For How Long Will Consumers Forgive Mobile Slowness?
Written by Evan SchumanWith its latest batch of M-Commerce performance benchmarks being published, online performance tracking firm Gomez noticed that the average mobile response time of major retailers is 3.7 seconds. That contrasts with an average response time of 2.2 seconds for those same chains’ Web sites and an industry ideal time of two seconds.
But are those differences meaningful? Indeed, even that two-second target is suspect. To make it meaningful, wouldn’t it have to factor in the consumer’s demographics? Is a 61-year-old corporate CEO going to have the same time sensitivities as 31-year-old unemployed painter? What about a teen-ager?
The truth is that the influence of age, income and background are trivial—in terms of projecting how long that person will wait for a page to display—when compared with something much more personal: Why is that person trying to access that Web page right now? If they’re in an airport trying to find an alternative to a snowed-in flight, they have little choice but to be patient. Or a 16-year-old trying to download a new hot song. And if it’s a consumer merely browsing to kill time? Almost any delay will make that consumer flee.
I mention this to point out that the differences of a few seconds may make a lot less difference than it might initially seem. On mobile response time, for example, Amazon and Best Buy came out on top, with response times of 2 seconds and 2.16 seconds respectively. At the very bottom of that list was Target (5.75 seconds) and Musician’s Friend (5.13 seconds).
The difference between the very top and the very bottom is 3.13 seconds. If the consumer is looking for something that is only found at Musician’s Friend or Target, do you actually think that extra three seconds will actually make a difference?
The newness of smartphone data applications—and M-Commerce in general—means that retailers get a healthy amount of leeway. Few consumers expect a full Web page to load on a mobile device as quickly as on a desktop device.
Over at Gomez’s Web site stats area, the differences are much more dramatic. The fastest Web site Gomez tracked was from Nike, clocking in at 0.58 seconds. Compare that with the slowest response in that area: Newegg at 5.76 seconds. That makes Nike about ten times faster, which most consumers would likely notice.
-Marc
January 10th, 2010 at 10:30 pm
The apparent (but incomplete. read on…) message is clear: capacity planning is very inconsistent across the industry. Performance under load is predicated on the site’s architecture and infrastructure. While money (and expertise) can address infrastructure bottlenecks, only foresight can produce a solid architecture that will help a system scale and distribute its load. Even the “cloud” is not a magic bullet. Cloud instances are virtualized server instances, nothing more, you still have to load-balance them.
But we can’t only lay the blame on mobile sites. Mobile devices are much more limited than modern computers both in terms of hardware and available bandwidth. Take SSL/HTTPS for example, the trusted “handshake” that takes place between the client and server is barely noticeable on a computer running a web browser. On the other hand, this operation (with requires several round trips) can take 3 seconds or more on the latest blackberry running on a 3G network.
So yes mobile sites could be faster but let’s remember that the device and the network themselves are bottlenecks.