The Estimation Accuracy Benchmark: Where Does Your Business Stand?

Ask a service business owner whether their quotes are accurate, and you will likely hear something like "pretty good" or "we're usually close." Press for specifics—actual percentage variance, average overrun, patterns by project type—and the answer becomes less certain. Most business owners believe they estimate well because they have stayed in business. But staying in business is a low bar. The question is whether estimation errors are silently eroding margins that could fund growth, better equipment, or higher wages.

The uncomfortable truth is that most service businesses do not know their estimation accuracy because they do not measure it. They quote, deliver, invoice, and move on. The gap between quoted cost and actual cost disappears into the general noise of running a business. Only at year-end, when the accountant delivers the profit figure, does the cumulative impact become visible—and by then, it is too late to trace back to individual projects.

Understanding where your estimation accuracy stands requires two things: benchmarks that provide context, and measurement that provides specifics. Neither alone is sufficient. Benchmarks without measurement tell you what is typical but not where you sit. Measurement without benchmarks tells you a number but not whether it is good or bad.

What the Industry Data Reveals

Research from Deltek found that only 25% of architecture and engineering firms deliver most of their projects on or under budget. That figure should give every service business owner pause. Three quarters of firms in industries built around project delivery routinely exceed their budgets. These are not amateurs. They are established businesses with years of experience, yet estimation accuracy remains elusive.

The pattern holds across sectors. Professional services firms commonly see project overruns of 15-25%, with the variance concentrated in revision cycles, scope additions, and underestimated complexity. Trades businesses operate on margins of 8-20% depending on the trade, which means estimation errors of even 10% can consume half the expected profit or more.

The gap between perceived accuracy and actual accuracy is particularly striking. Business owners who feel they estimate "pretty well" often discover, when they measure, that their average overrun sits at 12-15%. The projects that finished under budget fade from memory. The disasters get rationalised as exceptional circumstances. What remains is a vague sense of competence that the numbers do not support.

Consider a contractor who believed their estimates were reliable. Twenty years in the trade, a solid reputation, repeat clients. When they actually tracked quoted versus actual costs across 30 projects, the average overrun was 14%. Not on the difficult jobs or the ones with unexpected complications—across all jobs, as an average. Some came in under. Some came in significantly over. The 14% was the central tendency they had never seen because they had never looked.

Measuring Your Own Accuracy

Benchmarks provide context, but improvement requires knowing your own numbers. The measurement itself is straightforward, though it requires discipline to maintain.

For each completed project, record the quoted cost and the actual cost to deliver. Actual cost means total cost—labour hours at true hourly rate (including employer costs, not just pay rate), materials at purchase price, subcontractor invoices, travel, and any other direct expenses. It does not mean invoice amount. A project invoiced at £10,000 that cost £11,200 to deliver lost money regardless of what the client paid.

Calculate the variance as a percentage: actual cost minus quoted cost, divided by quoted cost, multiplied by 100. A project quoted at £8,000 that cost £9,200 has a variance of 15%. A project quoted at £12,000 that cost £10,800 has a variance of negative 10%—you overestimated.

Timeline accuracy follows the same logic. Quoted days versus actual days, calculated as percentage variance. A two-week project that took eighteen days overran by 29%.

Track both numbers for at least 10-20 projects before drawing conclusions. Individual projects vary. What matters is the pattern. An average overrun of 8% across 15 projects tells you something meaningful. A single project that overran by 30% tells you about that project, not about your estimation ability.

The discipline of recording this data, project by project, often produces insights before any analysis. The act of capturing actual costs forces attention to costs that were previously invisible—the three hours spent on client calls, the return trip for forgotten materials, the afternoon lost to a supplier error.

What Different Accuracy Levels Mean

Once you have your number, you need context for what it means.

Variance under 5% indicates strong estimation discipline. Projects occasionally exceed quotes, but the errors balance out across a portfolio. Margins are predictable. Cash flow forecasting is reliable. Businesses at this level can price competitively because they know their costs will not surprise them.

Variance of 5-10% is typical for well-run businesses. There is room for improvement, but the leakage is manageable. A business completing 80 projects annually at an average value of £8,000 with 7% variance is leaving roughly £45,000 on the table—real money, but not existential. Identifying the sources of that variance and addressing them is worthwhile but not urgent.

Variance of 10-20% represents significant margin leakage. The same 80-project business at 15% variance loses approximately £96,000 annually to estimation errors. That is a salary. That is equipment. That is the difference between growing and treading water. At this level, the problem is not random variation but systematic patterns that repeat across projects.

Variance over 20% indicates a fundamental problem with the estimation process. Quotes are closer to guesses than predictions. Margins are unpredictable, and cash flow problems likely follow. A business operating at this level needs to examine not whether individual estimates are wrong but why the entire approach is failing.

The financial impact scales with volume. A sole trader completing 30 projects annually at £5,000 average with 12% variance loses £18,000. A growing business completing 120 projects at £15,000 average with the same 12% variance loses £216,000. The percentage is identical. The pounds are not.

What Influences Accuracy

Estimation accuracy is not fixed. It varies by project type, estimation method, and the presence or absence of a feedback loop from delivery back to quoting.

Project complexity and variability matter. A business that delivers highly standardised work—same scope, same conditions, same timeline—can estimate more accurately than one handling bespoke projects where no two jobs are alike. A painter recoating identical new-build apartments has simpler estimation than one working on Victorian properties where every room presents surprises. This is not a criticism; it is a reality that should influence how much contingency to build in.

Estimation method matters more than experience. A business owner with 25 years of experience who quotes from memory and gut feel will often be less accurate than one with five years who quotes from data. Memory distorts. Recent projects loom larger than old ones. Dramatic successes and failures are remembered; average projects fade. Data does not suffer from these biases.

The feedback loop from completed projects to future quotes is the highest-leverage factor. A business that systematically compares quoted versus actual, identifies patterns, and adjusts future quotes accordingly will improve over time. A business that does not—regardless of experience—will repeat the same errors indefinitely.

A consultancy illustrates this dynamic. They tracked their projects for twelve months and discovered an average variance of 18%, concentrated in the revision cycles of brand strategy work. Clients consistently requested more rounds of feedback than quoted. Armed with this knowledge, they adjusted their scoping process—adding a revision round to standard quotes and pricing additional rounds explicitly. Within a year, variance dropped to 7%. Not through better intuition or harder work, but through measurement and adjustment.

The Starting Point

Benchmarks tell you that poor estimation accuracy is common. The Deltek figure—only 25% of firms delivering most projects on budget—confirms that struggling with estimation does not make you unusual. It makes you typical.

But typical is not aspirational. The businesses that improve their margins, stabilise their cash flow, and grow sustainably are those that measure their own performance and act on what they find. They discover that their "pretty accurate" estimates actually run 14% over. They identify which project types are predictable and which are volatile. They build contingency based on evidence rather than arbitrary percentages.

The number you find when you measure may be worse than you expect. Most business owners are surprised, and not pleasantly. But knowing the number is the prerequisite to improving it. A 15% variance that you can see is more valuable than a 15% variance that remains invisible until year-end. Visibility creates the possibility of action.

Start with your last ten completed projects. Calculate the variance. See where you stand. The benchmarks provide context. The measurement provides the starting point for improvement.