The Hidden Cost of Bad Estimates

I once watched a partner quote a data migration at $180,000. His gut said $180K. The team who’d actually do the work said $280K. He went with his gut because the client had budget constraints and he wanted to win the deal.

He won it. The project came in at $310,000 in actual cost. The team worked evenings for two months. Two people quit within six months. The client was unhappy because the last three weeks felt rushed and sloppy. They didn’t renew.

That $100K gap between the estimate and reality didn’t just cost $100K. It cost the firm closer to $600K when you add up the turnover, the lost renewal, and the reference that never happened. But nobody tracked it that way. In the system, it just showed up as a project that went a bit over budget.

The Anatomy of a Bad Estimate

Let’s be honest about where bad estimates come from. It’s not stupidity. It’s not incompetence. It’s almost always one of three things:

Optimism bias. We remember the project that went well, not the five that didn’t. When we estimate, we’re usually picturing the best-case scenario and calling it “realistic.” Every experienced consultant does this. We know we do it. We do it anyway.

Anchoring to the client’s budget. The client says they have $200K. Suddenly every scope configuration magically comes in around $195K. We’re not estimating the work. We’re reverse-engineering a scope that fits a number. That’s not estimating. That’s fiction writing.

No data to counter intuition. When the only input to an estimate is someone’s gut, there’s nothing to push back against bias. No historical actuals. No comparison to similar projects. Just vibes.

The Spiral Nobody Sees

Here’s what makes bad estimates so destructive. The damage isn’t contained to one project. It spirals.

Phase 1: The Overrun

The project goes over. This is the part everyone sees. The margin evaporates. Maybe you eat the cost, maybe you have a painful conversation with the client about a change order. Either way, the project is now underwater.

Phase 2: The Team Absorbs It

Someone has to do the extra work. Usually it’s your best people, because they’re the ones who care enough to stay late and fix things. They don’t complain, at least not at first. They just quietly start updating their LinkedIn profiles.

The cost of replacing a senior consultant is roughly 1.5 to 2x their annual salary when you factor in recruiting, onboarding, and the six months it takes before the replacement is fully productive. But we never connect that cost back to the bad estimate that caused the burnout.

Phase 3: The Client Loses Trust

Clients aren’t stupid. When a project runs over, they notice. They might not say anything. They might even acknowledge that scope changed. But internally, they’ve filed your firm under “can’t estimate their way out of a paper bag.” The next RFP goes to someone else. The reference call never happens.

Phase 4: The Pattern Repeats

And here’s the kicker: because nobody captures what actually happened, the next partner quoting a similar project makes the exact same mistake. Different client, different year, same underestimate. The firm pays the same tuition twice. Three times. Indefinitely.

“A bad estimate isn’t a mistake. It’s a tax you pay on every subsequent project you fail to learn from.”

Unpopular Opinion

Most service firms should raise their estimates by 20% across the board and see what happens. I’m serious. Most firms systematically underestimate because they’re terrified of losing deals. But here’s the math nobody does: if you raise prices 20% and lose 10% of your deals, you’re still making more money. And the deals you lose are probably the ones that would’ve gone over budget anyway.

The clients who walk away over 20% are the ones who were going to squeeze you on scope, fight every change order, and leave a lukewarm review. Let your competitor have them.

What Accurate Estimating Actually Requires

I’m not going to pretend there’s a magic formula. There isn’t. But I’ve watched firms go from chronic underestimation to reasonable accuracy, and they all did the same three things:

They started tracking actuals against estimates

Not in a spreadsheet that someone updates quarterly. In the system, automatically, on every engagement. When you can see that your ERP implementations consistently take 1.4x your estimate, you stop guessing and start adjusting.

They decomposed estimates into components

“This project will take 800 hours” is a guess. “Discovery will take 120 hours, design will take 200, build will take 320, testing will take 80, training will take 80” is five smaller guesses. And smaller guesses are more accurate. More importantly, when the project goes over, you can see which component was wrong.

They separated the estimator from the seller

The person responsible for winning the deal should not be the only person responsible for estimating the work. They have every incentive (conscious or not) to make the number smaller. A second voice, someone who’ll actually do the work or who’s at least accountable for delivery, changes the conversation entirely.

The Math That Should Scare You

Let me put some rough numbers on this. Say your firm does 100 engagements a year at an average of $150K each. That’s $15M in revenue. If your estimates are off by an average of 15% (which is conservative for most firms I’ve worked with), you’re burning $2.25M in unplanned cost annually.

But that’s just the direct cost. Add the downstream effects:

• Turnover from burnout: If you lose two senior consultants per year partly because of chronic overwork from underestimation, that’s another $400-600K in replacement costs
• Lost renewals: If bad delivery caused by bad estimates costs you even three client renewals, that’s potentially $500K+ in recurring revenue gone
• Reference damage: Every client who had a rough experience is a reference you can’t use, and a story they’ll tell to peers considering your firm

Add it up and you’re looking at $3-4M in annual damage from estimates that are “only” 15% off. On $15M in revenue, that’s 20-25% of your top line being destroyed by a problem most firms treat as a minor operational annoyance.

Bottom Line

Here’s what you can do tomorrow: pull your last ten completed engagements. Compare the original estimate to what actually happened. Hours, cost, timeline, all of it. I’ll bet at least seven of the ten came in over estimate. And I’ll bet the pattern is consistent by engagement type.

That pattern is your firm’s estimating bias. It’s real, it’s measurable, and until you see it in black and white, nobody will believe it exists. Once you see it, you can’t unsee it. And that’s when things start to change.