Data Illustrates the Impact of Load Times on Conversion Rates

Mobile page speeds in the US are lagging best practices, and a recent study from SOASTA [download page] (now part of Akamai) demonstrates how lengthier page loads can have a significant detrimental influence on e-commerce conversion rates. In fact, a mere 100-milliseconds slower than a peak time can potentially result in a 7.1% drop in smartphone conversion rates.

The study looked at the load times for sites with peak conversion rates:

  • 1.8 seconds for peak desktop conversion rates, which stood at 12.8%;
  • 1.9 seconds for peak tablet conversion rates, which stood at 7.2%;
  • 2.7 seconds for peak smartphone conversion rates, which stood at 3.3%.

A couple of quick takeaways from the above results:

Moving on, the SOASTA study then looked at conversion rates across these devices on sites that loaded more slowly. Specially, 100 milliseconds more slowly, 1 second more slowly and 2 seconds more slowly.

As mentioned above, even 100 milliseconds appears to have an impact, though it’s much more pronounced on smartphones than other devices.

The impact – per this analysis – of a 1-second slowdown figured to be around the 20% range across devices, while sites that loaded 2 seconds slower than the optimal load time experienced a massive 36.5% lower conversion rates on desktops. (See the chart above for the full results.)

Bounce Rates

Load times also have an impact on bounce rates, according to the study. In this case, a 100-ms delay was not associated with any significant increase in bounce rates, but other slowdowns did.

Once again, the effect was most pronounced on smartphones, where a 2 second slowdown was linked to a bounce rate twice as high as the optimal time. (In the above-referenced study, Google found that the probability of bounce on a mobile site more than doubled when page load time increased from 1 to 6 seconds – a much bigger delay. However, that analysis measured a range of sites, not just e-commerce sites.)

Head on over here to download the full study, which also contains data on start render times, session lengths, and third-party scripts.

About the Data: The results are based on an analysis of 1 month’s worth of beacon data from leading retail sites, comprised of customers who had given permission for their data to be anonymized, aggregated and used in the research. The study represents 27.7 billion beacons’ worth of data – which equates to roughly 10 billion user visits.