DMS Data Cleanup Before Migration

Every duplicate, dead number and lost opt out you carry across becomes somebody else's problem for the next seven years. Fix it in the file.

There is a moment in every system change when your entire customer database exists as a set of files on a desk rather than as a live system. That moment is the only cheap opportunity you will get to fix it. Once the data is loaded and your team is working in the new system, every correction is a live edit competing with real work, and the appetite for a cleanup project evaporates within about two weeks of going live.

Dealers know this and skip it anyway, usually because the migration timeline is already tight and cleanup feels optional next to getting the doors open on the new system. It is not optional. Whatever you carry across, you keep. A duplicate customer record created a decade ago will still be generating confusion a decade from now, because nobody ever goes back.

The cost of skipping it is not obvious on day one either, which is part of the trap. It shows up later as two salespeople calling the same customer, as a mailer going to a household three times, as a report that undercounts repeat buyers, and eventually as a compliance problem when somebody gets texted who told you two years ago to stop.

Budget the time explicitly. On a mid sized store, a competent cleanup pass is a week or two of somebody's focused attention, and it is the highest return work in the entire project.

Before you clean anything, decide what belongs in the new system at all. Most stores over migrate out of anxiety and then complain that the new system is cluttered.

Work through the categories separately. Customer and contact records almost always move. Vehicle and deal history usually moves in summary form. Accounting history generally stays behind in the old system or in an archive, because reconciled financial records do not belong in a fresh general ledger and your accountant will tell you the same thing. Service history depends on how far back is useful and what your fixed ops team will actually reference.

Then set a cutoff. A customer who last transacted in 2009 with a disconnected phone and a returned mail flag is not a marketing asset, and moving that record costs you storage, search noise and duplicate risk forever. Many stores draw the line at seven to ten years for full records and archive everything older as a read only extract.

The archive decision deserves its own conversation, and migrating historical records to a new DMS covers what to bring live, what to archive and how long to keep read only access to the old system.

Every dealership database has duplicates and the count is always higher than management expects. They come from the same places every time: a customer who bought once and serviced under a different spelling, a lead that came in twice through two sources, a household where two people share a phone number, and years of salespeople creating a new record because searching felt slower than typing.

Match on more than name. Name matching alone will merge two different people named Maria Garcia in the same ZIP and miss the same person entered as Bob and Robert. The reliable combinations are normalized phone number, email address, and address plus last name. Run each pass separately and review the results rather than merging automatically.

Decide your survivorship rules before you merge anything, and write them down. Which record wins on conflicting addresses, the most recent or the one attached to the most recent transaction? Which email survives? What happens to notes and history from the losing record, and are you certain they are being appended rather than discarded? Merges are hard to undo, so run them on a copy first and check a sample by hand.

Households are worth handling deliberately too. Two people at one address who both buy from you are two customers with one mailing relationship, and treating them as duplicates is its own kind of error.

Contact data is where cleanup pays back fastest, because it directly determines whether your campaigns reach anyone.

Phone numbers. Normalize to one format and hold the line. Our own rule inside the CRM is that a phone is stored as a valid ten digit US number in a single consistent format or it is stored blank, because a field containing 555 1212 ext 4, please call after 6 is not a phone number, it is a note pretending to be one. Strip extensions into their own field or into notes. Flag obvious placeholders such as 000 000 0000 and 111 111 1111, which exist in every dealership database because they satisfy a required field.

Email addresses. Remove the joke addresses and the placeholders that staff typed to get past validation. Fix the obvious domain typos. Then run the surviving list through validation before the first send rather than after, because a bulk send into a stale list is how a store damages its own sending reputation for months.

Addresses. Standardize to postal format and keep the returned mail flags you already have. A record marked undeliverable three years ago is information, and losing that flag in migration means paying to mail it again.

The CRM data cleanup guide goes deeper on the marketing side of the same work.

This is the part of cleanup that is a legal exposure rather than a tidiness issue, and it is the part most likely to be handled badly because opt out data often lives somewhere other than the customer record.

Find every place a do not contact instruction is stored. Email unsubscribes. Text stop requests. Do not call entries. Written requests scanned into a folder. Verbal requests noted by a receptionist. Then consolidate them into one authoritative list with dates, and confirm the new system can accept that list before you load a single customer record.

The order matters. Load the suppression list first, then load customers. If you do it the other way around, there is a window where your new system holds contactable records with no suppression applied, and all it takes is one campaign going out during that window.

Keep the dates. An opt out is a record of when a consumer told you to stop, and you may need to prove it. A migration that preserves who but not when has lost half the value.

And check that opt out applies across channels the way you intend. A customer who unsubscribed from email has not necessarily revoked text consent, but a customer who sent STOP absolutely has. Getting that distinction wrong in either direction causes trouble, and TCPA compliance for dealership texting covers the texting side specifically.

Vehicle and deal data has its own failure modes, and they are quieter than contact data problems because nobody notices a wrong number until they need it.

Check VINs for length and character validity. A seventeen character VIN with an O in it is wrong, because VINs do not contain the letter O, and that record will never match anything again. Look for VINs that appear on multiple open records, which usually means a unit was entered twice.

Check mileage fields for impossible values, and check dates for the classic errors: a delivery date before the sale date, a two digit year that landed in the wrong century, and records with a date field of 1900 or 1970 that mean the field was actually empty.

On money fields, confirm the format is consistent and that nothing is stored as text with a currency symbol in it. Mixed formats in a numeric field are how a report ends up silently excluding rows, and by the time somebody notices the total is wrong, the migration is a distant memory.

Reconcile counts before and after. Open deals, units in stock, active customers. If the numbers do not match on both sides of the move, find out why before you go live rather than accepting a close approximation.

Orphaned records are the ones assigned to somebody who left. In a store with normal turnover, a five year old database can easily have a quarter of its records owned by people who no longer work there, and those records are invisible in practice because nobody's task list includes them.

Build the user list first. Every person who ever had a login, whether they are active, and who inherits their records. Then reassign deliberately rather than dumping everything on the sales manager, which is functionally the same as deleting it.

While you are in there, use the moment for an access review. Anyone who left and still has a working login is a real risk, and system changes are exactly when those accounts get discovered, usually by accident. Decide the role and permission model for the new system now rather than retrofitting it onto two hundred live users later. Our CRM security checklist covers the access questions in order.

Keep the historical assignment data even as you reassign forward. Knowing that a customer originally came from a salesperson who left in 2022 is useful context, and it is not the same thing as leaving the record assigned to them.

Never clean production data directly. Export, clean the export, and load the cleaned file. That way a bad merge rule costs you an afternoon instead of a customer database.

Keep an untouched original of every export, stored separately and left alone. Keep a log of every transformation you apply, in enough detail that somebody else could repeat it. When a manager asks in three months why a record looks the way it does, that log is the only answer you will have.

Then validate before the real load. Run a subset into a test environment and have two or three people who know the customers well go look at records they recognize. They will find things a script cannot: the customer who is now in the system twice under a married name, the note that got truncated mid sentence, the vehicle attached to the wrong household.

Treat the export files themselves as sensitive. They are your entire customer database as downloadable files, sitting outside your normal controls during the exact period when the most people have their hands on them. Encrypt them, keep them off personal devices and consumer file sharing, limit who holds a copy, and delete the working copies when the project closes.

We should be clear about our position, because this page is about a system we do not sell. LeadLocate is not a dealer management system. There is no general ledger, no accounts payable or receivable, no payroll, no deal posting to accounting, no parts, no repair orders and no title work. If you are changing the system that runs your accounting and your shop, evaluate that on its own merits and do not let any CRM vendor tell you otherwise.

What we do sit alongside it. The CRM handles lead management, communication, follow up, desking and reporting, and it runs independently of whichever system your store operates on. No DMS integration and no inventory feed is required, which turns out to matter enormously during a conversion: your sales and follow up side stays stable while the back end is being replaced, instead of both going down at once.

The cleanup tools are real and worth naming. The phone validator and the email validator check contact data before you send anything. Contact import brings records in, blacklist import brings your do not contact list in first, and the customer profile stores personally identifiable information encrypted with access controlled per user rather than open to anyone with a login.

CRM Only is $199 a month if that is all you need, month to month with no long term contract. Details on the pricing page, and contact us if you would rather just ask whether we fit alongside what you are moving to.

How long should data cleanup take before a migration?

On a mid sized store, plan a week or two of focused attention from someone who knows the data. Compressing it into cutover week is how stores end up carrying duplicates and lost opt outs into a system they will run for the next decade.

What should we clean first if time is short?

Opt out and consent records, because they are a legal exposure. Then duplicates, then phone and email formatting. Vehicle and deal field validation comes after those three if the calendar allows.

Should we clean the data in the old system or in the export file?

In the export file, on a copy, with the original preserved untouched. Cleaning production directly means a bad merge rule damages live records your team is working in, and merges are difficult to reverse.

How far back should we bring records?

Most stores bring seven to ten years of full customer records live and archive anything older as a read only extract. Accounting history usually stays in the old system or an archive rather than loading into a fresh general ledger.

Does LeadLocate sell a dealer management system?

No. There is no general ledger, accounts payable or receivable, payroll, deal posting, parts, repair orders or title work. We provide the lead, communication, follow up and desking layer that runs alongside whichever system your store uses.

Do we need a DMS integration for the CRM side to work during a conversion?

No, and that is genuinely useful during a change. Neither a DMS integration nor an inventory feed is required, so your sales and follow up side keeps running while the back end is being replaced.

The CRM, the leads and the follow up run independently of your DMS, so a conversion does not take your sales floor down with it. Month to month, no long term contract.