He had already spent about £20,000 with an agency that sent him leads from companies which never use an outside firm. On a desk this narrow, the whole problem is knowing which companies count, so that is what we wrote down first.
Dan runs a specialist recruitment firm placing AI, cybersecurity and tech talent. He gets about ten messages a week from people promising him meetings. When we first contacted him we were one of those ten, and he says so himself.
He had already paid for that promise once. About £20,000 went to an agency that sent him low quality leads, from companies that did not need an outside firm in the first place. The money went out before anyone had agreed what a real target looked like.
A meeting with a company that never uses outside recruiters is not a slow start on his desk. It is nothing at all.
His stated worry was not lead volume. He says he was more concerned about his brand reputation and his messaging. He is the person whose name sits on whatever goes out, and he had already watched one agency spend his money without protecting either.
The second problem was where the failure had happened. The previous leads came from companies with no need for an outside firm, which means the mistake was made at the definition of a target, not in the copy. More volume against the same definition would only have produced more of the same.
The third was the partner problem. Dan says it is difficult to find a business development partner who can understand your niche and then deliver predictably, and that the previous ones burnt him out. We were starting from earned suspicion.
The £20,000 did not buy bad writing. It bought the wrong list. Plenty of companies hire AI and cybersecurity people, and almost none of the ones who hire in-house will ever pay a fee, so a list built on hiring signals alone produces meetings that cannot convert.
Reputation risk and qualification risk turned out to be the same risk. Every message sent to a company that does not use outside firms reads as untargeted to the person receiving it. His concern about brand was not separate from his concern about lead quality. It was the same failure seen from the other side.
Nobody had written down what a qualified company looked like on his desk.
AI and cybersecurity are broad labels, and a partner guessing at them guesses wrong in a way that stays invisible until the meeting starts. The knowledge sat with Dan and had never been moved into a document anyone else could work from. So the first deliverable was not outreach. It was the definition.
“I was more concerned about my brand reputation and messaging.”
0:25“If you're impatient, this might not work out. I had to wait at least 30 days for the machine to start running.”
0:57The knowledge of what a qualified company looks like sat with Dan. We put it on paper: we drafted the document from what he told us, he corrected it until it matched his desk, and it went into the agreement in his words. It names the decision maker, the live mandate and the fee level worth his time. It is the piece he singles out, and the reason is that qualification stopped being our guess about his niche and became his own standard, written down, with the right to reject anything that misses it.
His previous agency sent leads from companies that did not need an outside firm, so the target list is built to fail those companies early. Every account is checked against the written definition first, and anything that cannot show a live mandate plus a reason to go outside for it never becomes a message. Far fewer accounts survive that filter. The ones that do are the only ones his desk can convert.
Dan said he was more concerned about brand reputation and messaging than about lead count, so nothing goes out under his firm's name until he has read it. He sees the angles, the wording and the way his specialism is described, and he can change any of it. That costs a little speed at the start. It removes the risk he was actually worried about, which is the reason the previous spend hurt.
The agreement carries a monthly floor of qualified conversations, measured against Dan's definition rather than ours. If a month lands under the floor, we work the following period at no additional cost until the gap is closed. We do not guarantee placements and we do not guarantee fees, because neither is inside our control. The conversation count is, so that is the only thing we put a number on.
Dan can reject any meeting that misses the written definition, and each rejection is treated as information about the list rather than as a dispute. A rejected account tells us which part of the definition was read too loosely, and the filter changes that week. That loop is what turned an uneven first month into what he describes as totally stabilised by month three.
Dan says he had to wait at least 30 days for the machine to start running. That window is definition, list build and message approval, and it produces no meetings by design. He raises it himself as the reason an impatient owner should not do this.
Meetings were arriving before this point, but the month to month pattern was not yet something he could plan around. He describes the point where it settled as totally stabilised.
Dan says he is currently getting around 10 to 15 meetings a month. That is the level the machine settled at, measured against his own written definition, not a peak month.
Dan says that over the last five months he placed more than ten people out of these conversations. He and his team made every one of those placements. The conversations were the input.
The placement number is his work, not ours. Dan placed more than ten people over five months. We did not find those candidates, run those processes or close those fees. What we supplied was the conversation at the front, with a company that met the definition he signed off. Every step after the meeting was his firm. The five months describes the stretch he was talking about when he recorded this, not the length of the engagement.
The meetings number counts qualified conversations, not activity. The 10 to 15 a month are meetings that passed his written definition: a decision maker, a live mandate and a fee level he set. It is not a count of messages sent, replies received or calls booked. Meetings that fail the definition get rejected and do not appear in it.
Neither number is a forecast. Both describe what happened on one specialist desk, reported by the owner of that desk. The guaranteed part of our agreement is the monthly floor of qualified conversations. Everything downstream of the meeting depends on the firm, the market and the mandate.
Dan placed more than ten people. We do no candidate work of any kind, and we never guarantee placements or fees. The floor in the agreement is a number of qualified conversations, nothing further down the funnel.
He waited at least 30 days before anything ran, and describes it stabilising at month three. He raises this himself, and his own advice is that if you are impatient, this might not work out.
Dan holds rejection rights against his own written definition, and he uses them. The definition exists because meetings do miss. The count only includes the ones that passed.
One owner, one niche, his own account of his numbers. We have not audited his placement figures, we do not know what those placements were worth, and none of it is a projection for a different firm.
Dan's read is that it is difficult to find a business development partner who can actually understand your niche and then deliver predictably, and that the previous ones burnt him out. That is the problem this was built for. If you run a specialist desk and you have already paid for leads that were never going to convert, the question to ask a partner is not how many meetings they will book. It is who writes down what counts as a meeting, and what happens when they miss it.
Clientflow