Ludovic came back from six months of travel with €50,000 in recurring contracts signed. His product didn’t exist yet: he was selling on mock-ups. On the third episode of La Saturation, Célien’s podcast (in French), he opened up his whole acquisition machine, including the mistake that cost him the most: a badly defined ideal customer profile. If you prospect in B2B and your best campaigns don’t bring your best clients, this page is for you.
The starting point: zero budget, mock-ups and a mail merge
- He quits his job, goes travelling, and prospects from abroad with the cheapest technique there is: a Word mail merge. No sending tool, zero euros of stack.
Four lists of 300 contacts, enriched by hand. Around fifteen meetings. Two clients convinced on mock-ups. €50,000 in recurring revenue signed before the first line of code.
Only then does he buy tools: a database for his lists and a sending tool, about €2,000 a year, paid back straight away.
Remember the order: clients first, tools second. Plenty of founders do the opposite. They build a full stack, then find out three months later that their offer doesn’t get replies.
The ICP mistake: the segment that replies isn’t the one that pays
This is the most useful lesson of the episode.
His best open, reply and meeting rates came from one profile: innovation directors. On the dashboard, they looked like the perfect target.
Except those buyers signed 2 to 3 times lower, and didn’t use the product.
His real ideal customer profile was the mature market: people already equipped, already ready to pay, with time-consuming tasks to automate. Less responsive to prospecting, far more profitable once signed.
The lesson fits in one line: judge your campaigns on revenue signed, never on open rate. A campaign that gets lots of replies from a segment that doesn’t buy is worth less than a modest campaign on a segment that pays.
How to check your ideal customer profile is the right one
Three checks to run on your own numbers:
- Rank your clients by revenue signed, not by how fast they signed. The top of that list is your ICP. Not the profile that answered quickest.
- Look at who actually uses what you sell. A client who doesn’t use your product doesn’t renew and doesn’t refer.
- Compare segments on revenue signed per contact. A segment that replies half as often but signs three times higher is your best segment.
Precision beats volume, as we explain in what 15,000 LinkedIn messages taught us. But precision only pays if you’re aiming at the right target. And if a segment’s numbers look weak, run them through the LTV/CAC ratio before you scale anything.
An AI-run machine: around ten meetings a week
Today, his prospecting runs on an in-house machine driven by AI. It:
- reads results campaign by campaign;
- picks the targets;
- writes messages that are genuinely personalised;
- launches short campaigns of 50 to 70 contacts every day.
The result: around ten meetings a week for three partners, and one hour of maintenance at the weekend.
Asked for the secret of AI messages that still read as human, his answer had no magic in it: good prompts, volume, and no gimmicks.
What we take from it as an agency: AI is excellent for the admin and for short, targeted campaigns. But what follows a signal (the prospect who replies, clicks, raises a hand) deserves a human who calls back fast. That’s the principle behind the Saturation method: open cold, convert warm. And it’s why we don’t believe in fully automated prospecting.
The perfect product is a trap
He signed on mock-ups, and he still sees founders spend a year polishing their website.
His advice: talk to your buyers at scale, by email, on LinkedIn, on the phone, until you understand what they need. Products have never been faster to build. Market understanding can’t be bought.
That holds for a software company as much as for an agency or a consultancy. Prospecting isn’t only a sales channel. It’s the best market research you’ll ever run.
If he started from zero: the phone
Célien did the exercise in a video (in French): take everything away, the agency, the clients, the audience, the lists. He’s left with a laptop, a phone and 30 days to book meetings again. His plan: one target, an offer that gives something, a list of 200 companies, and 200 cold calls before the first email.
At the end of the episode, without comparing notes, Ludovic gave the same answer: if he started from zero, he’d pick up the phone. The feedback is instant and brutal, and it’s the best field test there is for an ICP and an offer. Cold calling is still the fastest channel to learn from.
Three questions to finish:
- Your campaigns: do you judge them on open rate, or on revenue signed?
- Your ICP: is it the profile that replies best, or the one that pays best?
- If everything were taken from you tomorrow, could you book meetings again within 30 days?
The full episode, in French, is on YouTube. If you want your ICP and your campaigns set on revenue rather than vanity stats, book your call.