Nobody has ever seen the 10%

What the misalignment statistics actually measure and what they can't.

A board deck has a number for almost everything. Bookings, pipeline coverage, cost per lead, retention — every euro that moved has a line, an owner, and a trend arrow. In the years I spent preparing and reading those decks, across commercial, product, and marketing roles, I never once saw a line called the leak.

And yet everyone can quote its size: around 10% of revenue, lost in the handoffs between marketing, sales, and customer success. It may be the most repeated number in B2B. So here is the uncomfortable part. Nobody has ever seen the 10%. Not the analysts who published it, not the consultants who quote it, not your CFO, and not me. Once you understand why, you will read every misalignment statistic differently. Including the one I use.

The number parade

Spend twenty minutes searching the topic, and you will collect quite a set. Companies lose 10% of revenue to misalignment. No, wait, 10 to 15%. No, wait, 38%, attributed to Forrester. A trillion dollars a year, attributed to everyone and no one. Seventy-nine percent of marketing leads never convert.

Now try to walk any of these back to an original study — the year, the sample, the method. Most trails go cold within two clicks. The 38% figure circulates widely with a Forrester label, yet I have never found the underlying research, and I have looked properly. The trillion-dollar figure is an aggregate that appears in vendor decks the way sea monsters appear on old maps: decorative, and drawn where the cartographer stopped knowing things. The "79% of leads" statistic comes from lead-management research that is almost old enough to vote, run on a world of web forms and cold transfers that barely resembles how complex B2B deals are bought today.

This is how statistics launder. A study makes a careful, narrow claim. A conference slide keeps the number and drops the caveats. A vendor blog cites the slide. A LinkedIn post cites the blog. Each retelling strips out the method and keeps the percentage, because the percentage is the part that sells. After enough hops, the number floats free of any evidence at all and becomes unfalsifiable, which is precisely what makes it so quotable.

For what it's worth, one figure does survive scrutiny better than the rest: a long-standing IDC (as cited) estimate that misaligned commercial functions cost companies 10% or more of revenue per year. It is roughly a decade old, and it is an estimate, not a fresh study. It is the only benchmark I use, presented as exactly that. But even the honest number has a problem the dishonest ones share, and it is a more interesting problem than sloppy sourcing.

A leak is something that didn't happen.

Here is the problem: a revenue leak is a counterfactual. It is the deal that would have closed if the handoff had held. The expansion that would have happened if onboarding had delivered what the proposal promised. The lead that would have converted if it hadn't gone cold in the queue between two teams' definitions of "qualified."

Your systems cannot record any of that, because systems record events — the meeting that was held, the stage that was changed, the invoice that was sent. A non-event has no timestamp, no owner, and no field. The lead that cooled in a handoff shows up as "no activity," which looks identical to a lead that was never any good. The account that churned because the deal was mis-sold shows up as a reason code picked from a dropdown, eleven months after the actual cause. The deal that was never created, because marketing and sales define the ideal customer differently, and the company slipped through the gap between definitions, shows up as nothing at all. The better your dashboards, the more convincingly they display everything except the thing that leaked.

The place I learned this wasn't a dashboard. It was a quarterly business review with every function head in the room, every number accounted for. Marketing had run several campaigns and delivered its lead count. Sales had closed one average-sized deal, but pointed to the big deals sitting in the pipeline, most likely to close this quarter. The retention team reported that a healthy number of customers had renewed, with a slight discount. Nothing anyone said was wrong. And I remember sitting there thinking: this engine is not moving.

What gave it away wasn't any single figure — it was that I had heard this meeting before. The campaigns and the lead count were last quarter's slide with new dates. The big deals had been "most likely to close this quarter" the previous quarter, too. The renewals with a slight discount had repeated quietly enough that the discount was becoming less an exception than a policy. Every function reported events, and every event was true. What no report could show was the non-events between them: the leads that never became pipeline, the pipeline that never became closings, the discount that never got questioned because a renewal counts as a win. The leak wasn't in anyone's numbers. It was the fact that the numbers didn't add up to motion, and no line in the deck existed to say so.

So, where does a figure like 10% come from if no company's ledger can produce it? From comparing companies with each other: firms that score as tightly aligned versus firms that don't, and the revenue gap between them. That is legitimate research work — and it yields an average spread across a population. An average across companies is a statement about the market. It was never a measurement of you. Quoting the 10% as your leak is like quoting the average household's electricity bill as your electricity bill. It tells you the category is real and roughly how big it runs. It tells you nothing about your meter, because you don't have one.

Three questions for anyone quoting a leak number

This suggests a simple test. When someone (a vendor, a consultant, an article, or me) tells you how much revenue you are losing to misalignment, ask three questions.

Where is the original study? Not the blog that cited it; the study. If the answer is a shrug, you have learned something about every other number in the deck.

What did it actually measure? An average spread between aligned and misaligned companies is a prior. A prior is useful — it tells you whether the problem is worth investigating at your size. It is not a finding about your engine.

Will you measure mine and re-measure it afterward? This is the question that separates an instrument from a sales pitch. A real answer names the evidence it will read (CRM records, financials, interviews, not a self-assessment survey), puts the result on a scale that stays constant, and commits to scoring the same scale again after the work, so the improvement is something you can show rather than something you are asked to feel. And a real answer accepts the risk that comes with measuring: the honest version of this work includes the possibility that your measured leak is smaller than the benchmark. If a diagnosis can only ever confirm the sales pitch, it isn't a diagnosis.

That is the standard I hold my own work to, and it is why I treat the industry estimate as a starting point and nothing more. The benchmark sizes the problem. Only your own data can locate it.

The line that isn't there

The leak will never appear in your board deck. Not because it isn't real — the research consensus, for all its laundered numbers, is pointing at something genuine — but because decks are built from what your functions record, and the leak lives in what happens between them, where nothing is recorded, and nothing is owned. That absence is not a reason to ignore it. It is the entire reason it has to be measured on purpose.

So, keep the 10%. Use it the way it deserves to be used: as a prior that says this is probably worth one afternoon of your attention. Then go get your own number, measured from your own evidence, on a scale you can score again. The board deck will still have a number for almost everything. The difference is that the line that was never there will finally have an answer, even if it never gets a row.

Want this measured for your own engine? Start with a €750 COS Efficiency Diagnostic.