Wait A Minute, Akamai, With Your 2-Second Page Load Report

Akamai recently published a study saying consumers lose patience with E-Commerce sites that take more than two seconds to load. The September 14 report, based on an Akamai-commissioned survey by Forrester, received a good deal of media attention. But the conclusion—that sites that used to have to load in four seconds now need to load twice as quickly—is not only not supported in the report itself, but it doesn't make any sense given the complicated reality that is today's Web.

Nobody will argue that when it comes to page display rates, the faster the better. However, we couldn't help but wonder about the accuracy, and meaningfulness, of the report. Did anybody actually measure how many seconds passed before an E-Commerce consumer clicked away from a retailer site? It seems hard to believe that somebody interested in shopping at, say, Macys.com would flee in a huff to Kohls.com if the Macy's site didn't load before three seconds had passed.

Indeed, given the tons of causes of slight delays in page loads today that are entirely beyond the reach of the retailer's IT department—including Internet congestion, delays at the consumer's ISP and unanticipated and momentary traffic surges for that retailer's site, not to mention the consumer's machine's speed, internal congestion and whether a backup or virus-checker program chooses to do something at the moment of that visit—it's hard to argue that a two-second delay could even be detected.

Many households have several computers sharing a cable modem or DSL line. All it takes is someone playing World of Warcraft for there to be a chunk of available bandwidth gone from the equation, a squeezing of the pipe that might make that retail shopping site appear to be a slow-loader when it actually appears in the wink of an eye on the neighbor's PC. There's also the processing power, available memory, running applications and open browser tabs that could slow a site's debut.

As it turns out, there was no real testing. The findings cited in the Akamai/Forrester report were based on answers to a survey E-Mailed to 1,048 people. We’re not suggesting the survey respondents lied. But can any real conclusions be drawn from the answers to multiple-choice questions asking people how long they wait for sites to load? The respondents weren't even hand-selected; they were in actuality self-selected based on whoever chose to fill out the form.

Even Akamai isn't exactly thrilled about the use of a survey to determine E-Commerce site user patience. "We have the same concern," said Akamai Marketing Manager for E-Commerce Margaret Rivera. "People aren't going to necessarily differentiate between fractions of a second. We looked into different approaches and, unfortunately, the consumer survey was the best way to go about it. I'm definitely open to other thoughts."

She pointed out that the survey is a measurement of "consumer perception," which might differ from real-world consumer action. However, Rivera jumped to a conclusion that we might not make: "The acceptable page-load time is probably even shorter" than two seconds.

The point is that, despite what they say on surveys, we assert that people are likely more willing to give a site a break than you would believe after reading the Akamai/Forrester report. It said that 47 percent of consumers "expect a Web page to load in two seconds or less." Akamai said that, in an identical survey it and Forrester conducted in 2006, most consumers expected pages to load in four or fewer seconds.Although Forrester "found that 40 percent of consumers will wait no more than three seconds for a Web page to render before abandoning the site," consider the flip side. The report could have easily been interpreted to say most consumers are rather patient when it comes to site load times, as 60 percent of the respondents said they were willing to wait four or more seconds for a site to load before giving up.

Also, 63 percent of those queried said they tend to sit there and watch while slow sites load rather than become engaged in other tasks. That should be good news to retailers fearful that distractions and multitasking can kill impulse shopping if their site doesn’t exactly explode onto a prospective customer's screen.

Another hard-to-swallow assertion from the report is its claim that people who experience a slow-loading retailer Web site said they would, in essence, punish the retailer by not only abandoning its site but also reducing their visits to its brick-and-mortar stores. However, the report pegged the number of folks who would do that at a less-than-jaw-dropping 27 percent, meaning that a very large majority of shoppers, logically, don't shun the local Wal-Mart if Walmart.com is sluggish. Also, we have to wonder if those consumer comments aren't empty threats to retailers to keep things moving quickly. It seems unlikely that a consumer who truly wanted to buy a TV at the neighborhood Target suddenly wouldn't because of a two-second delay in opening the Target.com homepage.

For all the report's faults, it does have value in that it can be compared to the survey conducted three years ago. Even if you don't buy into the truthfulness or accuracy of the answers, it makes sense to accept that whatever issues cloud this year's report also clouded the 2006 version. If that's the case, then Akamai and Forrester are correct in saying people expect more in terms of speed from E-Commerce sites now than they did then.

Rivera said Akamai knows this firsthand, because it has "a lot of customers who are setting goals around being able to decrease the response time on their sites, especially on transaction-related pages." She noted that Urban Outfitters, the $1.7 billion retailer with 130 stores, came to Akamai "with a goal of having pages load in less than a second" largely because its customers are young and impatient. "While people are adding content and making pages heavier, shoppers want sites to be faster and faster," Rivera said. "That's a big challenge for retailers."

It is. But encouraging retail E-Commerce execs to spend money to accelerate sites beyond a consumer's likely ability to notice would also be a big challenge, in addition to being a reckless way to spend shrinking IT resources in a day when stores are closing. Wouldn't pouring those dollars into additional functionality inside the site be a better approach?