A manufacturer's Meta lead gen stopped producing leads the sales team ignored

Representative examplesEvery case study in this library is an illustrative composite of the kind of engagement we deliver — written to show our method and standards, not to name clients.

Where they started

Machine shops — the specialty-equipment manufacturer's genuine buyers — form a small, identifiable population: a few thousand firms running specific equipment lines, reachable through exactly the trade spaces they already inhabit. The company builds configurable equipment sold by a direct sales team of three, each carrying a quota and each protective of their calling time. Meta lead forms had been running for a year on broad targeting, and the sales team's mornings had settled into deleting the overnight batch before coffee.

What it was costing

Lead forms produced 140 leads a month; sales called 20 of them. Most were students, job seekers, and the curious — the targeting was broad, the form asked nothing qualifying, and the sales team had learned to ignore the source.

What they could see

• Sales opened the lead sheet each morning to find students, job seekers, and one-time browsers mixed with the occasional real shop.
• Most form submissions had no equipment type and no volume — the form asked for a name, an email, and nothing else.
• The CRM tagged the source as Meta and the sales team filtered it out of call plans by default.
• Cost per lead looked enviable on paper, which made the sales team's refusal to call them look like stubbornness.
• Nobody could say what the pixel was actually counting — the Conversions API setup had never been verified against real events.

The constraints we worked inside

• The genuine buyer was niche — machine shops with specific equipment — and small targeting was the point, not the problem.
• Sales would judge the channel on the first batch; quality had to show immediately.
• The Conversions API setup had never been verified — attribution truth was unknown.

What had been tried before

Broad-match lead campaigns ran with higher budgets to let the algorithm find the niche.

The algorithm found volume, not machine shops — students and job seekers convert cheaply on forms, so broadening made the junk ratio worse.

Sales began emailing every incoming lead a qualification questionnaire before spending a call on them.

It pushed the qualifying work onto sales instead of the form; response rates were near zero and the team's disregard for the source hardened.

Lookalike audiences were built from people who had engaged with the page.

The engager pool was mostly students, so every lookalike inherited the same wrong profile and spent more to find it.

What we proposed

We proposed making the form do the qualifying. The three questions sales actually asks — equipment type, monthly volume, timeline — went into the lead form, so friction filtered the curious before a lead reached the CRM. Before any budget moved, the Conversions API setup was verified with real conversions so attribution meant something. Engaged non-converters became a small retargeting pool, and sales-accepted or rejected feedback flowed back weekly to steer targeting and questions.

Just as important is what we ruled out, and why:

• LinkedIn lead gen — The buyer exists there but the budget couldn't sustain its click prices for a niche of machine shops; Meta's reach inside the niche was cheaper to test first.
• Gated technical whitepapers as the conversion — A download converts students easily — it would have reproduced the exact junk pattern the form questions exist to filter.
• Search-first with high-intent keywords — Search demand for this equipment niche was thin and bid prices reflected the few bidders; Meta's targeting could reach shops not actively searching.

How the work ran

Delivered by the growth pod — paid social specialist over 6 weeks, with working increments reviewed with the client every week.

The stack, and the reasoning

Meta lead forms

Native forms with the three qualifying questions added friction exactly where it filtered — sales judged the first batch on quality, and the form was the quality control.

Conversions API (verified)

The setup had never been checked; verification with real conversions happened before budget moved so optimization never chased double-counted or phantom events.

Retargeting

A small pool of video watchers and pricing-page visitors gave the niche buyer a second touch without reopening broad prospecting.

CRM feedback loop

Accepted-or-rejected with reason, pushed weekly, taught the account which form answers predict real buyers — the qualifying questions earn their keep through this loop.

What went wrong

How we worked together

Cadence

A Monday call with the sales lead and the marketing coordinator; the week's accepted-or-rejected sheet reviewed line by line, without exception.

Client side

The sales lead co-owned the qualification questions — nothing shipped to the form until he signed that the answers matched how he screens calls.

Decisions

Rejected-lead reasons from sales drove every targeting and question revision; the account never moved on a metric sales hadn't corroborated.

They provided

CRM access for the feedback field, thirty minutes of a sales engineer's time weekly to review form answers, and the first batch judged without mercy.

What changed

The headline: leads per month — but sales-accepted leads rose from 20 to 31 — 140 → 52, read from CRM sales-accepted records. A second check: cost per sales-accepted lead at −38%.

The sales team calls Meta leads first now — the morning delete is gone, and the sales lead's screening questions and the form's questions have converged into one shared language. The marketing coordinator reads rejected-lead reasons as targeting advice rather than complaints. The company stopped paying for volume it had to apologize for. What changed most is the standing between the two teams: marketing brings sales evidence, sales brings marketing constraints, and the weekly sheet is where they meet.

The result was read from CRM sales-accepted records against the pre-engagement baseline over the stated window, with a guardrail check on cost per sales-accepted lead. Where platform-reported numbers and business outcomes differ, this record says which layer it is quoting.

What they own now

• The three-question lead form with the rationale for each question written down.
• A verified Conversions API configuration and the staging-domain exclusion that fixed the pollution.
• The CRM feedback field and the weekly accepted/rejected reporting view.
• The retargeting pool definitions — video watchers, pricing visitors — with refresh rules.
• A short training recording for the sales team on reading form answers.