The conversion rate dropped to under one percent. That was the win.

"My leads are flat. I need better keywords."

An owner-operator who sources hard-to-find and obsolete parts and resells them by quote came to us with a clear request. The leads felt flat, so he wanted sharper keywords, more parts in front of buyers, maybe some organic ranking so he could stop paying for every click. He had run the account himself for years, and his instinct was an experienced one.

Underneath the request was the feeling every owner knows: doing the work, watching a dashboard that reported hundreds of conversions, and still not seeing more come back. It is a quiet kind of stuck, the kind where you stop trusting your own read on the business and start second-guessing every call you make. Because the screen said the marketing was working, the natural conclusion was that the lever left to pull was a tactical one.

That belief feels safe because nearly everyone shares it. Trust the dashboard, add keywords, chase a better rate: the moves every operator reaches for. The catch is that the dashboard was grading itself, so more effort poured into a number that was never describing revenue. Safe is not the same as true.

What we found when we opened the scoreboard

So before touching a keyword, we opened the one screen most owners never look at: the list of events the account had been told to count as a conversion. That is where the story came apart.

It was counting phone taps. It was counting clicks on get-directions. It was counting an old purchase goal that had recorded almost nothing in years. Several of these were stacked, so a single visitor could register as two or three wins. The 11 percent was the sum of those false wins, not the sound of money arriving. Of the hundreds of thousands of keywords running, only 143 had produced a conversion in the last 90 days.

You already know this in your own work: a thermostat reading a sticky note instead of the room will run the furnace all day and still leave the house at the wrong temperature, no matter how hard it works. The account was that thermostat, locked onto a number that felt good, so every decision built on it leaned the wrong way. The harder he pushed, the further from the real room he got.

The intervention

We rewrote what the account was allowed to call a win. We removed the phone taps, the directions clicks, and the dead purchase goal, and we counted one event: a real quote request submitted by a real buyer. Nothing about the parts or the budget changed. We changed what the account was permitted to believe.

Then we set up the loop so a quote that turns into business could be fed back as the truth, the way a thermostat has to read the actual room temperature, not the number someone wrote on a sticky note. Until the account reads the real room, every adjustment is a guess. With one honest event in place, the gauges finally pointed at money, and we could see exactly where the remaining work lived.

The result, in context

Read quickly, that first number looks like a failure. It is the opposite. The 11 percent was never real. Sub one percent was the first honest baseline this account had ever produced, and an honest baseline is the only ground you can actually build on.

The number is the evidence, not the point. Over the same weeks the budget held steady while clicks and click-through climbed, which told us exactly where the remaining work lived: not in the traffic, but in the pages the traffic landed on. You cannot localize a leak until the gauges stop lying.

Figures above are the account's own in-platform reported numbers over a four-week window, agency measured and not third-party audited. Engagement targets and unit-economics models discussed at kickoff were planning figures, not results, and are not shown here.