The two-week CRM rollout for a small service team

Ask an owner why the last CRM did not stick and you will hear "it was too complicated." Push a little and the real story is almost always the same: someone tried to model the entire business before anyone had used it for anything. Three weeks in, the data was half-migrated, the team was running two systems, and the path of least resistance was the old one.

The fix is not a better tool. It is a smaller first version. Here is a plan that works for a 3–20 person service team.

Before day one: pick the leak

Choose the single biggest leak in the business. For most service teams it is one of four: leads not answered fast enough, quotes not followed up, invoices going out late, or repeat customers drifting away. Pick the one that costs you most, and be honest — the interesting one is rarely the expensive one.

This choice determines everything else, because it decides what "done" means in week one. Write it down as a sentence you can test: "Every quote we send gets three follow-ups without anyone remembering to."

Week one: one workflow, all the way

1. 01
   Days 1–2 — import only what is alive

   Active customers and open quotes. Nothing else. Nine years of history is not a prerequisite for anything, and it is the single most common place rollouts stall. You can import history in month three if you ever actually want it — most teams do not.

2. 02
   Day 3 — build the one workflow

   Just the leak you picked. Follow-up sequence, or the lead queue, or invoicing at completion. Resist every adjacent good idea; write them on a list you are allowed to look at in week three.

3. 03
   Day 4 — one hour with the whole team

   Not training on the software. Training on the workflow, on real records, with everyone in the room. Then have each person do it once themselves while you watch. The things they hesitate on are your actual configuration bugs, and you will not find them any other way.

4. 04
   Day 5 — go live and turn the old way off

   This is the step people skip and it is the one that matters. If the spreadsheet stays editable, the spreadsheet wins. Make it read-only, announce it, and mean it.

Week two: make it hold

1. 01
   Days 6–8 — watch and fix, do not add

   Every friction report this week is gold and expires fast. Fix the workflow you shipped. Add nothing. The temptation to bolt on a second workflow while momentum is high is the exact mistake that killed the last rollout.

2. 02
   Day 9 — check the number

   Whatever leak you picked, measure it against the week before. Median response time, quotes with zero follow-ups, days-to-invoice. If it did not move, the workflow is wrong and no amount of additional features will rescue it.

3. 03
   Day 10 — add the second workflow

   Now, and only now. Same pattern: build it, one hour with the team, turn the old way off, watch for three days.

What to deliberately not do

• Do not migrate all history. The value is in what happens next, not in nine years of closed jobs nobody will open.
• Do not build custom fields on day one. Use the system's opinions for a month. Half the fields you thought you needed turn out to be habits from the old tool.
• Do not integrate everything immediately. One integration, the one that removes real double entry. The rest can wait until you know how you actually work in the new system.
• Do not run two systems "just in case." This is the most reliable way to end up with neither.
• Do not let one skeptic run a private parallel process. Every business has one. Their objections are usually correct and worth fixing — but in the system, not outside it.

Who owns it

One person, named, with time actually cleared. Not "the team." A rollout owned by everybody is owned by nobody, and it stalls in week two when the season picks up. In a 3–20 person business this is usually the owner or the office manager, and it is roughly a day of real work spread across two weeks — not a project, but not free either.

What good looks like at day 14

1. Two workflows live, both with the old version switched off.
2. The metric you chose has moved measurably.
3. Nobody is maintaining a spreadsheet on the side.
4. A new hire could be shown the daily routine in ten minutes.

That is a successful rollout. Not a fully configured system — a business that has stopped leaking in one specific place and knows how to stop leaking in the next one. Everything after that is the same loop, run again: pick the leak, build the workflow, turn the old way off, check the number.