The EBITDA bridge in the IC memo was correct the day it was built. Entry EBITDA on the left, required exit EBITDA on the right, six labeled bars in between showing exactly how the business gets from one to the other: organic volume, pricing, cost and inflation, M&A and bolt-ons, one-time items, operational initiatives. It's a clean, confident chart, and it's usually the single most-referenced page in the deck.
Two quarters into the hold, it's already fiction. Not because the math was wrong -- because a bridge is a snapshot of a thesis, and a thesis is a living argument about a business that doesn't hold still. The pricing bar assumed a mid-year increase that got pushed to next quarter. The operational-initiatives bar assumed a procurement win that came in at half the size. Organic volume is actually carrying more of the number than anyone modeled, which sounds fine until you notice it means the things the deal was underwritten on aren't the ones doing the work. Nobody updated the chart, so the chart still says something that stopped being true weeks ago.
Building the EBITDA Bridge Was Never the Hard Part
Search for an EBITDA bridge template and the results are consistent: CFO.com and a scattering of PE-focused blogs explain what a bridge is, ValueBridge.net sells a static template as one line item in a library of a hundred private equity deliverables, Corporate Finance Institute has its own template resource, and a handful of template marketplaces and SlideTeam decks round out the field. Every one of them treats the artifact -- the six labeled bars, the waterfall shape -- as the deliverable.
The clearest evidence that the artifact was never the differentiator is Sourcetable, an AI-powered spreadsheet tool that now auto-generates an EBITDA bridge template on request. If a general-purpose AI tool can spit out the structure in seconds, the structure was never the hard part. Nobody needs help drawing six labeled bars between two numbers.
What all of them stop short of is the part that determines whether the deal hits its return: once the bridge exists, which of the six components is actually moving this quarter, and does that still add up to the number the thesis was underwritten on? A template is a static picture of a plan. A hold period is dynamic. Nothing found across two rounds of research does the ongoing work of re-attributing quarter-over-quarter movement back to the six components and checking it against the original thesis -- everyone stops at the picture.
Why the Drift Is Structural: Five of the Six Bars Have No Owner
The bridge decays for a boring, structural reason before it decays for any interesting one: you cannot put a person's name next to most of the bars.
"Cost and inflation" is not a human being. Neither is "one-time items," and neither, really, is "organic volume" -- it's an outcome that a dozen people influence and nobody is accountable for. Of the six standard components, exactly two are the kind of thing a named executive owns and reports on in a monthly operating review: pricing and operational initiatives. So those two get watched, chased, and escalated. The other four get reported once a quarter as arithmetic residue -- whatever's left after the two watched bars are accounted for.
That asymmetry produces the drift. The four unwatched bars absorb every surprise in the business, in both directions, silently. A price increase that was supposed to land in Q2 lands in Q3, but the P&L still shows EBITDA roughly on target -- because organic volume ran hot and quietly covered the shortfall. The aggregate number looks fine. The composition underneath it has already diverged from the thesis, and composition is what the next four quarters of capital allocation, hiring, and bolt-on decisions depend on.
Consider a composite pattern, illustrative of what shows up repeatedly across portfolio-company hold periods rather than any single company or engagement: the entry bridge attributes 40% of the required EBITDA gain to operational initiatives -- a vendor renegotiation and a push to lift revenue per employee -- and 30% to pricing. We size procurement opportunities at a 7% base case against addressable spend, which is a defensible number and also the reason it's the first bar to disappoint: 7% of a real spend base is a big enough figure to anchor a thesis on, and small enough per contract that a single stalled vendor conversation erases a third of it. By quarter three, that's exactly what's happened, and the revenue-per-employee work has been quietly shelved while the team fights a customer-retention fire.
The bridge in the board deck still shows the original 40/30 split, because nobody rebuilt it. What's actually closing the gap is an insurance settlement and a strong volume quarter from one large account -- both of which sit in the unwatched bars, and both of which evaporate the moment that account renews at a worse price or the settlement doesn't repeat. The thesis looks on track. It is being carried by two components that were never supposed to be load-bearing, and the sponsor won't find out until the year the settlement doesn't show up.
That's not a diligence failure or a bad original bridge. It's what happens to any point-in-time model left unexamined against a live P&L. The six components constantly trade places as the source of the number; a bridge built once at entry has no mechanism for noticing which one currently deserves the credit or the blame.
The Left Side of the EBITDA Bridge Moves Too
Here is the part almost nobody re-examines, and it's the one that matters most.
Every bridge treats entry EBITDA as a fixed anchor. It's the number on the left, it came out of diligence, it's settled. But entry EBITDA is not an observed fact -- it's a normalized figure built out of add-backs, and add-backs are predictions. Each one is a claim that a cost was genuinely non-recurring. Twelve months into the hold, some of those claims are simply wrong. The "one-time" legal spend recurred. The owner's compensation normalization didn't survive the retention package the new CEO negotiated. The system-implementation cost that was going to end, didn't.
When an add-back recurs, it was never an add-back. Which means entry EBITDA was overstated, which means the gap to required exit EBITDA was larger than the model said from the first board meeting, and every bar sized to close that gap was sized against the wrong target. This is why a serious quality of earnings review is worth revisiting after close and not just before it -- the add-backs you accepted at signing are testable hypotheses, and the hold period is the test.
Restating entry EBITDA when an add-back recurs is the single most valuable revision anyone can make to a bridge, and it's the one revision almost never made -- because it's the only change to the chart that makes the deal look worse. Every other update moves a bar sideways or reallocates credit between owners. This one mechanically widens the gap and puts pressure on everything downstream of it. Skip it and you get to feel good for several more quarters, then discover the whole gap at once -- usually in the year you were planning to run a process.
What Continuous Re-Attribution Actually Requires
Keeping a bridge honest through the hold means doing, every quarter, what the entry model did once: taking the period's actual EBITDA movement and classifying it back into the six components, then comparing that composition against what the thesis assumed. Four things make this different from re-running the original exercise:
- It has to reconcile against the original assumption, not just describe the quarter. "Pricing contributed $400K this quarter" is a fact. "Pricing was supposed to contribute $1.2M by now and has delivered $400K" is the finding that changes what the board does next. Price realization drifts quietly in exactly this way -- the increase was announced, the realized rate never followed.
- Every bar needs a named owner and an escalation threshold, including the ones that don't naturally have one. Somebody owns "cost and inflation," even if the honest answer is the CFO. An unowned bar is an unwatched bar, and unwatched bars are where the surprises accumulate.
- One-time items have to be flagged as structurally non-repeatable, and entry EBITDA re-tested against recurring add-backs. A component that closes the gap this quarter but cannot repeat next quarter is the most common way a bridge looks healthy while the thesis quietly fails.
- It has to run on the operating-review cadence, not the annual one. A bridge re-attributed once a year catches drift after twelve months of decisions have already been made on the wrong composition.
One practical trap worth naming: the planning bridge and the attribution bridge are not the same object. A deal-attribution bridge decomposes into organic volume, pricing, cost and inflation, M&A, one-time items, and operational initiatives -- it answers where did the number come from. A planning bridge decomposes into volume growth, price/mix, gross margin expansion, operating leverage, investment drag, and efficiency gains -- it answers what mechanism will produce the number. Both are legitimate and both have six bars, which is precisely why they get conflated. Reporting one against the other is the most common reason a quarterly bridge "doesn't tie" and gets abandoned by Q3.
And the number actually worth tracking isn't the whole bridge. It's the minimum viable bridge -- the subset of high-confidence components that on their own close the gap to base-case MOIC. Everything above that line is upside; everything inside it is the deal. A bridge of point estimates with no confidence rating attached to each bar isn't a plan, it's a wish list, and it gives you no way to answer the only question that matters at a quarterly review: not "are we on the bridge," which is always debatable, but "is the minimum viable bridge still intact," which has a yes or no answer every ninety days.
Performis's Value Creation Plan engagement builds the entry-to-exit EBITDA bridge this way -- initiative-by-initiative attribution with confidence intervals rather than point estimates, a named owner and RAG threshold on every component, entry EBITDA re-tested against recurring add-backs, and the minimum viable bridge identified explicitly -- delivered within 60 to 90 days of engagement start and then carried into the monthly operating review and quarterly board snapshot, alongside the 100-day plan it's built from. The Long-Range Plan engagement builds the six-component planning bridge for the annual cycle, as a formula-driven model rather than a slide, so the mechanisms are auditable. Both exist because a bridge that isn't re-attributed against the plan it came from isn't tracking anything -- it's a chart with last quarter's date on it.
If your board deck still shows the bridge from the IC memo, the real question isn't whether the chart is well built. A tool can build that part for you in seconds. It's whether anyone can tell you, this quarter, which of the six components actually moved -- and whether the entry number on the left is still true. Start a Value Creation Plan engagement and find out before the next board meeting asks the question for you.