The dashboard said $490 a lead. The real number was $977.

Two years, this account had been running. Two years of nobody stopping to ask if the number on the dashboard actually meant anything. We did. Turns out it didn't. The reported cost per lead was $490. Once we stripped out the conversions that weren't really conversions, the true number was $977. Double.

And the reason was almost funny, if it hadn't been costing them real money. Roughly half of what the account was counting as leads were job seekers clicking through to the careers page, not buyers, not anywhere close. The client had been paying to attract people who were never going to hand over a cent, and toasting a CPL number that didn't mean anything.

That was enough. The client signed on for the full engagement. First thing we noticed: zero presence on Meta, every dollar was sitting in one Google-shaped basket. So we built six new creatives on our own framework and put them in market. Three came back winners.

Sixty days later, the story had changed. Leads more than doubled. Cost per lead on the whole account was down over a third. Lead-to-sale conversions had doubled too. The winning creatives alone dropped cost per lead from $1,000 to $175, converting 50% better than the old Google-only ad, which we kept running on purpose, just so they could watch the difference for themselves.

Four things. Repeated every single day, no exceptions. We checked spend, revenue, and profit daily, nothing sat around for a week waiting on someone to build a report. New hooks went live in batches, whatever worked got fed more budget, whatever didn't got cut fast, no sentimental attachment to any of it.

We also went and figured out who was actually installing the thing and, more importantly, who was actually paying afterward, then wrote the ads for those people specifically instead of everyone with a phone.

And underneath all of it, one clear rule decided when to push spend up and when to pull it back. That's the part everyone skips, and it's the reason the account stayed profitable all the way up to $242K a month, not just at the small, safe-budget stage where profitability is easy for anyone.

Ninety days. That's it. The account went from a small test launch to $242K a month in ad spend, pulling in $508K in revenue and $266K in profit, $4.5M a year annualized, off 15.8M installs. And every single month on the way up stayed profitable. No dips, no "trust the process" quarter.