The 557 Work-Winning Gates: How UK Construction Work Is Actually Bought

Most UK construction work never appears on a tender portal. We hold a verified directory of 557 separate work-winning gates in the UK, and only a minority of them are buyers legally obliged to publish a notice. The rest fill work from lists they built before the job existed. Finding work is a directory problem, not a marketing problem.

I run a subcontracting model - win the work, sub the delivery, keep the spread - and the single biggest thing I have learned in doing it is that the constraint is almost never the ability to do the job. It is knowing where the job is bought. This post is the methodology behind that claim: what we counted, how we verified it, what the shape of the count tells you, and where it stops being true.

The thesis: work-finding is a directory problem#

Ask a contractor how to get more work and you will get marketing answers. Get a better website. Post on Instagram. Ask for referrals. Run some ads.

None of that is wrong, and all of it is downstream of a question nobody asks first: who buys the work you do, and what is the specific door they buy it through?

Almost every firm I have met can name three or four. A local authority they once bid to. A main contractor whose site they were on. A lead app. Then the list runs out, and the assumption is that the list itself is short - that there simply are not many places to find construction work, so the game must be about competing harder at the few you know.

The count says otherwise. There are hundreds of doors. Most firms are standing in front of four of them, wondering why the queue is so long.

The queue is long because everyone can see the door. The doors nobody can see have no queue at all.

What counts as a gate#

A count is only worth something if the unit is defined tightly, so here is the definition we work to.

A work-winning gate is a specific, reachable entrance through which a UK construction firm can be put in a position to be given work. Not a company that might need work done. Not an article about a company. An entrance: a supplier registration form, a portal signup, a framework expression of interest, a supply chain application, a network join request, an accreditation that other buyers filter on.

Three consequences fall out of that definition, and they matter more than the number.

A buyer with no route in is not a gate. During the last full audit we fetched 212 entries in detail and found 35 where the held page was reachable, the organisation genuinely buys construction services, and no public intake route existed anywhere on their site or sitemap. Large FM providers, several managing agents, a national housebuilder. We logged them honestly as cannot-determine rather than padding the count with a contact-us page.

One organisation can be several gates, or none. A housing association that runs its own portal, sits on two frameworks and buys reactive repairs through a managed network is reachable four ways with four different registration motions. Meanwhile a group of buyers who all procure through one shared platform is, from a contractor's point of view, a single registration.

A marketing page is not an entrance. This is the difference between a directory that looks useful and one that is. Anyone can list a hundred company homepages.

How the list is built and verified#

The number is the least interesting part. The verification is the asset.

Every entry holds a URL, and every URL is checked. At the last full status sweep of the 557 entries, 496 resolved cleanly to an automated fetch, 54 required a real browser to confirm because the page sits behind a bot gate, 6 could not be reached and are flagged for a hand check, and one holds no URL by design. Nothing in the set is a known-dead link.

Bot-blocked is not the same as broken, and we do not treat it as either. 33 entries in the audit returned a 403 or no response to an automated fetch. Several of those are large, obviously live organisations whose held URL is already the correct supplier registration page - the block is Cloudflare doing its job, not evidence of a bad link. Marking them dead would have been wrong. Marking them verified would also have been wrong. They carry a browser-check flag and get confirmed by eye.

We hunt for the deepest honest entrance, not the homepage. The same audit upgraded 42 entries from a page that describes a supplier relationship to the page where you actually apply. A corporate homepage became a supplier hub. A health and safety info page became the contractor onboarding portal it linked to. A portal login screen became the registration path behind it. That work is invisible in the count and it is most of the value: the difference between "this company uses subcontractors" and "here is the form".

Platforms die and companies rebrand, so the set is maintained, not published once. Lead platforms fold. Networks get acquired and fold into a parent's portal. A framework closes to new suppliers. Entries that stop resolving are re-hunted for a current route and culled if there is not one, and the research files keep a record of confirmed dead platforms so the same corpse does not get re-added by the next sweep.

That last point is the reason I distrust every "top 20 places to find construction work" listicle I have ever read. Not because the entries are wrong on the day they are written, but because nobody goes back.

The shape of the 557#

Here is where UK construction work is bought, counted as of September 2026. This is the map, not the list.

Where the work sits	Gates
Public sector - councils, NHS, universities, frameworks, consortia, devolved portals	112
Marketplaces, lead platforms, aggregators and master registrations	87
Labour routes - agencies, CIS engagement, job boards, per-trade schemes	72
Social housing landlords - housing associations, ALMOs, large private landlords	69
Private estates and infrastructure - commercial landlords, managing agents, utilities, rail, defence	61
Tier 1 and housebuilder supply chains	36
Insurance repair networks and facilities management giants	32
Accreditation schemes buyers filter on	19

That is 488. The remaining 69 came from two later sweeps - a gap-fill round into regional, retrofit and per-trade routes, and a best-entrance round that added better doors to organisations already held - and they cut across every family above.

Read the table for the asymmetry rather than the totals. The public sector column is the one everybody looks at, and it is the only one with a legal duty to advertise. Add the social housing landlords who run open portals and you have most of what a tender alert can ever show you. Everything else - the supply chains, the insurance networks, the managing agents, the private estates - buys work without publishing a word about it.

What it costs to get through them#

The other assumption worth killing is that the good doors are expensive.

Of the 557 gates, 253 have an open free signup and 119 are reached through a shared procurement platform where the registration itself is free. 73 charge - a listing fee, a membership, an accreditation. 40 are invitation only, 38 are closed to new suppliers entirely, and for 34 we could not establish the route with confidence and say so.

So the majority of the doors cost nothing but the hour it takes to fill the form. What they cost is readiness.

347 of the 557 demand at least one named accreditation or trade certification. 210 demand none. And 366 want an insurance position stated at registration. The badges buyers ask for most often, counted across the registration requirements we hold:

Demanded at registration	Gates
An SSIP scheme (any member scheme)	111
Gas Safe, where gas work is in scope	102
Constructionline	99
CHAS	64
SafeContractor	56
TrustMark	23
Achilles	14

This is the single most useful thing in the whole dataset, and it is the opposite of how most firms buy accreditations. The normal sequence is to buy the badge that a salesperson called about, then look for doors it opens. The sequence that works is to pick the twenty gates you actually want, read what those twenty demand, and buy only that. An SSIP scheme and a decent insurance position gets you further across this set than any single expensive certification, and if there is no gas work in your scope then the second most demanded item on the list is irrelevant to you.

Where the count stops being true#

I would rather state the limits than have you find them.

It is a snapshot with a decay rate. 557 is the figure as of September 2026. Some of those doors will have moved by the time you read this. The maintenance is the product; the number is just today's readout.

Coverage is deliberately uneven. Public sector and social housing are close to complete because they are legally visible. Private buyers are not, and cannot be. There is no register of every managing agent in Britain that onboards contractors. What we hold there is the large and mid-size operators we could verify, and the honest position is that the private layer is a floor, not a ceiling.

A gate is not a lead. Being on a list does not create demand. It puts you in the pool that gets called when demand appears in your patch. The compounding is real but it is slow, and it is why the firms that win at this treat registration as a standing weekly habit rather than a project.

The count says nothing about whether a buyer is worth working for. Payment terms, retention, the quality of their contract administration - none of that is in a signup form. That judgement is separate work.

The next step#

Do not try to do 557 of anything. The move is smaller than that and it is boring, which is why it works.

Pick the four families above that match what you actually deliver. Inside them, list the twenty buyers within reach of your patch. Read what those twenty demand at registration, assemble that document pack once, and then register with one of them every week until the list is done. In a quarter you will have done more for your pipeline than a year of posting.

For the mechanics of each route, start with what construction arbitrage actually is and the full demand map. Then work through the public side - which tender portal to use and how to set free alerts, the master portal registrations that cover many buyers with one account, and how to get onto a council approved contractor list, which is the same motion every private buyer on this list uses without telling anyone.

The verified directory itself, with the registration route held against each gate, lives inside Plan a Job, which is our own platform and the one I use to run the jobs once they land. The mapped version is the product. The method above is not, which is why it is here.

Last checked: 8 September 2026.