CRM playbook: import and clean your database

This track turns your CRM from a contact list into the operating system of your practice, and it starts with the least glamorous step: getting everything you have accumulated — a previous CRM export, a brokerage handout, spreadsheets of past clients, the residue of an old lead purchase — into one database you trust. If your list is mostly the people in your phone, the sphere import lesson covers that warmer, more personal path; this lesson is for the whole archive.

Step 1: inventory every source

List where contacts live today: the old CRM’s export, closed transaction files, open-house sheets, lead-vendor downloads, that “contacts final v3” spreadsheet. Export each to its own sheet before merging anything — you want to know, forever, where each person came from. Then apply the entry test: if you cannot place someone and they have never transacted with you, cut the row.

Step 2: clean in the spreadsheet, not the CRM

Fixing data is ten times faster in a sheet than record-by-record after it lands. Four passes: split names into first and last, spelled properly; keep one best email and one best mobile per person; dedupe across sources by email first, then phone; and stamp every row with two tags — the source it came from and the relationship it represents (past client, referral partner, inquiry, unknown). The broader week-one sequence this fits into is laid out in how to set up your CRM database.

Step 3: import in batches you can revert

In the CRM, start an import from Contacts: point it at a CSV, map your columns to fields, and study the preview before committing — a swapped column caught there costs seconds; caught after, hours. Each import commits as a job you can revert, so a bad batch is an undo, not an afternoon of manual deletion. Import one source per batch rather than one giant file: sources differ in quality, and — as the next lesson makes concrete — they differ in what you are legally allowed to send.

Step 4: keep provenance attached

Every batch should leave a trail: which system it came from and what you declared about it at import. Six months from now, “can I email this segment?” and “is this address any good?” are both answered by provenance. A sphere-of-influence database is only as durable as its labels.

What done looks like

One database. Every contact carries a proper name, a working channel, a source tag, and a relationship tag. Your import history reads as a clean series of named jobs, and no row survives that you could not explain to a colleague in one sentence.

The mistake that costs the most

The kitchen-sink import: every CSV dumped in one job, “to sort out later”. Later never comes. You end up with a database you do not trust, so you quietly stop using it — and three months in, you are running your business from spreadsheets again while paying for a CRM. Clean first, batch small, keep the revert option real.