> ## Content Index > Fetch the complete content index at: https://www.datadab.com/blog/llms.txt > Use this file to discover other available public pages before exploring further. # A quarter of your logo wall is invisible to the readers who count it most - URL: https://www.datadab.com/blog/a-quarter-of-your-logo-wall-is-invisible/ - Published: 2026-11-05T04:30:00.000Z - Updated: 2026-11-05T04:30:00.000Z - Description: Your customer logos are unreadable to about one mark in four, and the fix is a filename. What a census of 479 homepages found. - Author: Amit Ashwini - Tags: SEO *21.1% of B2B homepages show customer logos. About one mark in four is unreadable to an automated reader, and the fix costs a filename change.* Logo walls are the easiest proof a marketing team can ship, which is exactly why they are so often decorative. We coded every homepage in [the logo-wall census](https://www.datadab.com/research/b2b-saas-logo-wall-census): 479 B2B software companies, 101 of them (21.1%, 95% CI 17.7-25.0%) showing a customer wall. Then we asked a narrower question about the marks themselves. Could an automated reader name them? 74.2% carried an identifiable name in alt text, a filename or a link. The remainder, roughly one mark in four, were hashed asset filenames, light and dark variants, or bare marks with nothing to read. So a wall is not one asset. It is a wall, and roughly one mark in four is doing nothing for anyone who cannot see it. ![](https://www.datadab.com/blog/content/images/2026/10/a-quarter-of-your-logo-wall-is-invisible-01-hero-1.png) ![](https://www.datadab.com/blog/content/images/2026/10/a-quarter-of-your-logo-wall-is-invisible-02-the-median-wall-is-five-logos-an-1.png) ## The median wall is five logos, and that is the right size Five recognizable logos beat fifteen obscure ones. The typical wall carries five, buyers skim for names they already know, and a wall of fifteen small marks reads as an inventory rather than as evidence. The fragmentation confirms that collecting logos is not a winning move. No customer appears on more than 4% of walls. Amazon shows up on four, Clay, Netflix and Google on three each, and that is the top of a distribution with no top. If your best customer is not a household name, saying so with a story works. Putting them in a wall the buyer cannot recognise does not. ![](https://www.datadab.com/blog/content/images/2026/10/a-quarter-of-your-logo-wall-is-invisible-03-company-size-does-not-move-a-wal-1.png) ## Company size does not move a wall, and that is the tell Here is the comparison that makes the whole finding interesting. Story hubs climb steeply with headcount, from 33.3% at 1-10 employees to 90.2% at 200+. Comparison pages climb too, 24% to 51%. Logo walls ignore size entirely: 22.9% at 1-10 against 30.8% at 200+. A wall is a day-one asset, not a scale programme, and teams treat it as though it grows into something. It does not. It gets prettier. The practical consequence is that a wall is the weakest of the three proof assets at exactly the moment a mid-market buyer needs evidence, which is when they are comparing you to two other vendors. ![](https://www.datadab.com/blog/content/images/2026/10/a-quarter-of-your-logo-wall-is-invisible-04-what-does-move-a-wall-is-your-bi-1.png) ## What does move a wall is your billing model Flat-billing vendors lead at 37.6%. And among pages that show numbers but no tier structure, walls are at zero. That pattern points at the self-serve motion. When the page has to sell without a salesperson, proof goes on the page. When a human closes the deal, the proof moves into the conversation and the homepage stays generic. Which means the honest question is not "should we have a wall". It is "does my page sell itself, or does my team do the selling". ![](https://www.datadab.com/blog/content/images/2026/10/a-quarter-of-your-logo-wall-is-invisible-05-why-an-unreadable-mark-is-a-mark-1.png) ## Why an unreadable mark is a marketing bug, not an accessibility one The obvious framing is accessibility, and it is true: a screen reader gets nothing from a filename like \`logo-dark-03.svg\`. The less obvious framing is that automated readers are now a real audience for that strip. Answer engines and agent crawlers parse what a page states, in text. A mark that names no company states nothing. Both audiences win from the same cheap fix, which is why I keep sending it to teams instead of arguing about wall size: \- name the asset file after the customer, not the variant - put the customer name in alt text, and the fact if you have one - do not ship a wall of four versions of the same SVG with no labels - if a logo cannot be named, replace it with a sentence about that customer Four companies in our sample show investor logos where customer logos would normally go. That is a legitimate choice for a early-stage company and a confusing one for a buyer trying to judge fit, so label it as backers if you do it. ![](https://www.datadab.com/blog/content/images/2026/10/a-quarter-of-your-logo-wall-is-invisible-06-what-a-wall-cannot-do-for-you-1.png) ## What a wall cannot do for you It cannot carry a claim. A logo says "this company paid us". It does not say what broke, what changed, over what period, or whether the buyer was your size. That is why the same census points at story hubs for mid-market customers. If your customers are recognisable, the wall argues for you. If they are not, the wall is a decoration and the story is the argument. And it cannot substitute for the security page. In the same research programme, compliance proof on vendor sites outran review badges by more than three to one. A buyer who is close to signing is checking whether you can pass procurement, not whether your customer list looks like a conference sponsor wall. ![](https://www.datadab.com/blog/content/images/2026/10/a-quarter-of-your-logo-wall-is-invisible-07-small-walls-dominate-and-the-exc-1.png) ## Small walls dominate, and the exception tests the design Here is the count distribution across the 101 walls: 27 walls show 1-3 logos (26.7%), 35 show 4-6 (34.7%), 31 show 7-12 (30.7%), and 8 show 13 or more (7.9%). The great majority stay at 12 or fewer, and the median is five. One outlier shows 54 marks, which is a customer directory rather than a logo strip, and that distinction matters for anyone tempted to go long: past a dozen names you are building an index, not making an argument. ![](https://www.datadab.com/blog/content/images/2026/10/a-quarter-of-your-logo-wall-is-invisible-08-walls-hubs-and-attacks-are-one-s-1.png) ## Walls, hubs and attacks are one stack The three proof surfaces in this research programme stack by effort, and the cohort starts at the cheap end. Walls are cheap and flat across company sizes. Story hubs cost writing and climb with headcount. Comparison pages cost maintenance and climb the steepest. Of the 101 companies with a wall, 74 also publish comparison pages and 50 also run story hubs. That overlap is the finding a planner should care about: the vendors doing all three are running competitive content as a programme rather than a page, and the wall is where nearly all of them started. So it is a fine first asset and a poor only asset. ![](https://www.datadab.com/blog/content/images/2026/10/a-quarter-of-your-logo-wall-is-invisible-09-walls-open-stories-close-1.png) ## Walls open, stories close The sales reading of the same data is narrower than the design reading. Walls are flat across company sizes because they do awareness work: they tell a stranger that real companies pay you. They do not carry an enterprise deal. When the cohort invests in proof that scales with size, it goes into the story hub, which runs from 33% at 1-10 employees to 90% at 200+. Point buyers there rather than at the strip, and treat the wall as the door rather than the argument. ![](https://www.datadab.com/blog/content/images/2026/10/a-quarter-of-your-logo-wall-is-invisible-10-how-we-counted-a-wall-and-where--1.png) ## How we counted a wall, and where that has edges Method matters here because the number looks precise. All 479 homepages were fetched in October 2026 and 364 of them loaded. Walls were detected from trusted-by, customer and logo-wall patterns in server-rendered HTML, then validated per section by requiring at least two logo images. Investor strips were separated and never counted. That gives a floor, not a census. A wall rendered only by JavaScript after the page loads can be missed, and a logo carried as an inline SVG with no filename, no alt text and no link is exactly the mark our readability rule flags. We publish counts, never trademarks. No logo was republished in the study, and the per-wall data is reusable with the identifiability caveat attached, because a quarter of the marks are machine-unreadable by our own rule. If you quote these numbers, quote the caveat with them. ![](https://www.datadab.com/blog/content/images/2026/10/a-quarter-of-your-logo-wall-is-invisible-11-the-twenty-minute-audit-1.png) ## The twenty-minute audit Do this before you redesign anything, because it is cheaper than a debate about wall size. Save the homepage HTML and count the marks in the wall section. For each one, check whether a name appears in the alt text, the filename or the link. Any mark that fails all three is doing nothing for a crawler, a screen reader, or a buyer who tabs away and searches the logo later. Then check the ordering. The point of a wall is recognition, so the names your buyer knows go first, and a wall of fifteen unknown marks is an argument for the opposite of what you intended. Write down what an outsider would learn from the wall alone, in one sentence. If the sentence needs a salesperson to finish it, the wall is not doing the job you are paying it to do. Finally, decide what the wall is supposed to be proving. If the answer is "these companies are like you", the story hub is the asset. If the answer is "these brands trust us", the wall is fine, as long as it can be read. ![](https://www.datadab.com/blog/content/images/2026/10/a-quarter-of-your-logo-wall-is-invisible-12-who-this-is-for-1.png) ## Who this is for \* *Founders with a wall of six logos:*\* check whether each one can be read as text. Roughly one in four of the marks we coded could not, and that is a twenty-minute fix. \* *Designers:*\* five marks, named, with an optional number under each. Depth is not the goal, and the median is five for a reason. \* *Marketing leads planning proof assets:*\* the wall is the cheapest of three assets and the weakest at scale. Put the next week into the story hub or the comparison page, which both climb with company size for a reason. \* *Buyers:*\* if the wall is the only proof on the page, that is information about the page, not about the product. ![](https://www.datadab.com/blog/content/images/2026/10/a-quarter-of-your-logo-wall-is-invisible-13-a-note-on-how-confident-we-are-1.png) ## A note on how confident we are The sample is 479 YC-backed B2B software companies, so it skews developer and AI-native and venture-backed. Legacy mid-market software probably shows more walls, and I would not extrapolate the 21.1% to all B2B SaaS. The readability test is our own rule: an identifiable name in alt text, filename or link. A mark with a filename like \`acme-logo.svg\` counts as readable and a bare SVG path does not, which is a defensible line and still a line. We separate investor-logo strips rather than counting them, and only four companies in the sample did that, so treat the small counts as anecdote. Wall presence was read from the homepage only. A company with logos deeper in the site, on a case-study index for instance, counts as no wall here. ![](https://www.datadab.com/blog/content/images/2026/10/a-quarter-of-your-logo-wall-is-invisible-14-wrap-up-1.png) ## Wrap-up A fifth of homepages carry a wall, the median wall is five logos, no customer appears on more than 4% of them, and a quarter of the marks are unreadable to anything that has to read. Name the files and write the alt text. Then check whether the rest of your page is readable too, because a logo strip is rarely the only thing a crawler cannot parse. [Run the audit on the page a buyer opens first](https://www.datadab.com/tools/agent-audit). Coded from 479 YC-backed B2B software homepages, 101 of which show a customer wall. Readable means an identifiable name in alt text, filename or link, measured on the homepage only. [Check whether a reader can parse your page](https://www.datadab.com/tools/agent-audit)