A multichannel ad engine, built from scratch
The problem
A growth-stage software company had an outbound engine and no paid ads. Nobody had settled which buyer the ads should reach or what to say to them. There were no campaigns, no audiences, and no way to tell what a lead cost or whether it ever became pipeline. The team was small and wanted to stay small, so an agency or a full-time ads hire was off the table.
The engagement
Build paid ads from nothing, light enough for a small team to run. Settle the buyer and the message first, then build everything else on that.
What we did
We started with the brief. It names the buyer by role, company size, and industry; the problem that buyer is trying to fix; the one message that answers it; and the proof behind it, drawn from the company's own platform data.
The offer is the proof itself, a benchmark report built from that data. Audiences come in three tiers, so the same offer gets tested at three levels of fit: everyone who matches the role, company size, and industry; named target accounts; and the sales team's own prospect lists.
Each campaign has its own capture form, so every lead can be followed through the CRM from lead to qualified to sales-ready to customer. Spend per audience sits against those stages, which gives the two numbers that matter: cost per qualified lead and cost per sales-ready opportunity. Each week, every audience is marked on or off target, with a one-line note on what to change.
A builder walks a new campaign through audience, message, ad, offer, and budget, with AI drafting each step from the brief and the company's data. An optimizer watches live campaigns and proposes changes: a budget shift, a bid change, a new ad, a new angle. A person approves each one before it spends.
All of it sits in one hub in the company's go-to-market dashboard, next to the outbound engine and the CRM data they share. Campaigns, ads, tests, proposals, and daily numbers are stored the same way for every channel, so adding a channel means connecting it and giving it a capture form.
The impact
The engine took ten weeks to build. The first program then went from brief to live campaigns in two weeks. The company's first campaign, run on its own, cost $553 per lead; the engine's campaigns run at under $160. Cost per lead is only the first number; budget moves on cost per qualified lead and cost per sales-ready opportunity. Spend has grown from $5K to more than $15K a month, around a hundred leads a month. The team runs it in a few hours a week: approve proposals, read the weekly notes, move budget. The brief, the measurement, and the AI tooling are built once, so every channel and campaign after the first starts from them.