Estimate Your Next Project from Your Last Ten

Most teams estimate a new project based on what they remember from past ones, plus a little optimism. But the most accurate estimating data you have is already sitting in your finished projects: how long the work really took and where it ran over. As more projects move to fixed fees and outcome-based pricing, the teams that learn from that history will quote with confidence while everyone else keeps padding their numbers and hoping.

‍

The Number Everyone Remembers Wrong

Ask your team how long a 50-user Microsoft 365 migration takes. Someone will say three weeks. Someone else will say two, "if the client is ready." Your best engineer will say it depends.

Now pull up the last ten migrations you delivered. Say they averaged five weeks, with two taking more than eight.

That gap between what people remember and what really happened is where estimates go wrong. And it is entirely fixable, because the answer is already in your own data.

‍

Why Estimating Matters More in 2026

For years, time-and-materials billing hid a lot of estimating mistakes. If a project ran long, the client paid for the extra hours, even if nobody was thrilled about it.

That cushion is going away. TSIA's 2026 State of Professional Services argues that AI is eating into effort-based pricing, and that services organizations need to move toward pricing tied to outcomes. TSIA also notes that value-based and outcome-based pricing still make up only a small share of services revenue today. The shift is coming, but most teams are not ready for it.

Fixed and outcome-based pricing only work when you know what the work will really take. As TSIA puts it in its guide to services pricing, fixed pricing is convenient but carries the risk that project nuances throw revenue and effort out of line. Every hour you underestimate comes straight out of your margin.

MSPs are already feeling the squeeze. Among the MSP leaders we surveyed with The Channel Company, 44.4% called pricing and estimating a top challenge, and about half admitted their current project practices are hurting profitability.

‍

1. Pick Your Repeatable Projects

You don't need to analyze everything. Start with the three to five project types you deliver most often: migrations, onboarding, firewall replacements, device rollouts, or whatever makes up the bulk of your work.

These are the projects where history is most useful, because the shape of the work is similar every time. If you already standardize these projects with templates, you are halfway there.

‍

2. Look at Actuals, Not Plans

For each project type, gather what really happened on recent projects: total hours, calendar duration, and which phases ran over. Compare that with what was quoted.

This only works if time is logged against projects consistently. If your team treats time entry as a chore, it is worth making it easier to do before anything else, because estimates are only as good as the data behind them.

‍

3. Find the Pattern, Then the Outliers

Once you have ten or so projects of one type, the pattern usually jumps out. Maybe discovery always takes twice as long as planned. Maybe testing is consistently skipped in the quote but always happens in the field.

Look at the outliers, too. A project that ran 60% over usually has a clear reason: a client who wasn't ready, an unexpected legacy system, a hardware delay. Those reasons become questions you ask during scoping next time.

‍

4. Build Durations from What Really Happened

Update your templates so each task's duration reflects your actual history, not the best case. When task lengths are grounded in real delivery, the timeline built from them is far more believable.

Realistic durations also make conversations with clients easier. You can explain a five-week timeline because you have the receipts.

‍

5. Price from the Plan

When the plan is built from real durations and real effort, the price can come from the work itself instead of a gut feeling: what the work takes, plus the margin you need, with risk accounted for where your history says it lives.

It also changes how you handle buffers. Instead of padding every quote by 20% "just in case," you add contingency where your data shows it is needed. Your quotes get more competitive and more accurate at the same time. We have shared more on building prices that hold up for MSP projects specifically.

‍

6. Close the Loop Every Time

Estimating is not a one-time cleanup. After every project, compare quoted against actual and feed the difference back into the template.

In Moovila, templates carry durations and effort into every new project, and finished projects show exactly where time went, so the loop between what you planned and what happened stays tight without a spreadsheet exercise every quarter.

‍

Your History Is Your Advantage

AI will keep changing how services are sold and delivered. But TSIA's broader point holds: AI does not fix broken systems; it exposes them. The teams that win under fixed and outcome-based pricing will be the ones that already know, from their own track record, what their work really takes.

Your last ten projects have already told you. It is worth listening.

If you'd like to see how Moovila helps teams turn delivery history into better estimates, you can book a demo here.

‍