The pixels were counting sales that never happened.
Ask a stuck B2B team about dead and duplicated conversion pixels and most will tell you their tracking is fine, because the dashboard keeps reporting numbers as if everything still works. For one profitable training firm, the pixels were counting events that never happened, and the part of the budget that actually started deals was quietly being starved. They thought they had a lead problem. They had a counting problem, and once we fixed it, one month's cost per deal fell by half.
The instrument was lying, not the market.
Some conversion pixels were dead and counted nothing. Others were duplicated and counted events that never happened. Last-click crediting then starved the display and video work that actually started the deals.
We rebuilt the tracking to read the truth.
Pixels that fire only on real submissions, last-click crediting replaced, and every reported number reconciled against the deals that actually closed, so cost per deal could finally be judged honestly.
"The leads went weak. We need more of them."
A profitable, family-run firm that had spent three decades teaching engineering, pharmaceutical, government, and corporate teams to write better came to us off a record year, then a quarter of softer leads, with a fair worry that the market or the website had turned. The ask was a tactic: hire an SEO team, buy more qualified leads, price out a website rebuild. It was a reasonable read, and the prior work they had done was genuinely sound.
Underneath it was the feeling every founder knows: a strong business, steady demand, and the sense of working harder for thinner results. After three strong decades, a slow quarter naturally looks like demand cooling or the site aging, so the conclusion almost wrote itself. The leads had gone weak, so they needed more of them.
That belief feels safe because nearly everyone shares it. When the numbers soften, buy more leads, rank for more terms, rebuild the site. The catch is that more of an unmeasured thing only adds spend to a number you have never checked is real. Safe is not the same as cheap.
What we found when we opened the account
So before buying a single new lead, we opened the ad account and checked what the conversions were actually made of. The instrument was broken. Some of the conversion pixels were dead and counted nothing. Others were duplicated and counted the same thing twice, recording events that never actually happened. On top of that, last-click crediting handed all the credit to the final touch and starved the display and video work that started the relationship in the first place.
So the account was being graded on a number that was not real, and it learned to chase whatever produced that fake number. When we traced the spend, two-thirds of the paid budget sat on keywords that had never produced a deal. Across roughly two thousand dollars of spend, only six keywords had ever converted at all.
You already know this in your own work, because this firm teaches one rule above all others: write for the reader, not for yourself. Their ad account was breaking that exact rule. It was optimized to what looked good on its own dashboard, not to what a real buyer actually did. A speedometer reading thirty while the car sits parked is not a slow car, it is a broken gauge, and you would never press the pedal harder to fix it. The prior tidy cost-per-lead number looked like progress, but it was built on pixels counting the wrong thing.
The intervention
We repaired the gauge before we touched the engine. We rebuilt the conversion tracking to fire only on real submissions, removed the dead and duplicated pixels so the account stopped counting events that never happened, and replaced the last-click crediting so the display and video work that started deals was no longer starved.
Then we reconciled every platform number against the deals that actually closed and treated those closed deals as the single source of truth. With the signal honest, we moved budget off the keywords that had never produced a deal and toward the terms real deals came from. Nothing about the training, the product, or the team changed. We fixed what the numbers were counting, and where the money pointed.
Repair the instrument, then judge cost per deal.
Dead and duplicated conversion pixels teach an account to chase the wrong visitor, and last-click crediting starves the work that started deals. Rebuild the tracking to fire on real conversions, reconcile every reported number to the deals that actually closed, and only then judge cost per deal. The gap between the reported cost per lead and the real cost per deal is almost always where the B2B budget leaks.
The result, in context
Once the pixels were rebuilt, the April numbers told a different story than the dashboard had been telling. Cost per deal fell by half from one month to the next, paid leads rose 27 percent, and scheduled-call deals climbed sharply. That 50 percent drop is a specific April result, not a permanent state, and the months after it moved around as spend was adjusted. The number is the evidence, not the point. The account could finally tell the truth, so it started chasing the visitor who buys instead of the one that merely fired a pixel.
Figures are real and client-attested (agency-measured), the cost-per-deal drop framed as a single April result, not third-party audited.
"It is not about what we like. It is about what our reader needs. The moment we started a marketing question with what we thought, that was already the wrong starting point."
Who this is for
If you run a profitable B2B service business, your leads have softened, and your first instinct is that you need more of them, this is for you.
Not because more leads is wrong. But if you have never checked what actually fires the conversion your ad account is counting, you do not yet know whether the number is real. A pixel can be dead, or duplicated, or firing on the wrong moment, and the dashboard will keep reporting as if nothing is off.
Here is the thing most teams have never stopped to look at. You teach your own customers, in your own way, to serve the buyer and not yourselves. An ad account that counts what looks good instead of what a real buyer did is breaking that same rule on you. You can probably name the number you trust most right now, and you have probably never checked what fires it.
Picture opening your own account and your pipeline side by side and, instead of asking where the deals went, seeing exactly which spend started the ones that closed. That is the shift. You stop being the founder who hopes the dashboard is right and become the one who knows. Not more leads, an instrument that finally agrees with how your buyers actually buy. The truth is only worth something once you can see it, and right now you cannot.
We open your ad account and your pipeline side by side and show you the gap, on your own numbers, before anyone touches a campaign. It is not a sales call. It is a Second Opinion for B2B firms that have real demand and cannot see why more spend is not producing more deals.
Clarity is only the start. When you can see how your business really works, the hard calls get easier and the thing begins to run on its system instead of on you. That is what we are actually building toward.
Before you buy more leads
How do dead or duplicated conversion pixels hurt my ads?
A dead pixel counts nothing, and a duplicated pixel counts the same event twice or counts events that never happened. Either way the ad platform optimizes toward a number that is not real, so it learns to bring the wrong type of visitor. The dashboard can look busy and healthy while the deals do not follow.
What is last-click attribution and why does it starve my best campaigns?
Last-click attribution gives all the credit to the final touch before a sale. That starves the earlier campaigns, such as display and video, that actually start the relationship, because they rarely get the last click. The platform then pulls budget away from the work that began the deal and feeds the touch that merely finished it.
How do I tell if my ad account is counting the wrong thing?
Open the conversion the platform is counting and check what actually fires it. If it fires on a form open, a duplicated tag, or a pixel that no longer works, the number is not real. Reconcile it against the deals that actually closed before you judge cost per deal. In one B2B account, two-thirds of the paid budget sat on keywords that had never produced a deal, while a single exact-match term quietly ran a 4.83 dollar cost to acquire a customer against 266 dollars on broad.
Wondering if your ad account is counting the right thing?
The first step is seeing what your conversions are actually made of. The diagnosis is independent and yours to keep. There is no obligation to have us rebuild anything, and no half-answers that end in a referral list. It is a process, not a pitch.
Start with a Second OpinionA representative engagement from the Business JetPack Diagnosis Library. Identifying details have been removed or changed to protect client confidentiality. Figures are real and client-attested, not third-party audited, and reflect a specific engagement; the cost-per-deal change described occurred in a single month and is not a sustained state. Results vary.