Growth engineering: what >1000 weekly signups taught me about funnels

TL;DR

Multiposting across 8+ platforms became our most profitable channel: >1000 weekly signups, >70% wizard conversion, <€10 CAC, 3x revenue in 90 days. The levers were rotation logic, per-channel CAC, and a wizard instrumented step by step. Growth engineering is plumbing with a scoreboard.

The target was >1000 weekly signups from job multiposting. We hit it, and the channel became the most profitable acquisition source we had: >70% conversion through the signup wizard, under €10 cost per signup, and a 3x revenue multiple within 90 days.

None of that came from a growth hack. It came from treating the funnel as an engineering system.

Distribution was the product

Multiposting means pushing one job to 8+ platforms. The naive version fires the same payload everywhere and calls it reach.

The engineered version treats every platform as its own funnel:

• Different field requirements, different review queues, different latency
• Rotation logic so budget flows to platforms that convert, not platforms that look big
• Per-platform cost tracking, so CAC is known per channel, not blended into mush

The rotation logic mattered more than any landing page. Budget follows evidence.

The wizard is a measurement device

A signup wizard with >70% conversion isn't "good UX" in the abstract. It's a series of steps where every drop-off has a name:

• Which platform sent the user
• Which step lost them
• What they did instead of converting

We instrumented all of it. When conversion dipped, the funnel told us where before anyone guessed. The dashboard wasn't a vanity project. It was the steering wheel.

FROM PROD

The most expensive myth in growth: "let's just get traffic first and fix conversion later." Traffic without conversion is a way to pay platforms to disappoint people at scale.

What I'd tell anyone building an acquisition channel

1. Know your cost per signup per channel, always. Blended CAC hides the channel that's burning money.
2. Rotation is a policy, not a setting. Revisit it on data, on schedule.
3. The conversion path is never done. >70% wasn't a finish line; it was the number that made the next experiment obvious.
4. ROI has a half-life. We tracked 90-day revenue multiples because signup volume alone lies.

The part nobody markets

Most of this work was plumbing: integrations, cost exports, automated reports, edge cases where a platform rejected a job for reasons nobody documented.

Growth engineering is mostly unglamorous systems work with a scoreboard. The plumbing is 90% of it. The scoreboard is why anyone tolerates the plumbing.

The scoreboard is why it's worth it. When the channel prints its own numbers (signups, CAC, ROI), the argument for the next investment makes itself. And if my numbers can't make that argument, I don't get to call it a growth channel. It's a cost center with confetti.