Introduction — The Hundred and Forty Leads That Were Contacted on a Thursday
A volume target met by loosening a filter nobody had written down.
Dara agreed a price per lead.
The pilot was straightforward. A landing page, a small paid budget, a form with five fields, and a filter that removed anybody outside the client's stated sector. Two weeks produced thirty-one leads. The client was pleased. A monthly volume was agreed at the same price per lead, and the arrangement began.
Month one delivered a hundred and forty leads.
The client accepted ninety-six. Forty-four were rejected — some for sector, some because the company was too small, and eleven because the person who filled the form was not a decision maker, which had never been part of the definition. Dara had hit the volume target, and hitting it had required loosening the filter in week three, because the qualifying pool at the original setting produced about ninety a month rather than a hundred and forty.
Of the ninety-six accepted, the average time to first contact was four days. Thirty-one of them were contacted for the first time more than a week after they were delivered. Some were never contacted at all — they sat in a spreadsheet the client's salesperson opened on Thursdays.
At the end of the month the client said the leads were poor.
Dara had no way to answer that. The speed-to-contact figure existed in the client's CRM and nobody had ever asked for it. There was no record of which leads had been contacted, when, or by whom. There was no threshold document, so the rejections could not be checked against anything. And the paid channel's costs had risen by roughly forty per cent through the month, against a price per lead that had been fixed for six months — which meant the final three weeks were delivered at a loss that only became visible when the card statement arrived.
That is the arithmetic this book exists to prevent.
The interesting thing about it, again, is that nobody behaved badly. The client genuinely believed the decision-maker requirement was obvious. The salesperson genuinely intended to call everybody and genuinely ran out of week. The platform did not raise costs to hurt anybody. And Dara loosened the filter for an entirely rational reason: there was a number to hit, and loosening the filter is the only lever that moves the number this month.
Every part of it was ordinary. That is why it is worth a book.
What went wrong, named precisely
The volume target was met by loosening the definition. This is the central trap of the trade. Volume can always be produced. Quality cannot be seen for weeks. So the lever that fixes this month is the lever that ruins next quarter, and it is pulled by reasonable people under mild pressure.
The threshold was never written down. Forty-four rejections, against a standard that existed in somebody's head. Eleven of them against a criterion nobody had ever stated.
The handoff was never measured. Four days average, and a third of the leads waiting more than a week. Intent decays in hours, not days — and every one of those leads was judged as a bad lead rather than as a late call.
The cost base was not the price base. A fixed price per lead, sitting on a variable cost that somebody else controls. That is not a pricing error; it is a structural one, and it is specific to this trade.
The four rules this book is built on
One — a lead is a claim you are making about a person.
Every record you deliver asserts that this person is a plausible buyer and that they may reasonably be contacted. You are making both claims, in your client's name, about somebody you have never spoken to, at a volume of hundreds a month.
Two — a lead perishes, and it perishes in the gap you do not control.
The moment you deliver, the record enters a process that is not yours. Speed to first contact is the largest single determinant of whether your work looks good, and it belongs to somebody else. Measuring it and reporting it is not defensiveness — it is the only way the conversation about quality can be had honestly.
Three — you can always produce more leads, and quality is invisible for weeks.
There is no month in which you cannot hit a number. There is only a later month in which the loosening becomes visible, by which time it will be attributed to your work rather than to the target.
Four — you are building a machine, and everything it runs on is rented.
Data sources change their terms. Platforms reprice. A tool changes its fields and a form stops capturing something. Underneath the price you fixed sits a cost base that belongs to other people, and it moves without telling you.
Where this went, over twenty-six months
Dara is a composite. The specifics below are what the pattern looks like when the four rules are taken seriously rather than discovered one at a time.
Month two: the lead definition written down as a threshold — six checkable criteria, each with a worked example either way, and an agreement that changes apply forward rather than backward.
Month four: the volume band restructured. A build fee, a monthly retainer covering the system, and a stated band of qualified leads with an agreed above-band rate — so a volume request became a commercial conversation rather than a threshold adjustment.
Month six: speed to first contact instrumented and reported. It turned out to be the single highest-return change of the whole arrangement, and it required no change to the leads at all.
Month nine: platform costs moved again. This time the agreement had a pass-through clause and a floor, and the month was still profitable.
Month twenty-six: cost per qualified lead had fallen by roughly two-thirds. The accepted-lead rate had moved from roughly 69% to roughly 94%. Average speed to first contact had moved from around four days to under two hours on routed leads.
Revenue mix at that point: retained systems 38%, profile and threshold projects 14%, list building and enrichment 18%, page and capture builds 16%, reactivation projects 14%.
None of it came from producing more leads.
What this book will not tell you
No rates. Every financial figure is blank, in U.S. dollars, and yours to fill. A price per lead without a source cost, a threshold and an accepted-lead rate attached is not information.
No law. No data-protection position, no consent rule, no lawful basis, no source-terms interpretation, no scraping position, no employment-status test, no tax treatment. All of it differs by where you are, where the client is, where the person on the record is, and where the data came from. Chapter 15 gives you the questions; a qualified professional gives you the answers.
No promise about outcomes. Not a conversion rate, not a cost per acquisition, not revenue. You build the top of a funnel and hand it to people whose work you cannot see.
How to use the prompts
There are 148 of them, four per chapter, and every one carries that chapter's constraints inside it, marked with ⚠. That is deliberate: the model works under the rules rather than around them, and each ends with you supplying your own profile, your own threshold, your own sources and your own figures.
None supplies a rate.
The one that matters most in this trade: never let a model produce, complete or infer a data field on a lead record. In lead generation the dangerous output is not a badly written page. It is a plausible-looking company size, job title, email address or phone number that was generated rather than found — delivered as fact, in a client's name, about a real person, in a batch of two hundred where nobody will check.
©2026 James Henderson / https://localhandyman.work