The New Definition of Web Performance Monitoring, Part 2

This blog post is the second in a three-part series that explores (WPM) from three directions:

  1. What is WPM exactly
  2. Why WPM is important to a wide variety of business functions from sales to security
  3. How WPM mitigates performance risks and helps direct future optimization. Today we’re focusing on the business of web performance.

As of this writing, Mark Zuckerberg continues to delay the launch of online video ads on Facebook. Never mind that the holdup is costing Facebook $2 million a day; Zuckerberg is worried about performance.

Issues like load times and streaming interruptions are all the more challenging when your user base of 1.1 billion people is accessing the site via all manner of devices and connection speeds. Facebook engineers have toiled over how to make the ads fast enough so as not to alienate users and, ironically, cause a social media disaster — especially during launch.

The delay comes despite the fact that some potential Facebook advertisers created their video ads in anticipation of a summer rollout. They have had to make alternative marketing plans for time-sensitive products. Nevertheless, Zuckerberg would rather disappoint the advertisers than his all-important user base.

Zuckerberg is likely familiar with some of the more spectacular web performance disasters, as many were front-page news around the world.  For example, in April 2011, the Olympic Games in London made 6.6 million tickets available online — all at once. Their website was slowed to a crawl late on the last day under heavy load conditions, and the deadline to purchase had to be extended by several days to prevent the games from being stuck with unsold tickets.

Or consider one of the most famous — when Dr Pepper promised that if Guns n’ Roses released the “Chinese Democracy” album in 2008, they would give everyone in the US a free Dr. Pepper. There are 300 million people in the United States. You can imagine how that played out.

Dr. Pepper’s website was flooded with customers, who had 24 hours to fill out a voucher for a free soda on Nov. 23, the day “Chinese Democracy” was released. When Dr. Pepper’s site failed to respond to the traffic, the company extended the offer for another day.

Guns N’ Roses lawyer Alan Gutman summed it up in an L.A. Times blog post, “What happened on November 23 was a complete fiasco. In what could only be characterized as reckless indifference or complete stupidity, Dr. Pepper was completely unprepared for the traffic to its site. Most visitors were greeted with error messages. Some people who got through to Dr. Pepper’s servers were told to call a toll free number, few of whom got through. Many walked away angry at Dr. Pepper and soured on release of ‘Chinese Democracy’.”

Conversions, satisfaction and customer retention

In a widely quoted study of e-commerce sites, the Aberdeen Group found that a one-second delay in page response time results in a 7% reduction in online customer conversion, 11% fewer page views and a 16% decrease in customer satisfaction. In monetary terms, that means if your site earns somewhere around $100,000 each day, you could be losing nearly $7 million annually simply because of this modest lag in performance.

While this may seem that users are just being persnickety, recent science shows that they can’t help it. Last year, Glasgow Caledonian University strapped EEG caps on users to monitor their brain waves and aimed EOG devices at them to track eye and facial movements while they shopped online.

The participants were separated into two groups. One had a constantly throttled 5 Mbps connection; the other a paltry 2 Mpbs connection. Users with the slower connection had to concentrate 50% more to complete the same tasks and showed significantly greater agitation and stress.

In media and e-commerce companies, the impact of performance on revenue can be huge. In the SaaS industry, it can be even bigger. Worldwide SaaS revenue is forecast to reach $14.5 billion in 2012 – up 17.9% from 2011 – with long-term projections topping $22 billion by 2015.

In the highly competitive SaaS world, the importance of lowering your Customer Acquisition Cost is second only to customer retention. Profitable SaaS companies don’t begin to make money on a new subscriber until month 18 on average. Driving down that number and keeping customers longer is therefore key.  And both are directly dependent on web performance.

Internal testing of client SaaS applications accelerated by site optimization company Strangeloop shows that a 50% increase in response time boosts conversions rates by 5% and page views by 26% while decreasing bounce rates by up to 14%.

Because SaaS customers have very little invested in their chosen solution, switching costs are low. Increasingly, the battle for a competitive advantage among SaaS providers is being fought and won over better performance.

The Global Performance Gap

Web performance is a matter of individual perception, and that perception depends entirely on where you sit.

For all the performance challenges experienced in the U.S., they are likely exacerbated overseas.

The global average connection speed in the first quarter of 2013 was 3.1 Mbps. That’s less than half of the 8.6 Mbps enjoyed in the U.S. and it includes developed countries like Italy, Spain and Portugal. In fact, the U.S. is faster on average than every country except for Sweden, the Netherlands, Switzerland and South Korea.

This is “on average” because there is wide disparity within the U.S. If you’re online in Virginia or Vermont, for example, your connection speed is likely twice that of someone in Arkansas or Alaska.  But connection speeds are getting faster, right?

Yes, but not at the same rate everywhere. In parts of Europe and Asia, the gap is widening.  And while connection speeds are getting faster, Web pages are getting bigger. The average homepage today contains 98 page objects, a 13% increase from last year.

Let’s shift focus to business continuity from the standpoint of security.

We’ve talked about performance monitoring’s impact on page-views, conversions, customer satisfaction, retention and revenue in general. What about when the game is all or nothing?

Just over a month ago, the Syrian Electronic Army hacked the DNS settings for both the New York Times and Twitter.  The hacks were virtually identical — phishing expeditions led to hacked email accounts where server passwords were exposed — but the results were very different.

Both hacks happened on a Tuesday. Twitter quickly restored service, but by Thursday afternoon, Times users will still having trouble accessing the site. According to InformationWeek, the difference was that “Twitter actively monitors its DNS settings and thus learned of the hack very quickly.”

By now, we hope you see the kernel of your own business case for Web Performance Monitoring. As mentioned, this post is the second in a three-part series on the subject, so be sure to stop by for our next post where we’ll talk about how WPM mitigates performance risks and helps direct future optimization.


  1. barryherne says:

    Thank you for the substantial article with lots of grea ideas. But it is not a problem at all if you rely on good monitoring software which you install in the office. We have the software Anturis and we feel more protected than before.

Speak Your Mind