Niche x Zone x Audit: the only local prospecting method we keep
A narrow trade, a precise geography, an observation the prospect can verify. Three constraints that take a reply rate from 2 to nearly 10%.
Most local outbound campaigns fail before the first send. Not because of the subject line, not because of the tool. Because of fuzzy targeting: "local businesses", "SMBs around here". Fuzzy targeting produces a fuzzy message, and a fuzzy message ends up in the trash.
After dozens of campaigns, we stopped testing email variants and wrote down a method. Three constraints, in order, before writing a single line: Niche x Zone x Audit. It is the foundation of everything we publish on this blog.
- Niche: a narrow trade, not a category. "Dentist", not "healthcare". This is what makes the audit credible.
- Zone: one geography, sized to 300 to 1500 businesses. Enough to sustain a campaign, small enough to exhaust.
- Audit: an observation verifiable in 10 seconds becomes your opening line. This is the lever that moves replies from 2.1% to 9.7%.
- Tool-agnostic. The full checklist and the opener prompt are below, copyable.
The real problem: fuzzy targeting kills the message
When your target is "local businesses", you can only write a generic message. You know neither their trade, nor their constraints, nor what costs them money. So you flatter: "I love your work". The prospect has read it a thousand times.
The method inverts the problem. We do not hunt for the best email, we hunt for the best starting constraints. Once the niche and zone are fixed, the right message becomes almost obvious, because you know exactly who you are talking to.
You do not personalize an email. You pick a target so precise the message cannot help but be relevant.
The Niche x Zone x Audit method, in three parts
Order matters. The niche shapes the audit, the zone shapes the volume. Do not skip a step.
One trade, not a category. "Dentist", "web agency", "HVAC plumber", not "healthcare" or "construction". Selection rule: the prospect has an auditable site, and a defect on that site directly costs them money (a lost lead is a lost quote). If the observation does not hurt the prospect, change niche.
A single zone: a city, a metro area, a department. Count the pool by crossing the niche's NAF code with the zone on SIRENE or directories. Below 300, the zone is too small to justify the effort. Above 1500, tighten the geography, keep the trade. The goal is to exhaust one zone before opening another.
Open the site, note ONE defect confirmable in 10 seconds: broken form, slow mobile, no online booking, CTA below the fold. That observation becomes your first sentence, 18 words maximum, no superlatives. The prospect must be able to verify the fact on their own screen.
The proof: what the Audit does to reply rate
The Audit part is not a stylistic detail. Across 40 local outbound campaigns (~18,400 emails, French B2B niches, first half of 2026), we measured reply rate by the personalization level of the opener.
The verifiable audit is 4.6x the generic opener and doubles plain merge fields. The full breakdown of this study is in what 40 local cold email campaigns taught us. Here, keep the mechanics: the niche constraint is what makes the audit possible, and the audit is what makes the email credible.
Why the Niche multiplies the Audit effect
The same observation is not worth the same across niches. Where a site defect directly costs the prospect a customer, the audit opener explodes. That is why the niche choice comes first.
The same email goes from 5 to 11% just by changing niche. For more on picking the local niche, see how to choose a local niche that replies. And for the specific case of web agencies, our guide 50 leads for a web agency without LinkedIn applies the method end to end.
The full method, as a copyable checklist
Here is the entire method as a checklist. You run it in order, box by box, for each new campaign. It is tool-agnostic: nothing in it requires a particular tool.
NICHE x ZONE x AUDIT METHOD: CHECKLIST
1. NICHE (pick a narrow trade)
[ ] One trade, not a category ("dentist", not "healthcare")
[ ] The prospect has a site you can audit
[ ] A site defect costs them money (a lost lead = a lost quote)
[ ] You can state the observation in one sentence
2. ZONE (one geography, 300 to 1500 businesses)
[ ] A single zone (city, metro area, department)
[ ] Count the pool: NAF code + zone on SIRENE / directories
[ ] Too small (<300): widen the zone, not the niche
[ ] Too large (>1500): tighten the zone, keep the niche
[ ] Goal: exhaust the zone before opening another one
3. AUDIT (a verifiable observation becomes the opener)
[ ] Open the site, note ONE fact checkable in 10 seconds
(broken form, slow mobile, no online booking, CTA below the fold)
[ ] Banned: vague praise, superlatives, "I love your work"
[ ] The prospect must be able to confirm the fact on their own screen
[ ] One sentence, 18 words max, neutral tone
SENDING
[ ] 3-touch sequence, only the first sentence changes
[ ] 20 to 40 sends / day / inbox, ramped up gradually
[ ] Measure reply rate per niche, not per emailThe trickiest part is the audit opener. Here is the exact prompt we use to generate it from a prospect's public data. The 18-word constraint and the superlative ban are what stop the AI from sliding back into flattery.
You are an SDR. Here is a prospect's public data:
- Trade: {{trade}}
- Zone: {{city}}
- Site: {{url}}
- Audit findings: {{site_score}}/100, {{defect_list}}
- Google rating: {{rating}} ({{review_count}} reviews)
Write ONE cold-email opening line (18 words max) that:
1. cites a VERIFIABLE fact from the site, confirmable in 10 seconds (no praise),
2. uses zero superlatives ("great", "amazing", "leader"),
3. is specific to a {{trade}} in {{city}}, not generic,
4. makes the reader feel a human actually opened their site.
Return only the sentence, no quotes.Open the demo campaigns and judge the lead quality yourself, no card required. Worst case, you lose ten minutes.
Where AutoLeads fits
None of the above requires AutoLeads. You can list a niche across a zone by hand via SIRENE and directories, audit each site with Lighthouse, write the openers yourself, and send with the tool you already know. AutoLeads industrializes the three parts: it lists the niche across the zone, runs the audit on each site, generates the opener, and pushes it all ready to send. We accelerate the method, we do not replace it. The same system applied solo is detailed in our 100-euro solopreneur prospecting system.
If you remember one thing: never start with the email. Start with the niche, calibrate the zone, open the site. The message flows from the constraints, never the other way around.
Frequently asked questions
Why such a narrow niche? I am shrinking my market.
You shrink the market, not the revenue. A narrow niche makes the audit observation precise and the message credible: a dentist knows you are talking to dentists. A message for "SMBs" talks to no one. You exhaust one niche, move to the next, and reuse 90% of the work.
How do I count the 300 to 1500 businesses in a zone?
Cross a NAF code (your trade's INSEE activity code) with a geographic zone on SIRENE, professional directories, or a sourcing tool. Below 300, your zone is too small to sustain a campaign. Above 1500, tighten the geography rather than widening the trade.
What is a verifiable audit observation?
A fact the prospect confirms by opening their own site in 10 seconds: a broken contact form, a slow mobile site, no online booking, a call-to-action below the fold. Not vague praise. The audit is not flattery, it is proof you actually looked.
Do you need AutoLeads to apply the method?
No. You can list a niche across a zone by hand via SIRENE and directories, audit each site with Lighthouse and a look at the mobile view, and write the openers yourself. AutoLeads industrializes those three steps so you do not do them by hand on 800 leads, but the method works without us.
Is B2B email prospecting legal in France?
Yes, under conditions. In B2B, email prospecting is allowed without prior consent if the message relates to the recipient's professional role, with a clear opt-out. See the CNIL guidelines on commercial prospecting. In B2C, prior consent is required.
How long before I see results?
Count on a full zone, so 2 to 3 weeks of staggered sends, before judging a reply rate. A one-week campaign on 50 leads proves nothing statistically. The method is read at the scale of a whole niche, not a batch.