Why operating problems almost never live where they look like they live

Mike was sitting with his Q3 P&L on a Monday morning, and the numbers weren’t adding up.

Gross margin was down 4 points from Q2, about $80,000 on the quarter against an $8M annual run rate. The project mix was the same. Remodels in the $80K to $500K range, the work he’d been doing for twelve years. The crews hadn’t changed. Suppliers hadn’t changed. The only thing different was the gap between what each job was supposed to make and what it actually made, and that gap was widening on jobs his estimators had priced.

By lunch he’d decided he needed to retrain them. By Friday he was going to start interviewing replacements. By the following Monday he hadn’t done either, because something one of his lead carpenters said over coffee made him stop.

The carpenter said: “It’s not the estimate. It’s what we’re building that’s not in the estimate.”

This is the version of a Monday morning that almost every founder in a project-based service business has had. The names change. Agencies, consulting firms, design and engineering shops, specialty contractors. The numbers change. The function that catches the blame changes. But the shape doesn’t. There is a visible problem in the P&L, there is a person or function the founder believes is responsible, and the actual cause is somewhere else entirely.

The work is figuring out where they actually live, and most businesses don’t have a way to do that.

There are three reasons this happens, and they show up in nearly every diagnostic we run.

The first is that the visible symptom is downstream of the cause. By the time something is bad enough to show up in margin or schedule, the thing that caused it happened several steps earlier, often in a completely different function from the one that’s getting blamed. Mike’s estimators looked like the problem because their numbers were attached to the losing jobs. But the estimators were inheriting decisions and assumptions that had been made before they ever opened a takeoff. The estimate was where the bleeding was visible. The wound was upstream, and it had been bleeding for a while.

The second is that operators are pattern-matchers, not investigators. Founders who built the business by being great at the work develop sharp instincts for which things tend to go wrong. They develop weaker instincts for why those things go wrong, because the why was usually held in their head and never had to be examined. So when something breaks, the pattern-matching instinct fires immediately. “Estimating problem.” And the investigation never starts. The instinct feels like knowing. It usually isn’t.

The third reason is the one that matters most. The team usually knows. The people closest to the work can sense exactly where the friction is. They feel it every week. But they don’t have language for it in a way that travels up to the founder, and even if they did, there is rarely a structured moment where that language gets surfaced. So the founder makes decisions based on what’s nameable, not on what’s true.

Operators see the result, not the cause. That isn’t a failure of intelligence. It’s a structural feature of how project-based businesses work.

Here’s what we found when we looked.

The named concern was estimator accuracy. Mike’s estimators had two decades of combined experience and a reputation in the local market for getting the numbers right. They were not, on any reasonable measure, the problem.

A remodeling project doesn’t get priced once. It gets priced in layers. A budget range during the first conversation. A refined number after the site visit. Allowances built in where selections aren’t final. Trade bids gathered and leveled against the scope. And finally an internal margin review before the price goes to the client. The estimate that lands in production at the end of all that is the result of dozens of decisions made by different people across multiple stages. Each of those stages is an opportunity to introduce a gap between what the business is selling and what the business will actually build.

When we sat with the estimators, what they said was: “We’re estimating against assumptions we didn’t make. By the time the project gets to us, sales has already given the client a range. The client has anchored on it. We’re working backward to make the numbers match.”

When we sat with the sales side, which in Mike’s shop meant Mike himself and one other person, what we heard was: “The client wants a number before they’ll let us in the door. If I don’t give them something close, we lose the opportunity. So I give my best read based on what they’ve described, and we sort out the details in design.”

When we sat with the production lead, what we heard was: “I get the budget by cost code. I don’t always know what was in the estimator’s head when they built it. The first time I find out about a verbal commitment is usually when the client mentions it on site.”

Three different conversations. Three different views of the same project. No structured moment where they ever reconciled.

We pulled twelve months of completed jobs and walked the variance. The pattern was specific, and once we segmented the data, the average stopped telling the whole story. Roughly 40% of Mike’s jobs were running close to target gross margin. The other 60%, the ones with verbal scope commitments that hadn’t made it into the written estimate, or allowances set below the client’s actual taste level, or scope assumptions production discovered mid-build, were running 8 to 11 points under. The 4-point quarterly average was just where those two populations met.

That distribution matters. It meant the problem wasn’t a small, evenly distributed leak. It was a sharp concentration on a specific kind of job, hiding inside a quarterly number that looked like general drift. The verbal scope misses alone accounted for roughly $50K of the Q3 hit. Under-set allowances accounted for another $20K. Production discovering scope after starting drove the rest, usually as field decisions, rushed trade coordination, or rework crews absorbed without it ever showing up as a change order.

Annualized at the trailing rate, the structural leak was running closer to $300K a year. Not in any one place a P&L would flag, but spread across a recognizable category of job that the team was quietly absorbing.

The variance wasn’t randomly distributed. It was concentrated on the projects where the gap between what got promised, what got priced, and what got built was widest. The estimators were estimating accurately against the scope they were given. The crew was building to the actual promises. The margin was the difference between the two.

The driver wasn’t estimating. The driver wasn’t sales. It wasn’t production either. The driver was the absence of a layer that sat across all three. A recurring practice where the verbal commitments from sales, the assumptions and allowances built into the estimate, and the scope production was about to build all got reconciled into a single document before the price went to the client, and re-reviewed before production picked it up.

That layer didn’t exist anywhere in Mike’s business. Not because the team didn’t want it. Because nobody’s job was to build it.

The fix wasn’t training. It wasn’t a new pricing tool. The fix was three structured moments that hadn’t been happening, and the documents and decision rights that made those moments work.

The first was a scope reconciliation between sales and estimating before any number went back to the client. Verbal commitments captured, allowance levels set deliberately based on the client’s actual taste level, exclusions named in writing.

The second was a margin and risk review before the proposal. Not just “is this priced right,” but “where is the risk in this scope, and are we pricing it in or pricing it out.”

The third was a production handoff that wasn’t just “here’s the budget by cost code.” It was a conversation that explained why the estimate was built the way it was, where the allowances sat, and which assumptions production needed to defend in the field.

Three meetings. A few short documents. Decision rights named clearly enough that the team knew who owned each call.

Six months later, Mike’s margin variance was back inside historical range and his project-to-project consistency had tightened. He didn’t fire anyone. He didn’t buy a new system. He surfaced a layer that should have existed and hadn’t.

That’s what most of this work looks like once you can see it. It’s quiet. It’s specific. And it’s usually buildable from pieces the business already has.

What happened in Mike’s shop isn’t a remodeling problem. It’s a project-based service business problem, and it shows up in roughly the same shape across very different industries.

We worked with an agency where the named concern was project profitability. Specific accounts kept losing money despite billable hour targets being hit. The founder believed the problem was their billing rate. When we dug in, the actual driver was that scope, pricing, and delivery were governed by three different people who never reconciled their assumptions in writing. Account managers were absorbing client requests they didn’t feel authorized to push back on. The design team built to those requests. The numbers came in three weeks later. The fix wasn’t a new rate card. It was a clarified decision right on scope changes, a short script for account managers, and a recurring touchpoint where account, design, and delivery reviewed live projects together.

We worked with a consulting firm where the named concern was missed deadlines. The founder believed the problem was project management discipline. When we mapped it, three of the four delayed projects were waiting on the same senior partner for the same kind of approval. A decision that had quietly migrated upward over the years and never been redistributed as the firm grew. The bottleneck wasn’t PM discipline. It was a decision-rights gap that crossed every active engagement. The fix wasn’t a project management tool. It was identifying which kinds of approvals could move down a level, writing the criteria for when they could, and giving the senior partner permission to let go.

The pattern, across all three: the surface concern lives inside a function. The driver lives at the seams between functions, or in a missing layer that should have sat across them.

So what changes when you can see the seam?

The big-feeling problem becomes a small number of specific moves. A recurring meeting that didn’t exist. A document that needed writing. A decision right that needed to move. Operators in this position don’t need a transformation. They need to see the right thing.

When we start with a client at Currenza Lab, we don’t try to solve the named concern. We start by separating what was stated from what was surfaced. What the operator believes is wrong, and what the work and the team actually reveal when we look carefully.

The first two weeks of an engagement are a diagnostic. We talk to the founder. We talk to the people closest to the work, usually three or four conversations of forty-five minutes each, across sales, estimating, design, production, and delivery depending on the shape of the business. We pull whatever the business already has in the way of project records, P&Ls by job, scope documents, allowance schedules, trade bids, the kind of artifacts that exist whether or not anyone is reading them. We look for the gap between what those documents say and what the team says.

The output is small on purpose. A map of three to five drivers, ranked by leverage, with a concrete first move on each one. Not a hundred-page report. Not a strategy deck. A short document that points at the specific things that, if changed, would change the outcome.

Most of the operating wisdom inside a project-based business already exists. It’s distributed across the team and buried in the work. The job isn’t to bring new wisdom in. The job is to surface what’s already there, in a form the business can actually use.

That’s most of the work. The rest is helping the business turn the map into something operational. A recurring touchpoint that didn’t exist. A document that needed writing. A decision that needed to move. Practical. Specific. Usable on a Monday morning.

Mike didn’t need to fire his estimators. He needed three structured moments that weren’t happening, and the documents and decision rights to make those moments work.

The point isn’t the story. The point is that there’s almost always a version of those moments missing inside a project-based service business, and the question is whether the business has a way to find them. Most don’t. Not because the people aren’t smart. Because the structural layer where it would surface doesn’t exist, and nobody’s job is to build that layer.

That’s the work we do.

If this is the kind of thinking you want applied to your business, the door is at currenzalab.com.