Start 90 days before the renewal date, not 90 hours

Most dealership software renewals happen the same way. An email lands three weeks out, the rep offers a "loyalty discount" if you sign by Friday, and you sign because nobody has time to prove the tool isn't working.

Then you find out in month four that two departments are paying for the same texting function, the equity mining tool hasn't been logged into since March, and your CRM never got the DMS write-back turned on — so your BDC has been typing deal numbers by hand for a year.

An audit fixes that, but only if you do it before you're under time pressure. Pull every auto-renew date into one list, sort by date, and start work on each one 90 days out. That's enough runway to cancel inside most notice windows and enough time to actually watch people use the thing.

Here's the audit itself. Six columns. One row per platform. No exceptions for "that one's just included."

Build the inventory first, and build it from invoices

Don't build your list from memory. Build it from the accounting payables detail for the last 12 months, filtered to anything that smells like software, plus your OEM statement, plus anything billed through your DMS vendor as a line item.

You will find things nobody remembers buying. A lot vendor that still bills $340 a month for photo hosting. A dead chat tool. A reputation platform your old GM signed that nobody's logged into.

Three rules for the inventory:

  • Include the free ones. OEM-provided tools still cost you training time, duplicate entry, and data leakage. They go on the list.
  • Include per-seat tools. Twelve seats at $89 when you have eight salespeople is real money nobody is watching.
  • Note the renewal date and the notice window in the same row. "Auto-renews 4/1, 60-day notice" tells you the real decision date is 1/31.

A 15-rooftop group will land somewhere north of 30 platforms. A single store usually lands between 12 and 25. If your dealership technology stack list is shorter than that, you missed some.

Column 1: The workflow it serves

Write one sentence, in plain language, describing the job this platform does. Not the marketing category. The job.

Bad: "AI-powered customer engagement platform."

Good: "Sends the first text to an internet lead within two minutes, 24/7, and routes a reply to the on-duty BDC rep."

If you can't write that sentence without looking at the vendor's website, that's your first finding. A tool nobody can describe is a tool nobody is running.

When you finish all the rows, sort by workflow. Now you can see the overlaps. You'll probably find four tools that all claim "follow-up," three that send texts, and two that pull soft credit.

Column 2: The named owner

One human name per platform. Not a department. Not "the BDC." A person who owns adoption, owns the vendor relationship, and gets the call when it breaks.

Then verify the ownership is real by asking the owner three questions:

"When was your last call with this vendor, and what was it about?"

"What report do you pull from this, and how often?"

"If I turned it off Monday, what breaks first?"

If the answers are "I don't know," "I don't," and "nothing," you've found a renewal candidate for cancellation — not because the software is bad, but because nobody is driving it.

I'd rather own six tools well than twenty tools badly. A dealer software audit usually ends with fewer platforms and more results, because attention is the scarce resource, not features.

Column 3: Utilization evidence

This is the column that gets skipped and the column that matters most.

"Evidence" means a number from the admin panel, not an impression. Every vendor can show you one. If they won't, that's data.

What to ask for, by tool type:

  • CRM: logins per user per week, tasks completed vs. tasks assigned, percent of leads with a logged call attempt in the first hour.
  • Texting/chat: conversations started, conversations that got a human reply, average first-response time.
  • Equity mining: alerts generated vs. alerts worked vs. appointments set from those alerts.
  • Desking/menu: deals desked in the tool vs. total deals delivered. If it's 40%, half your store is desking on paper.
  • Inventory pricing: price changes accepted vs. recommended.
  • Video tools: videos sent per rep per week, open rate.

Then pick the number that ties to money and write it down. Example: "Equity mining generated 1,140 alerts last quarter. 212 were worked. 19 appointments. 6 deliveries. Cost: $1,500/mo = $4,500/quarter = $750 per delivery."

That's a defensible number either way. Six deliveries at $750 acquisition cost may be a bargain. But now you're arguing about a number instead of a feeling.

The "two reps" test

Before the renewal meeting, sit with two reps — your best and your most average — and ask each to do the core workflow in front of you.

Watch for hesitation. Watch for the alt-tab. Watch for the sticky note. If your best rep can't find the function in under ten seconds, your average rep never uses it.

Column 4: Duplicate entry

Walk one deal end to end and count every time a human retypes information that already exists somewhere in the building.

A typical path: lead comes in → CRM → desking tool → DMS → F&I menu → lender portal → delivery checklist → service appointment. Customer name, phone, email, VIN, and trade details can get typed five or six times on one deal.

For each duplicate, write down who does the typing and roughly how long it takes. Two minutes per deal across 110 deliveries a month is about three and a half hours of pure retyping — and more importantly, it's where phone numbers get transposed and the follow-up dies.

Duplicate entry is almost always a dealership CRM integration gap that was never turned on, not a technical impossibility. Which brings us to the next column.

Column 5: Integration gaps

For each pair of connected tools, record three things: does data flow, which direction, and how fast.

The common gaps in automotive retail technology:

  • CRM ↔ DMS one-way. Deals push from CRM to DMS but delivered-deal data doesn't come back, so your reports never reconcile.
  • Service ↔ Sales invisible. Your sales team can't see that the customer has a service appointment Thursday. That's your easiest equity conversation, and you're missing it.
  • Lead source tagging lost in transit. Everything shows up as "Internet," so your cost-per-sale by source is fiction.
  • Texting outside the CRM. Conversations live in a separate app, so when a rep quits, the thread goes with them.

Ask each vendor one question before renewal: "What's the exact scope of your integration with [other tool], and what's included at my current tier?" Get it in writing. "It integrates" is not an answer.

Column 6: Overlapping cost

Now total the spend by workflow instead of by vendor.

Write it as: Follow-up texting — $1,890/mo across 3 platforms. Inventory merchandising — $2,400/mo across 2.

Then per-unit it. Total software spend ÷ units sold per month. Most dealers have never calculated that number and are surprised by it. Whatever yours is, it's now a thing you can manage.

Turn the grid into three decisions

Every row ends in one of three buckets, with a name and a date:

  1. Renew as-is. Owner is active, utilization is real, no overlap.
  2. Renegotiate. Usually seat count, tier, or term length. Walk in with your utilization numbers: "We're paying for 14 seats and nine people logged in last month. Let's true that up."
  3. Consolidate or cancel. Pick which tool wins the workflow, set a kill date, and assign the migration.

For anything you keep but that scored badly on utilization, set a 60-day adoption target with a specific number — "desking tool used on 90% of delivered deals by March 31" — and put it on a manager's one-on-one agenda. If it doesn't move, it's a cancel next cycle and you already have the evidence.

The whole audit takes a focused afternoon for the inventory and about two weeks of watching to fill in utilization. That's cheap compared to one more year of paying for a platform nobody opens.

And if you want a shortcut on the utilization column, the recordings and email threads you already have will tell you which tools actually show up in real conversations — that's the kind of thing MoreSignal surfaces when it scores calls against your process.

Build the grid once. After that, each renewal is a 20-minute conversation with a number attached.