The First Week of Every Project Is the Most Expensive One You're Not Billing.
October 28, 2026

The First Week of Every Project Is the Most Expensive One You're Not Billing.

Let's do some quick math, and I want you to use your own numbers when you're done reading, because mine are made up. Say your engineers burn 6 hours of discovery on a typical project that never shows up on an invoice. They poke around the environment, read through old documentation, track down who set up the firewall in 2019, and sort out what's really going on before they can say what the project needs. Now say you run 30 projects a year, and your engineers cost you about $175 an hour once you add up everything that goes with them. That's 6 hours times 30 projects times $175, which comes out to $31,500 a year, and nobody on your team ever made a decision to spend it.

That's the version where discovery stays at 6 hours. In our experience working with MSPs it rarely does, because nothing is holding it there.

It usually goes like this. A project gets sold, a date goes on the calendar, and the engineer gets told to get familiar with the environment and come back with a plan. Nobody says how long that should take. Nobody bills for it. The hours either get buried in the project budget or they don't get tracked anywhere at all. Most MSPs treat discovery like a warm-up, when it's really the first part of the work.

And without a limit, discovery stretches. A thorough engineer keeps digging because there's always one more thing to check. A careful engineer keeps digging because if they miss something, it's on them. Neither one is slacking off. The setup just never gives them a reason to stop. Time you don't price and don't limit gets spent, and it gets spent every time.

The fix has three parts and none of them are complicated. Bill it, box it, and end it with a plan.

Start by making discovery its own line item with its own hours and its own price. This works on the client side too, which surprises some owners. It tells the client the work is real, and it moves the surprises to the front. You find the problem in discovery and scope it with them up front, instead of finding it halfway through the project and sending a change order that makes everybody angry. If a client won't pay anything for discovery, that tells you something about how the rest of the project is going to go. If you're worried about sticker shock, plenty of MSPs credit the discovery fee toward the project once the client signs, and that works fine.

Next, put a clock on it. Pick the number of hours before discovery starts and hold to it. When the engineer gets to about 80 percent of the budget without a clear plan, that's the trigger to say so, and it needs to be normal to say so. If flagging it early feels like admitting a mistake, people will just keep quietly going until the hours are gone, and then you're right back where you started.

Last, and this is the new step, discovery ends with a written engineering work plan. The plan is how you know discovery is actually finished. It doesn't have to be fancy. I'd want to see the confirmed scope, the tasks in order, the estimated hours on each task, who owns each piece, what the client has to provide, any dependencies, the risks, how we cut over and how we back out if something goes wrong, what done looks like, and some target dates. One or two pages is plenty, and if it's longer than the project deserves, something's off.

Have one person own the plan, and have a second engineer or your service manager read it before the client sees it. That's not a committee, it's a quick check for gaps. It also catches the underestimates before they turn into overruns, which is where a lot of the margin goes.

Now here's the part I care about most as someone who manages technicians. Putting a finite window on discovery and requiring a written plan tells your team something they might not have heard clearly before. It tells them that thoroughness gets measured against a clock, that finishing on time is part of doing the job well, and that there's an approved way to ask for more time when the environment turns out to be worse than anyone expected. A lot of engineers, especially the good ones, would honestly love to stop exploring and just be told when they're done. This gives them permission.

Once the project wraps, compare the hours in the plan to the hours you actually spent, and find out why they're different. After a handful of projects your estimates get a lot better, your discovery windows get tighter, and the team starts to see the plan as theirs. Discovery hours are also billable hours, so they count toward utilization, which ties right into the weekly scorecard stuff I wrote about in another post.

Sometimes discovery turns up something much bigger than you thought. When it does, the plan gets rewritten and you go back to the client with the new numbers in front of both of you. You don't quietly absorb it. That keeps the plan honest, so it stays a baseline you can actually measure against.

A few things to watch for. Plans can turn into paperwork nobody reads. Discovery can be priced so high that clients walk. And a time box can be so tight that your engineers rush and miss things. The fix for all three is the same, keep the plan short, price it fairly, and adjust the box based on what the actuals tell you.

All of this belongs in your field guide, the discovery line item, the time box rule and the work plan template, so every project starts the same way no matter who's running it. If you read my post on scoping a PSA project you've seen the same idea at the company level, a clear scope with a hard date. This is that idea applied to each client project. If you want help building it out, that's what we do at www.builttorunmsp.com.

 

Frequently Asked Questions

Should an MSP charge for project discovery?

In most cases yes. Billing discovery as its own line item with a fixed number of hours signals that the work is real, protects project margin, and surfaces problems before the project is underway. Many MSPs credit the discovery fee toward the project if the client signs.

What should an engineering work plan include?

Confirmed scope, tasks in order, estimated hours per task, owners, what the client needs to provide, dependencies, risks, cutover and rollback steps, a definition of done and target dates. One to two pages is usually enough.

How long should project discovery take?

Set the number of hours before discovery starts and base it on past projects of the same type. When the engineer reaches about 80 percent of the budget without a clear plan, they should flag it so you can decide whether to extend or rescope.

 

Why does a time limit on discovery help with technician management?

It tells engineers that thoroughness is measured against a clock and that finishing on time is part of doing the job well. It also gives them a clear, approved way to ask for more time when the environment is worse than expected.

About the author
Adam Kuester

Adam Kuester

Adam Kuester has a PhD in genetics and a career built inside managed services, an unusual combination that shapes how he works. He spent time designing operations at an MSP before joining Bruce McCully to build Galactic Advisors, where he's served as VP of Special Projects. His focus has been operational: finding gaps, building systems, and turning expertise into tools MSP owners can use across a partner base of nearly 1,000 companies. Built to Run MSP is that same work in a different form, practical frameworks for MSP owners who are good at winning business and want to get equally good at running it.