Running Two CRMs During a Dealership Migration

Dual running is the safest way to change systems and the fastest way to lose leads. The difference is one rule, applied without exceptions.

Every dealership that has cut a CRM over on a Friday night with no fallback has a story about the following Tuesday. A lead source that was never repointed. A phone number that still forwarded to a dead extension. Three days of internet leads sitting in a system nobody was logging into anymore.

Paying two vendors for four to six weeks feels wasteful right up until it saves a month of lead flow. Put a number on it. If your store works two hundred internet leads a month and a bad cutover costs you a week of them, the overlap subscription is the smaller number by a wide margin. That is the whole argument, and it usually ends the debate with a controller faster than any risk narrative.

The reason dual running gets a bad reputation is not the cost. It is that most stores do it without rules, so leads get worked twice, notes end up split across two systems, and the reporting becomes unusable in both. That is a discipline failure rather than an argument against overlap, and it is fixable with one decision made up front.

This is the rule that everything else hangs from. For any given lead, exactly one system is the place it gets worked, and everyone in the building can say which one without checking.

The failure mode when you skip this is specific and predictable. A lead exists in both systems. A salesperson works it in the new one and logs a call. The BDC sees it untouched in the old one and calls the same customer forty minutes later. The customer now thinks your store is disorganized, which they are correct about, and one of those two people stops trusting the software. Once trust goes, salespeople go back to their personal phones and your activity data becomes fiction.

Say it out loud in the sales meeting, write it on the board, and repeat it every week of the overlap. Nobody works a lead out of the non authoritative system, even if they are certain it was missed. If they think it was missed, they tell the pilot owner, who checks. That single habit prevents most of what goes wrong.

There are two ways to divide the flow and only one of them survives contact with a real store.

Splitting by person sounds fair: half the team on the new system, half on the old. It falls apart the moment a customer calls in and gets whoever answers, which is most of the time. Now the lead exists on both sides, owned by two people using two systems, and the rule above is already broken.

Splitting by source works. Route a specific lead source, or a specific set of sources, entirely into the new system, and leave everything else on the old one. The routing happens upstream at the source itself rather than inside either CRM, so there is no ambiguity about where anything landed. Walk in traffic and phone ups go wherever your ADF and your phone forwarding point, so decide those deliberately rather than letting them default.

The cleanest version of a source split is one whole team taking one whole set of sources, so the internal handoffs stay inside one system. There is more on configuring the routing itself on the lead routing rules page, and on round robin distribution if that is how your floor works.

Four places, in rough order of how often they bite.

Unrepointed sources. Every third party that posts leads to you has your old destination stored somewhere. Third party lead providers, your website forms, chat, ADF feeds, marketplace listings. Make the complete list before you start rather than discovering it one vendor at a time over three weeks.

Phone. Tracking numbers, forwarding rules, after hours handling and voicemail boxes are the most commonly forgotten piece and the most expensive one, because a missed call is a customer who was ready to talk right then. Missed call text back closes some of that gap automatically by turning an unanswered call into a live text thread.

Automations still running on the old side. A drip sequence that keeps firing out of the system you are leaving will contact customers your new system thinks it owns. Inventory those before cutover and turn them off deliberately.

The gap between business days. Cut over midweek in the morning, never on a Friday afternoon. If something is wrong you want a full staffed day to find it, not a weekend of unanswered leads and a Monday of apologies.

Duplicate records are the visible symptom. The one that costs you real money is duplicate contact.

Start with opt outs, because this is the piece that carries actual exposure. A customer who unsubscribed from texts or email in the old system has opted out, full stop, and the new system needs to know that before it sends anything. Export the suppression list first and import it as a blacklist on the new side, and do it before the first campaign rather than after. Blacklist import exists precisely for this. The rules are summarized on TCPA compliance for dealership texting, and treating them as a migration step rather than a legal afterthought is the correct instinct.

Then handle the ordinary duplicates. Decide a match rule before you import, usually mobile number first, then email, then name plus address. Normalize phone formats on both sides before you compare, because a list stored as one long string and a list stored with separators will match nothing and you will conclude, wrongly, that you have no duplicates at all. See duplicate lead management for the ongoing version of this problem once you are live.

The order matters, because each step should be reversible until the one after it is proven.

1. Inbound lead sources first, one at a time. Repoint a source, confirm a live lead arrives correctly formatted with the right source label, then move to the next. Do not batch this.
2. Website forms and chat. Test each form yourself with a real submission and watch where it lands, including the confirmation the customer receives.
3. Phone. Tracking numbers, forwarding, IVR trees and after hours routing. Call every number from an outside line, including the ones on printed material nobody has updated since 2019.
4. Outbound automation. Rebuild follow up processes and drips on the new side, then disable the old ones the same day so nothing runs twice.
5. Reporting last. It is the only piece that can wait, and trying to do it first delays everything that actually touches a customer.

If the migration also involves a system change on the accounting side, sequence them apart rather than together. The parallel run guidance for a dealer management system conversion covers that, and doing both at once is how a store loses a quarter.

For the duration of the overlap your numbers live in two places, and pretending otherwise produces bad decisions.

Report the pilot slice on its own, against the same slice's history from the old system, rather than reporting a store total that mixes both. Comparing a partial month on the new system to a full month on the old one makes the new system look terrible and tells you nothing.

Three measures are enough during an overlap: median time to first response, follow up touches per lead, and appointments set. Those are readable in weeks. Gross and closing rate are not, because the deals delivering this month came from work done before the change.

One advantage worth naming. Because texting and calling happen inside the platform rather than on personal cell phones, and call recording with transcription means a manager can read a twelve minute call in half a minute, the activity data on the new side is real rather than self reported. That usually makes the new numbers look worse at first, since the old ones were flattering. Expect it and explain it in advance, or somebody will use it as an argument to abandon the migration in week two.

Four to six weeks covers most single store migrations. Groups roll store by store, so each individual rooftop still gets its own four to six weeks rather than the whole group carrying a six month double bill.

Set the end date before you start and put it in writing. Open ended dual running is how stores arrive eighteen months later with two half maintained systems, notes split between them and no clean history in either. Once you drift past the plan, the second system stops being insurance and becomes a liability.

The end sequence, in order: repoint the last remaining sources, pull every export you will ever want from the old system while you still have access, confirm the exports actually open and contain what they claim, then give notice. Not the other way around. Your leverage and your access are both highest before you send the cancellation, which is the argument the data export checklist is built around.

Check the notice period in your current agreement too. Plenty of automotive contracts require notice well ahead of renewal, and a store that plans a six week overlap around a thirty day assumption can end up paying for a year it did not want.

We price per store, month to month, with no long term contract, which means an overlap costs you one or two extra months and nothing else. CRM Only starts at $199 a month if you have your own lead sources. Plans that include exclusive local leads start at $799, and combined buyer and seller programs from $1,599. Full detail on the pricing page.

There is no setup fee structure that punishes a short trial and no minimum term that makes stopping expensive. That is deliberate. A vendor whose commercial model depends on locking you in before you have seen it work has told you how confident they are about month four.

Most stores get here by piloting first. Run one rooftop or one team for three to four weeks, fix the process problems the pilot exposes, then use the overlap plan above to move the rest. The sequencing is on how to pilot a new CRM at one dealership, and the full step by step is in the CRM migration checklist. We cannot guarantee a painless migration, but we can tell you where they usually break.

How long should we run two CRMs at once?

Four to six weeks for a single store. Groups roll rooftop by rooftop, so each store gets its own overlap rather than the whole group carrying a long double bill. Set the end date before you start.

Should we split the team or split the lead sources?

Split by source. Splitting by person breaks the moment a customer calls in and gets whoever answers, which puts the same lead in both systems with two owners. Source splits are decided upstream and are unambiguous.

What is the single biggest mistake during dual running?

Letting both systems be authoritative for the same lead. That produces double contact, split notes and salespeople who stop trusting either system. One system per lead, stated out loud and repeated weekly.

How do we avoid texting a customer who already opted out?

Export the suppression list from the old system and import it as a blacklist on the new side before the first campaign goes out. Opt out status has to travel with the migration, and it is a compliance issue rather than a convenience.

When should we give notice on the old vendor?

After every source is repointed and after you have pulled and opened every export you will want. Check the notice period in your agreement early, because many automotive contracts require notice well ahead of renewal.

Will our numbers get worse during the overlap?

The reported numbers often do, because activity captured inside a platform is real while activity self reported from personal phones was flattering. Explain that before it happens, or somebody will read it as evidence to abandon the change.

We will map the lead sources, set the authoritative rule with your managers, and put an end date on the calendar. Month to month, no long term contract.