Before any system can draft a quote, book a job or chase a customer, it needs one true copy of the facts it works from. Customers, prices, availability, job status. When three apps each hold a different version, the system either stops or guesses, and a guess sent to a customer is worse than no reply.
That is why the first useful build is so often a data job, and not the exciting one.
Why does it stall after the demo?
The demo runs on clean sample data. Your business runs on the real thing. The customer's phone number is in the job app, their email is in the inbox, and the latest price is in someone's head. Pointed at that, a system has nothing solid to stand on.
The fix is dull, and it works.
How do you fix the records?
Four moves:
- One home for each kind of record. Decide which app holds the true copy of contacts, which holds prices, which holds job status. Every other app reads from that home.
- Connect the apps a system needs to act. The diary, the inbox, the job and accounting software, so a change in one place shows up everywhere it should.
- Give access on purpose. A system reads what its job needs, and changes as little as it can. Reading a customer's history is low risk, but sending an invoice is not.
- Write the few rules down. Your pricing, your booking policy, how you speak to customers. Rules in someone's head cannot be followed by anything else.
What does it cost to leave it messy?
Time, first: every retyped detail is a few minutes, many times a day. Then mistakes: the quote sent to the old email address, the booking made against a diary that was wrong. Then trust, when a customer gets two different answers from the same business.
What goes wrong?
Trying to clean everything at once. Pick the records one job needs, fix those, and let the next job pull the next set in.
And cleaning once. Records drift back to messy the moment two apps disagree again, so the fix has to include which app wins.
One step for this week
Write a two-column list: each kind of record your business keeps, and the one app that should hold the true copy. Where the answer is "a few of them", that is your first fix. When to buy and when to build covers the choice between a tool and a build.
Integration and data sync is one of the eight things we build. See it in section 02.


