Portent: 1-second sites convert 2.5x higher than 5-second sites

Portent, a digital agency, went back to a question it has revisited across several years of client data: does page load time actually move conversion rate, or is the relationship mostly folklore repeated from one blog post to the next. Its April 2022 analysis pulled from 20 of its own client sites and found conversion rate dropping steadily as load time rose, in both directions of its dataset.

What did the aggregate data show#

Portent split its sample into 6 B2C e-commerce sites and 14 B2B lead-generation sites, then bucketed 5.6 million sessions by load time in one-second increments. For e-commerce, conversion rate fell from 3.05% at a 1-second load to 1.68% at 2 seconds, 1.12% at 3 seconds, and 0.67% at 5 seconds, a 2.5x gap between the fastest and the 5-second mark. Lead-generation sites moved from roughly 40% at 1 second to about 34% at 2 seconds, 29% at 3 seconds, and roughly 13% at 5 seconds, close to a 3x drop over the same range. By 10 seconds, B2B conversion had fallen further to roughly 8%, about a fifth of its 1-second rate.

Load time	B2C conversion rate	B2B conversion rate
1 second	3.05%	~40%
2 seconds	1.68%	~34%
3 seconds	1.12%	~29%
5 seconds	0.67%	~13%
10 seconds	not separately reported	~8%

Portent’s own write-up is direct about the shape of the curve: “a site that loads in 1 second has a conversion rate 3x higher than a site that loads in 5 seconds” for B2B, and “2.5x higher than a site that loads in 5 seconds” for B2C e-commerce.

Why B2C and B2B sit so far apart#

The two conversion curves fall at similar rates proportionally, but their absolute levels are not close. Portent attributes that gap to what counts as a conversion in each case: “goal conversions on websites are generally achieved at a higher rate than e-commerce conversions,” since a form fill is a lower-effort action than a completed purchase. That framing matters for anyone comparing this report against a different one. A retailer benchmarking against the roughly 40% B2B figure would be comparing against the wrong curve entirely; the e-commerce numbers, topping out under 4% even at a 1-second load, are the relevant comparison.

Portent also flags a limit in its own dataset rather than smoothing over it: the 10-second bucket for B2C carries a smaller sample than the faster buckets, because fewer of the 27,000 landing pages in the sample were that slow to begin with. The agency reports the B2B 10-second figure but does not publish a parallel B2C number at that load time, and this piece follows that same boundary rather than filling the gap.

Why we care#

Conversion sensitivity to speed is not a one-off finding. Portent’s aggregate numbers track the same direction reported elsewhere, including a separate 2019 Deloitte study of 37 brand sites that found conversion lifting when mobile speed improved. Two datasets built from different methodologies pointing the same way is a stronger basis for a speed budget than either alone.

Site type changes what “good” looks like. A team running an e-commerce checkout should not benchmark its conversion rate against a lead-generation site’s, and should not expect the same percentage-point stakes. The multiplier, not the absolute rate, is the number worth carrying into a business case for a performance budget.