site and its functions (shopping cart, login, checkout, etc.), it’s going to be too late.

To help retailers gear up for what will hopefully be a profitable holiday season, we have three key suggestions to help ensure all shoppers will receive a high quality Web shopping experience.

1. Load testing to ensure your Web store can scale to meet expected user traffic.
2. Ongoing Web store monitoring to benchmark the performance profile of your current systems and resources.
3. Having a Plan B, so that when things go wrong–and they will–you can react gracefully and with minimal impact to the business.

Over the next few weeks, I’ll delve into each of these suggestions a little further in a series of blog posts and podcasts. Let me know if you have any specific questions or concerns you’d like addressed.

But DéjàClickeliminates these difficulties and allows you to truly experience your Web site from the perspective of its users.

Update 11/4/09: Since writing this, we posted an in-depth overview on DéjàClick to our webinar page.

No, this isn’t one of my quotes. Our marketing or sales teams didn’t say it either. This quote came directly from a blog post written by our customer, AgencyNet Interactive. AgencyNet is an award winning digital advertising and marketing agency that has worked with some of the world’s largest brands, like Ford, Pepsi, Grey Goose, etc.

The post highlights the importance of Web performance monitoring, and specifically explains how AlertSite services helped them analyze and improve their sites. The post also includes a copy of one of our reports.

It’s a great testimonial on how Web performance monitoring in general, and AlertSite tools more specifically, can optimize the user experience.

Thanks to AgencyNet for sharing their experience!

How many Web sites do you visit in any given day? And how many of them are built partly or entirely based on Flash?

Adobe claims that Flash content is present on more than 85 percent of the top 100 Web sites, and that approximately 75 percent of all Web-based videos use Flash. But video, images, scrolling text—all the features that make a Web site rich, engaging and interactive—have been a monitoring challenge for network administrators, developers and Web designers trying to create a high-quality, reliable experience for Web users.

DéjàClick 2.0, the latest version of our Web recorder, uses new TrueScreenâ„¢ technology to capture and replay Flash in real time and requiring no extra steps in the recording process. TrueScreen captures all of a user’s movements on a site…any click, drag, drop, movement or size adjustment is recorded as it is performed, and then played back at the click of a button.

This capability is rare in the Web performance industry, and it brings Web performance monitoring and management to a whole new level. Try it for yourself. The DéjàClick recorder is available as a free download for Firefox users.

It all started with the 8-second rule. Back in 2001, Zona Research released a study which stated that users will stick around for up to eight seconds waiting for a page to download.

Then, a November 2008 report by Aberdeen suggests that business performance begins to decrease after a response time delay of 5.1 seconds.

NetForecast’s APDEX uses 4.0 seconds as the dividing line as to where users become frustrated.

And the most recent study, conducted by Forrester Consulting, suggests that two seconds is the new threshold in terms of an average online shopper’s expectation for a Web page to load, while 40 percent of shoppers will wait no more than three seconds before abandoning a retail or travel site.

I do agree that the user experience is more important than ever, AND that providing a good user experience is paramount to online success, AND that this number has decreased vastly, or rather, that user expectations for what represents an acceptable page response time has increased. This is especially true of today’s broadband connected users.

I strongly believe that this response time number, the one that represents where the user experience begins to be impacted negatively, is not quite so black and white.  It is different based on where you are in the process. For example, PPC landing pages, the home page, search results, and the product catalog really need to be snappy. Latency there may have a significant impact on the shopping experience.

But there really is more to it. Once a user has mentally committed to the purchase and has started the checkout process, I think users are accepting of a somewhat slower response time. It’s ok if it takes five or six seconds to complete the login portion of the checkout process, or if it takes 10 or 15 seconds to process payment and finalize purchase.

This is especially true if the Web store or application provides good visual feedback that indicates the progress of the transaction instead of just leaving site visitors guessing.

There may not be a magic number that holds true for all Web sites but one thing we can be certain of is that a speedier Web experience is a better Web experience.

But more than a dozen major disasters in eight countries spanning around the globe can lead to major bandwidth complications, as more people try to access resources online, and the Red Cross needs to coordinate volunteer and supply efforts around the world.

To ensure that it’s Web site, IT infrastructure and internal response sites for volunteers and employees are accessible and running efficiently enough to meet this need, the Red Cross has taken measures to prevent performance degradations and service disruptions including:

1. Being proactive by assessing technology needs and addressing business requirements that are changing by the minute
2. Strengthening their IT infrastructure for speed, reliability and availability
3. Dedicating more server support to handle frequent communication with those impacted and the general public

A principal component of this effort is using AlertSite to monitor the availability and performance of the Red Cross’ internal SharePoint sites, which they use regularly to transfer information from volunteers and relief efforts “on the ground” during and after a disaster back to the Disaster Operations Center, the nerve-center of Red Cross in Washington, D.C. The Red Cross also closely monitors performance levels for the main Red Cross Web site and the Red Cross Newsroom, which serve as the primary online resources for up-to-the-minute information on relief efforts, support for locating missing loved ones and information on how the public can help.

Knowing that time is of the essence during disasters, and that the reliability of the Red Cross’ internal and external Web sites is critical to their global relief efforts, we were happy to support the organization by donating these Web monitoring services.

We spoke to Michael Spencer, Technical Lead of Infrastructure Services and Microsoft SharePoint for the American Red Cross National Headquarters, about this effort and he shared with us that, “our Newsroom is our way of communicating and getting information out to the disaster victims and their family and friends while also alerting the public about our relief efforts. Being able to ensure reliability of this site so it can be accessed at all times without complications is a huge benefit to the Red Cross and disaster victims who count on us to be there for them in their darkest hours. This is a great example of how technology is helping us take care of disaster victims.”

The American Red Cross demonstrates how an organization is successfully using the real-time Web along with the real-time features of SharePoint, and the pivotal role of Web monitoring to ensure reliability and availability of these updates. People want information and want it fast. Just look at how much we rely on Twitter and Facebook feeds to stay on top of what’s happening as soon as it happens. ReadWriteWeb gives some great examples of other technologies that make use of the real-time Web. End user experience and uptime is paramount in each of these examples. The real-time Web is here to stay and will continue to evolve. It’s up to organizations to take the necessary measures to ensure accessibility.

Here’s how it works: One credit gets you a one-time look at the performance of a single Web page from one of our North American monitoring locations. For two credits, you can check the performance of one full page from a monitoring location outside North America.

And it’s affordable–each credit costs less than a penny.

The idea behind this plan is to give customers more flexibility over their monitoring. You can ramp up during sales/promotions or pause monitoring during scheduled downtimes. You can also try different services with minimal commitment (credits can be used with any Monitoring Suite Service).

To learn more or switch over to this new model, contact your Account Manager at any time. Prospective customers can call 877-302-5378 or e-mail Sales@alertsite.com.

AlertSite’s monitoring showed about 27% availability during that time, but it was only that high because of a few positive results from some of our monitoring locations earlier and later during that window. The site was pretty much unavailable during the mid-section of that time period.

Here is a screenshot of the error message a browser would have seen:

Here is our detailed report for that time period:

Load testing allows you to stress test your systems and determine how many users your Web site is capable of accommodating before it breaks. This can help eliminate or at least control the negative effects of a site outage.

With the peak shopping season of the year only a few months away, now is the time to start preparing your sites for the holiday rush.

Join me for a FREE live webinar on Tuesday, Sept. 29, during which I’ll explain to you what Web load testing is, why it’s important, the best times to test, and where to test–either internal or external to your network.

Best part is: you’ll have three chances to listen. I’ll be hosting the discussion at 9 a.m, 4 p.m. and 8 p.m. EST.

Registration is open, so sign up now!

With info security such a pressing topic these days, particularly in banking, insurance, healthcare, government, etc., we’re always looking for ways to increase the reliability and privacy of our tools and solutions, and help businesses protect their data.

Here’s a quick video I recorded to walk you through each of DéjàClick’s new security options. Hope you find it useful!

To the dismay of myself and I’m sure many others, it appears that Twitter balked a bit under the load of all those bloggers and tweeters during the music event yesterday.  As GigaOM points out in this post, the data shows that at certain points during Steve Job’s talk, Twitter was only a little more than 50 percent available.  Apple and Steve Jobs were the top trending topics, and I was getting plenty of those dreaded HTTP 503 errors.

It’s hard to complain about Twitter availability given the price. However, we all seem to be relying on it more and more for keeping a finger on the pulse of our favorite topics.

Users are terribly fickle today, meaning landing pages, home pages and search pages need to provide very respectable page load times to offer a proper user experience and ensure site visitors stick around to see what your company has to offer.

Offering excellent performance and availability is especially important if you have invested a lot of resources in SEO to drive traffic to your site, or if you are spending money on PPC (pay per click) to drive traffic.

Ongoing monitoring by a real Web browser can help you understand the kind of page load times your site offers visitors, and how those times vary throughout the day. And if necessary, it can alert you to adjustments that need to be made to your site in order to accommodate the traffic your SEO and marketing efforts are driving.

We’ll keep you updated here on new monitoring locations and IP addresses added to our network. For example, starting Saturday, Sept. 12, AlertSite will be monitoring from an additional IP address in Mumbai, India: 202.87.46.210.

If you are an AlertSite customer and interested in seeing the full list of locations we monitor from, including IP addresses, look under the Help Menu of your Control Panel, under “Monitoring Locations.”

Cost is also a factor. Amazon Web Services and Google App Engine are already delivering the ability to build and deploy an application without any capital investment in hardware and software. Further, some experts estimate that large scale computing environments can be run at about a 70 percent lower cost factor than individual data centers for small and medium businesses.

The challenges include things like disaster recovery. How would you recover in the event your cloud vendor went dark? What about portability? If you choose Google App Engine would it be possible to port to another cloud service if you needed to? Lastly, security concerns arise, including fear about who might have access to my business data.

Despite the potential concerns, many of us have started to embrace SaaS and cloud computing, and are already utilizing hosted CRM, web analytics, Amazon Web Services or Google App Engine. Once hardware and network operations are outsourced to the cloud, the single most important responsibility is to ensure that the application delivers high levels of service to the end-users.

The most accepted method of doing this today is to setup remote external Web application performance monitoring. In an environment where every resource is virtualized (CPU, disk I/O, memory, and network), the only way to understand the end-user experience is to measure where the “rubber meets the road,” at the Web application interface layer.

But doesn’t that seem a little like the fox watching the hen house to you? I’m not saying that your hosting partner doesn’t really want to do the best they can. However, when your Service Level Agreement (SLA) states that the provider may have to provide service credit or suffer financial penalties, there is simply too much of a conflict of interest to rely solely on that monitoring.

Hosting providers usually monitor your Web site from inside the data center. The problem with this is that users come to your site from the world outside the data center. So, the tests your providers are performing won’t identify many of the potential issues that can affect the end-user experience—for example, those related to DNS, peering, routing or Internet weather.

What do they test? If your Web server is being pinged, you’ll only determine if the server is on. If your homepage is being monitored, you’ll only know that the http server is running.

These server statistics are frequently unrelated to Web application performance and availability for end-users. They don’t provide any metrics regarding the availability and performance of your applications from locations around the country or the world. It doesn’t show you if all the moving parts of your Web store, such as add-to-cart, checkout or login, are working. So then how do you know?

As one of your partners, it’s important to trust that your hosted provider’s technical resources are doing all they can to help you deliver a good experience to your users. But it’s also important to monitor and verify on your own.

If your business depends on the interactions you have with users online, then only you can take custody of understanding how Web site application performance impacts the end-user experience.