Last year we wrote about how website carbon testing helps strengthen your Environmental score for B Corp renewal. The argument is that your website is a slice of Scope 3 emissions most B Corps never look at, and that measuring it gives you evidence for the B Impact Assessment that very few of your peers are bringing to the table.
That post made a claim we hadn’t actually tested: that most B Corp websites have room to improve.
So we tested it. We took a random sample of 70 UK B Corps in professional and financial services: consultancies, accountants, law firms, recruiters, investment managers, insurers — and measured the carbon footprint of their homepages.
Here’s what came back.
Two-thirds of the sites are over 1 MB
The median homepage in our sample weighed 1.11 MB. Sixty-six per cent were over 1 MB, and nearly a quarter were over 2 MB.
That matters because of what a mobile phone can actually cope with. Alex Russell’s Performance Inequality Gap analysis sets the realistic budget for a JavaScript-driven site at around 1.3 MB total if you want it interactive within five seconds measured on a mid-range Android handset (a Samsung Galaxy A51, the sort of phone most people actually own) over a median global mobile connection of 7.2 Mbps.
Set against that budget, four in ten of the sites we measured are already over it, 23 of 58. Weight isn’t an abstract environmental metric. It’s the same thing as your site feeling slow to the person trying to use it.
The spread was the striking part. The lightest site in the sample came in at 0.13 MB. The heaviest was 17.1 MB, 135 times larger, and both are UK B Corps selling professional services to broadly similar buyers. There is no version of the work these two businesses do that justifies a 135x difference in the cost of loading their front page.
For context, the median mobile page across the whole web was 2.31 MB in the HTTP Archive’s 2024 Web Almanac, roughly twice our sample’s median, and only 10 of our 58 sites exceeded it. So B Corp professional services sites are meaningfully lighter than the web at large, and that’s worth saying plainly. The point isn’t that this sector is bad. It’s that a group of businesses who have formally committed to environmental accountability are still, four times in ten, over the budget their own visitors’ phones impose.
Only 12% scored an A
We graded every site A+ to F on emissions per visit:
| Grade | Sites |
|---|---|
| A+ | 3 |
| A | 4 |
| B | 12 |
| C | 20 |
| D | 13 |
| E | 2 |
| F | 4 |
Seven sites out of 58 (12%) scored an A or better. Nineteen scored a D or worse.
The bulk sits in the middle, at C. These aren’t disasters. They’re ordinary business websites that have accumulated a few too many uncompressed images, a handful of tracking scripts nobody has audited in three years, and a font stack that loads six weights when it uses two.
The improvement is smaller than you’d think
Here’s the finding that surprised us most.
If every site in the sample simply matched the best quartile of the sample itself, 0.77 MB, a figure fifteen of these businesses already achieve total page weight across all 58 sites would fall by 60%.
Nobody would have to do anything clever. No one would need to beat a best-in-class benchmark or hire a performance specialist. Every site would just have to be as good as the better quarter of its own peer group already is.
The median site is only 1.4x heavier than that quartile. Most of the gap is closed with unglamorous work: compressing images properly, serving modern formats, removing scripts that no longer earn their place, and letting fonts subset.
There’s a concentration effect too. The six heaviest sites in the sample account for 40% of total page weight across all 58. If you’re one of those six, this is the cheapest environmental win available to you this year, and it isn’t close.
Consultancies are the heaviest; investment managers the lightest
Breaking it down by sub-sector, with the caveat that these are small groups and directional only:
| Sub-sector | Sites | Median |
|---|---|---|
| Management consultancies | 13 | 1.90 MB |
| Recruitment & HR | 5 | 1.65 MB |
| Equity investing | 4 | 1.49 MB |
| Professional, scientific & technical | 14 | 1.27 MB |
| Other financial services | 4 | 1.00 MB |
| Accounting & auditing | 4 | 0.96 MB |
| Investment advising | 4 | 0.78 MB |
The pattern is roughly what you’d expect, and it’s about design convention rather than industry virtue. Consultancies and recruiters sell on brand and culture, so their sites lean on full-bleed photography, video headers and team galleries. Accountants and investment advisers tend to build plainer, more text-led sites and land lighter almost by accident.
Which is worth sitting with. The lightest sub-sector in our sample didn’t get there through an environmental programme. They got there by not adding things.
What this means for your BCorp renewal
The practical value here is evidence. B Lab standards get more demanding every cycle, and “we looked at our digital footprint, measured it, and acted on it” is a genuinely differentiated thing to bring to an assessment because on this sample, most of your peers haven’t.
A website carbon audit gives you three things the BIA can use: a baseline figure in CO₂e, a documented set of changes, and a re-measurement showing the reduction. That maps onto Climate, and depending on your hosting decisions, it can touch Energy/Water/Materials and Suppliers too.
At the median in our sample, 0.325 g CO₂e per visit, a site taking 10,000 visits a month emits around 39 kg CO₂e a year. At the worst site we measured, the same traffic would produce 598 kg. Neither number will transform your carbon accounts on its own. But both are real, both are measurable, and unlike most of your scope 3 emissions, this one is entirely within your control and fixable in weeks rather than years.
How we measured this
We drew a random sample of 70 from 847 UK-certified B Corps listed in the B Corp Directory across 19 professional and financial services industry codes, using a fixed seed so the sample can be reproduced. Homepages were scanned on 7 August 2026 using our own website carbon testing tool.
Twelve sites were excluded and are not counted in any figure above: six blocked automated scanning, four returned responses too small to be a real page, one could not be fetched, and one failed DNS resolution. That’s a 17% exclusion rate, which is normal for automated scanning at this scale and worth stating plainly, sites behind bot protection can otherwise be scored as near-zero and quietly flatter the results.
One limitation to be clear about: page weight is measured without executing JavaScript, so sites that load assets dynamically after the initial page load will be under-counted. That makes these figures conservative. The real numbers are likely worse, not better.
This is the first wave. We intend to run it annually and to widen it to other sectors, so we can say something useful about whether the picture is improving.
Want to know where your own site sits? You can test it yourself in about thirty seconds. If you’d rather talk through what the result means for your renewal and what the fixes actually involve, book a 15-minute call.
