Every referral passes five checks before it reaches you. We publish all five, and the billing rule that follows. Ask a leads site for theirs.
One partner per area in trade and local-service markets. In US legal, one buyer per call, never resold, under the state bar rules that govern the receiving firm.
Three things decide whether a bought enquiry is worth the money, and almost none of it is public. None of the five biggest UK platforms publishes a full price list, and Rated People and TrustATrader publish nothing at all. Opaque supply keeps you comparing star ratings instead of cost per job won.
Not the average. This one. Checkatrade’s model shares a single enquiry with 10–20+ trades, and SolarReviews’ own page states its leads are sold to 2.3 companies on average, with no returns and 93–97% that will not close.
Both of those are honest disclosures, and neither tells you what happened to the name on your invoice this morning. In fairness to Checkatrade: it holds 4.6/5 on Trustpilot from 75,000+ reviews, mostly homeowners rather than trade members.
We have not found a lead vendor, directory or agency that publishes its qualification criteria in full. So the buyer guesses, and the seller decides after the fact what counted as a good lead.
Ours are the five below, in fixed words. They do not change per campaign and they do not change when a month goes badly.
This is the one that costs money, and it is the one nobody writes down. Bark sells credits at £1.80 that expire after 3 months, which is a billing rule, just not one about whether the lead was any good.
Ours is mechanical. Fail a check, not sent, not billed. If one gets through anyway, you flag it, we credit it, and we trace which check let it through.
Five checks, in this order, on every referral. Each one is a stop condition: fail it and the referral is not sent, so it never reaches your invoice.
Before anything else happens, we reach the person. The number rings, a human answers, and they know why we’re texting. No form-fill ghosts, no dead lines, no “enquiries” that were only ever a mistyped phone number.
If we can’t reach them, you never could have either, so it stops with us.
They describe the job over SMS, themselves. Not a category picked from a dropdown, not our summary of a form, their own sentences, forwarded to you as they wrote them.
You know exactly what they asked for before you ring, and they know a call is coming. That’s the difference between a referral and a name on a list.
SMS · confirmed in their own words
“Yes, back gutters are leaking over the door, want it sorted in the next 2 weeks. Not had anyone else out yet.”
They tell us when they want the work done. “Next two weeks” is a job; “someday, maybe” is a browsing session, and we bin it.
A leads site counts the someday crowd because volume is the product. Ours is the opposite: fewer names, each one with a date attached.
We ask the question outright: has the work already been done, or a contract signed? A closed job is binned, never sent, never billed.
A customer still comparing quotes is a live job, and you will know exactly what they told us when you ring.
You get the referral. Nobody else does, not a second partner, not a “backup”, not the same job resold next week. One area, one partner, one handover.
Ring them knowing no competitor has their number. They’re expecting your call, and only yours.
Scope: one partner per area applies in trade and local-service markets. In US legal we sell exclusive delivery per call, one buyer per call, never resold, with no unconditional one-firm-per-area promise, because state bar advertising rules govern how a firm takes work in its own jurisdiction.
Four things, on one page. What gets checked before a lead is sent, what fails, what happens to the bill when it fails, and how many buyers receive the same enquiry.
Every test a lead has to pass before it is sent, in fixed words that do not change per campaign.
What gets binned and why. A standard with no stated failure conditions is a slogan.
What happens to the invoice when a lead fails. Credit, replacement, or nothing at all.
How many buyers receive the same enquiry. Per enquiry, not as an average across a year.
The format on this page is free to copy. We are not asking anyone to match our checks, only to write theirs down. Any lead vendor, directory or agency is invited. Publish yours and we link to it from this page, competitor or not. We have not seen one published yet.
The same rule applies to what we publish about results. In the Gutter Plus case study our system record is set out separately from what the client reports, so you can see which number came from a database and which came from a person.
One flag, one credit, one trace. If a referral reaches you having failed any of the five, you flag it, we credit it, and we tell you which check let it through.
Billing is driven by the check result. A referral that fails is stopped before it is sent, so in the ordinary case it never becomes a line on your bill. The rule is enforced by the system that raises the invoice, not by how a phone call goes.
When one gets through, you do not have to argue it. Name the check it failed. We credit it and trace the failure back, which is the only way a published standard stays honest instead of becoming a page nobody reads.
The five checks answer most of it. Pass all five and a referral is sent and billed. Fail one and it is neither.
That rule is published here in full: the five checks, and what happens to the bill when one fails. We have yet to find a lead vendor who publishes theirs.