Finding UK companies in Clay: why Companies House data isn't enough
If you've run UK prospecting in Clay, you already know the quiet problem.
You build a beautiful enrichment workflow, point it at a UK ICP, and the table comes back lighter than it should. Half the companies you know exist don't show up. The ones that do are the obvious names — the FTSE 250s, the venture-funded SaaS, the agencies with slick LinkedIn pages. The long tail you actually wanted is missing.
This isn't Clay's fault — and it's not a bug you can fix by tweaking your search filters. It's a coverage problem with the underlying data. And if you're selling into the UK SME market, it might be the single biggest constraint on your pipeline.
This piece is for Clay users who want to fix it. We'll cover four things: where Clay's UK coverage gap actually comes from, why the matches you do get often aren't what you searched for, why falling back to the Companies House API isn't the answer most teams hope it is, and what a complete UK enrichment layer looks like — and why we built Firmbase to be one.
Clay's UK coverage gap is real (and bigger than you think)
Quick test. Run a Clay search for "small UK hotels, 1–49 employees". You'll get 2,265 results.
Run the same search in Firmbase: 10,000+ results.
That's roughly 78% of small UK hotels — the independents, the boutiques, the regional chains, the family-run guesthouses that haven't built a Series-A-friendly LinkedIn page — invisible to your Clay workflow.
It's not just hotels. We re-ran a few common UK B2B searches across both platforms:
The veterinary practices row is the one that should jolt you. There are roughly 5,500 independent vet practices in the UK. Clay can see 282 of them. If your ICP is veterinary clinics — or any comparable UK SME vertical — your prospecting pipeline is starting at around 5% of your total addressable market, and you probably didn't know it.
The B2B SaaS row is worth pausing on for the opposite reason. SaaS is Clay's home turf, exactly the kind of company that Apollo and ZoomInfo categorise well. And even there, Firmbase still finds 56% more. The pattern holds across categories — it's just less extreme when the underlying enrichment is US-shaped to begin with.
Why does this happen? Because Clay's enrichment layer is built on partner data sources — Apollo, ZoomInfo, LinkedIn-derived providers, web-crawl databases — that are optimised for the US and skew heavily toward companies with strong digital footprints. They love venture-funded SaaS. They love companies that ship press releases. They struggle with owner-managed UK SMEs that don't buy programmatic ads or hire growth marketers.
The UK has 5 million+ active companies registered at Companies House. A typical commercial enrichment dataset captures a small fraction of them.
If your ICP is "UK businesses that look like a US tech company", Clay's native enrichment is brilliant. If your ICP is anything else — and most UK B2B sellers' ICPs are — you've got a hole.
And the matches you do get aren't always what you asked for
Coverage is only half the problem. The other half is precision — the companies that do show up in your Clay table often aren't what you searched for in the first place.
Run a Clay search for "UK boutique hotels, 1–49 employees". Among the results, you'll find:
- Oyo UK — not a boutique hotel. Oyo is a SoftBank-backed hotel franchise platform with thousands of properties globally. It surfaces in the results because its profile mentions "hotel" and its UK headcount sits inside the employee filter. Pitch boutique-hotel software to Oyo and you'll get a polite "wrong number" inside fifteen seconds.
- Shackle — also not a hotel. Shackle is a hotel software company that bills itself as "the greatest hotel app on the planet". It's in the results because the underlying enrichment dataset categorises it under "hotels" rather than "hotel software". One of these is a potential buyer of your product. The other is a competitor.
Both are wrong matches for the same underlying reason: Clay's enrichment layer infers industry from LinkedIn descriptions, web-crawl signals, and self-reported categories. Anything with "hotel" prominently in its profile gets flagged as a hotel — franchise platforms, SaaS vendors, marketing agencies, insurance brokers, the lot. Employee-count filters help a bit, but Apollo and ZoomInfo headcounts are estimates that go stale fast and don't always reflect actual operating headcount, especially for franchise structures and UK subsidiaries.
The result: your "3,000 boutique hotels" Clay search is actually a couple of thousand real hotels plus a long tail of companies that have nothing to do with running one. You find the false positives the slow way — an SDR sends an email, gets a confused reply, and the row gets marked as junk. Multiply that across a full campaign and the wasted-effort cost is real.
So the actual usable pool of UK boutique hotels in your Clay workflow is even smaller than the headline coverage gap suggests. Half-empty and partially mislabelled.
"We'll just use the Companies House API, then"
Most teams notice this gap and reach for the obvious fix: build a Clay HTTP column that calls the Companies House API. Companies House is free, public, and has every active UK company on file. Problem solved, right?
In a demo, yes. In production, no. There are four reasons it falls apart, and each one matters.
1. SIC codes are broken
The UK's industrial classification system — SIC — was last meaningfully revised in 2007. It was built for a manufacturing-heavy economy, not a services-and-software one. There's no clean SIC code for "B2B SaaS". There's no distinction between consumer ecommerce and enterprise marketplaces. There's no way to separate fintech from traditional finance. Many companies pick the most generic SIC available at incorporation and never update it. Others file three or four SIC codes hoping one fits.
The result: if you filter Companies House by SIC, you'll get a noisy, incomplete list — full of companies in adjacent industries and missing the ones that picked the "wrong" code at registration.
A taxonomy built today, from the way companies actually describe themselves on the web, doesn't have these problems. SIC tells you what someone ticked at incorporation. A modern taxonomy tells you what a company actually does.
2. Rate limits make production workflows fragile
The Companies House API caps you at 600 requests per five minutes. That's fine for ad-hoc queries. It's not fine for a Clay table running enrichment across thousands of accounts.
You'll hit 429 errors. Your workflow will break mid-run. You'll spend more time debugging request timing than prospecting. Anyone who's tried to operationalise the Companies House API at scale has the scars to prove it.
3. No financial health signal
Companies House gives you raw accounts data — balance sheets, P&Ls, charges, employee counts, dormancy flags. What it does not give you is a usable signal on whether a company is healthy, growing, or quietly running on fumes.
To extract that signal, you'd have to download accounts, parse them, normalise across reporting years, handle micro-entity exemptions, and build your own scoring model. That's a six-month engineering project, not an afternoon in Clay.
For UK B2B sellers, this matters more than it would in the US — because UK companies file accounts publicly, the signal exists. It's just locked up in PDFs and inconsistent XBRL filings. Most teams can't extract it, so they ignore it.
4. No natural-language ICP search
Companies House lets you query by company number, by name, or by SIC. That's it.
There's no "show me UK SaaS companies, 20–100 employees, headquartered in London, founded in the last five years, healthy financials" query. You'd have to build that intersection yourself, across multiple data sources, and re-run it every time you tweaked the criteria.
This is the single biggest gap. Modern outbound is defined by ICP iteration speed. If it takes you a week to test a new ICP hypothesis, you'll test maybe four a quarter. If it takes you ten seconds, you'll test forty. The teams who can iterate ICPs faster simply win more pipeline.
What a proper UK enrichment layer looks like
This is what we built Firmbase to be. Five things, each addressing one of the gaps above.
A modern taxonomy, built from web data — what companies actually say about themselves on their own sites, what their customers say, what roles they hire for, what their tech stack signals. B2B vs B2C. Vertical. Business model. Stage. Crucially, it distinguishes companies that operate in a sector from companies that serve it — a hotel operator from a hotel software platform, a law firm from a legaltech startup, a recruitment agency from a recruitment-tech vendor. Oyo doesn't show up when you ask for boutique hotels. Shackle doesn't show up when you ask for hotels at all. The kind of segmentation that maps onto how sales teams actually think — not how a 2007 government classification system thinks.
A financial health score — synthesised from Companies House filings, normalised across years, scored on a clean 1–10 scale. Lets you filter for "growing UK SaaS with healthy cash position" without parsing accounts yourself. This single feature changes the kind of ICP a UK seller can credibly target.
Natural-language ICP search. Type "UK SaaS, 20–100 employees, London-based, raised in the last 18 months, healthy financials, B2B not B2C". Get a list back. Tweak it. Push to Clay. The whole flow is seconds, not days.
No rate limits at the Firmbase layer. Run enrichment across thousands of Clay rows without watching your workflow break.
Full UK coverage. Every active company at Companies House, enriched with web, people, and funding data layered on top. That's the gap closing in the table above — 282 vet practices becomes 5,453, 2,265 small hotels becomes 10,000+, and even Clay's strongest category (B2B SaaS) gains 50%+ more accounts.
That last one is worth dwelling on for a second. There are 5 million+ active UK companies. Your Clay table currently sees a small slice of them. The slice you can't see is exactly the slice your competitors also can't see — which is exactly the slice where the easy pipeline is hiding.
What this looks like inside Clay
Here's the workflow we're building toward.
You're a UK sales team prospecting independent hotels for a hospitality SaaS product.
In Clay today: you search "small UK hotels", get 2,265 rows, enrich what's there with people and email data, push to outbound. The 7,700+ hotels that don't appear in Clay's enrichment partners simply don't exist in your pipeline. You'll never email them, because you don't know they're there.
In Clay with Firmbase: you query Firmbase in natural language — "UK independent hotels, 10–50 employees, healthy financial position, outside major chain ownership" — get a complete list back, push it into Clay, layer on Clay's existing people, web, and AI enrichment, and run outbound as normal. The hotels Clay couldn't see are now in your pipeline. Better-segmented than they would have been with a SIC filter, and pre-scored on financial health.
The output side of your Clay table is what changes most: more rows, better-segmented rows, with financial health and modern taxonomy fields you can sort and filter on. Everything downstream of that — sequencing, AI personalisation, dispatch — works exactly the way it does today.
How it's going to work
Firmbase will integrate with Clay through Clay's HTTP column when our API ships. That means any Clay user will be able to wire Firmbase into their tables in a few minutes — paste an API key, map the inputs and outputs, run enrichment. A native Clay marketplace integration is on the roadmap after that.
We're early. The Firmbase API is in private beta, and we're currently working with a small group of UK sales teams to refine it. If you want first access — and a chance to shape what gets built — the link to book a call is at the bottom of the page.
Built to work with Clay, not around it
Worth saying explicitly: we love Clay. It's the best go-to-market tool of its generation, and the team there has built something that genuinely changes how revenue teams operate.
We built Firmbase to extend Clay where it's currently thin — UK long-tail coverage and Companies House-derived enrichment — not to replace any part of it. The two are complementary by design. Firmbase is the data layer that knows the UK; Clay is the workflow layer that turns data into pipeline.
If you're prospecting the UK and your Clay tables feel half-empty, that's the gap we're here to close.
Ready to see it in action?
We're onboarding a handful of UK sales teams into Firmbase early access right now. On a 20-minute call we'll walk you through:
- The kind of UK ICP queries that aren't possible anywhere else
- A live look at the coverage difference vs Clay's native sources on your verticals
- How the Clay HTTP integration will work when the API goes live
- A path onto the early access list