The New School of Sales: How to Measure the ROI of Dealership Technology

When this post first ran in January 2020, its message was that the "new school of sales" runs on tools, because shoppers arrive having done their homework online. That is still true. What has changed is the number of tools on offer, and every one of them promises more sales and sends an invoice.

This guide is for the dealer principal, GM or sales manager deciding whether a tool pays for itself: what to do before you buy and after you go live, with a worked example.

If you are still deciding which tools your store needs at all, start with the guide to planning a powersports dealership tech stack. This article assumes you have a tool in mind and want to know whether it works.

The method in five steps

1. Baseline the metric the tool is supposed to move, before anyone touches the tool.
2. Count the full cost: license, setup, integration, training time and the staff time to run it.
3. Run a pilot with a comparison group that keeps working the old way over the same weeks.
4. Check for attribution traps before you credit the tool with anything.
5. Review at 90 days and apply a keep, extend or cancel rule you wrote down before you signed.

Step 1: Baseline the one metric it should move

Every tool is bought to change something. Write that something down as a number, and pick one primary metric, not five. A texting tool should move the share of leads that reach a two-way conversation. An inventory merchandising tool should move how fast units go from arrival to online. A desking tool should move how long a customer waits for numbers.

Then measure that number for the last 90 days, using the same definition you will use later. As one dealer trade article put it, "To measure ROI, you have to know where you are now," and it advised comparing performance at 30, 60 and 90 days, noting that "Not all improvements will appear in the same time frame."

A good primary metric has three traits:

• It comes from a system, not a memory. Your CRM, phone system or inventory platform can report it without anyone counting by hand.
• It moves within 90 days. Repeat purchases take years to show up; appointment rates do not.
• It sits close to money. Logins, messages sent and tasks completed measure activity. As one CRM vendor put it plainly, "They measure activity, not value."

Step 2: Count the full cost

The license fee is the number on the proposal, but it is rarely the whole cost. Count every line for the first year.

Loaded hourly cost means pay plus benefits and payroll costs, not the wage alone. U.S. Bureau of Labor Statistics data show that for private industry workers in June 2026, wages and salaries were 70.0 percent of total compensation; benefits made up the rest. A simple way to estimate a loaded rate is to divide the hourly wage by 0.70. A $24.50 hourly wage becomes $35.00 an hour ($24.50 / 0.70).

Staff time is the line most stores forget, and it often turns a cheap tool into an expensive one.

Step 3: Run a pilot with a comparison group

A before-and-after comparison is the weakest test a dealer can run, because powersports is seasonal. A tool switched on in March will look like a hero by May, and a tool switched on in October will look like a failure by December, whatever it actually does.

The fix is to run the old way and the new way side by side, over the same weeks:

• Split by lead, not by person, when you can. Alternate internet leads between the new process and the old one (for example by odd and even lead number). Both groups then see the same season, the same inventory and the same ads.
• Split by team or rooftop when you must. If the tool cannot be split by lead, pilot it with one team or one store and compare against a similar one. Note any differences between them before you start.
• Fix the rules before day one. Write down the primary metric, the definition, the pilot length and the decision rule. Do not change them halfway.
• Ask the vendor for pilot terms. A short pilot, a reduced first-term price or an exit clause tied to your decision date is a reasonable request. Check the notice window on any auto-renewal.

Ninety days is a sensible pilot for most sales tools: long enough to cover a full sales cycle for most leads, short enough to end before a renewal.

The worked example: a lead follow-up tool

All numbers below are hypothetical. Replace them with your own.

The store: 200 internet leads a month (2,400 a year). Average gross profit after sales commission: $1,200 per unit.

First-year cost:

Break-even, set before signing: $9,880 / $1,200 = 8.2, so the tool must produce 9 extra units a year. On 2,400 leads, that is a lift in close rate of 0.375 percentage points (9 / 2,400).

The 90-day pilot: 600 leads, alternated between groups.

If the sales lift is real: 1 extra point of close rate x 2,400 leads = 24 extra units a year. 24 x $1,200 = $28,800 of extra gross. First-year ROI = ($28,800 - $9,880) / $9,880 = 191%. In year two, with only the license and running time ($6,840), ROI would be ($28,800 - $6,840) / $6,840 = 321%.

Now the honest part. Three extra sales out of 300 leads is a small difference, and a standard two-proportion test puts it well inside normal chance variation (a z-score of about 0.44, where about 2 is the usual bar). The appointment lift, 24 extra appointments on the same leads, scores about 2.16, so that one is probably real. The sales lift is not proven yet.

The decision: the tool clearly moved appointments, and break-even needs only 9 units a year, but the sales evidence is thin. Under the rule below, this is an extend: keep the split running for another 90 days, with the cancel date booked before the renewal notice window.

Attribution traps that flatter a tool

Before you credit any tool with a result, check for these.

• The attention effect. People work harder on whatever the GM is watching, so some early lift comes from attention, not software.
• Funnel multiplication. Vendor models often multiply an appointment lift by your old close rate. In the example, 24 extra appointments at the control's appointment-to-sale rate (24 / 75 = 32%) would predict about 8 extra sales. The pilot produced 3. Extra appointments are often weaker ones, so count sales, not projected sales.
• Double counting. If a new ad campaign, a price drop and a new tool all start in the same month, each vendor will claim the sales. Credit only the difference against your control group.
• Last-click credit. Website and ad reports usually give all the credit to one touch. Google's own Analytics help page explains that its "Paid and organic last click" model gives 100% of the conversion value to the last channel clicked. A tool that touches the customer last looks better than it is.

Which metric each tool should move

For the tools themselves and what good looks like for each, see our list of sales tools for powersports salespeople. For where the technology is heading, see powersports dealership technology trends. If the tool under review is a desking system, our desking software checklist covers what to test during the pilot.

What most ROI guides leave out

We read the pages that rank for this question before writing. Most are vendor articles, and they share three gaps.

No comparison group. Almost every guide compares "before" with "after", which in a seasonal business mostly measures the weather. A control group over the same weeks is the biggest single improvement a dealer can make.

No allowance for small numbers. A single rooftop rarely has enough sales in 90 days to prove a modest close-rate lift, as the worked example shows. Pick a primary metric that moves faster, and extend when the sales signal is thin.

No cancel date. Guides explain how to calculate ROI, but rarely say what to do with the answer. Without a decision date that sits before the renewal notice window, an underperforming tool simply renews.

The 90-day review and keep-or-cancel rule

Book the review on the day you sign. Bring the pilot table, the cost table and the adoption numbers.

First, check adoption. If fewer than about 8 in 10 of the intended users used the tool as designed, the pilot did not test the tool; it tested the rollout. Fix training, rerun for 30 days, and review again once. This threshold is our recommendation, not an industry standard.

Then apply the rule:

Write the decision and the numbers on one page and file it with the contract.

This is general information, not financial or legal advice. Check contract terms with your own advisers.

Where PSX Ascend fits

Measure us the same way. The AI-powered powersports CRM is built around clear stages, ownership and next actions, and helps managers "manage the performance signals that lead to a sale". PSX records and tags every inbound call, and outbound calls where state law permits, with recordings in the customer record and a dashboard view for manager review. ALMA, the Automated Lead Management Assistant, re-engages stalled prospects and brings them back to a live conversation with a person; it makes no AI voice calls.

On the inventory side, PSX Ascend inventory management lets teams "see which units are fresh, gaining attention, or need a pricing and merchandising decision", which supports the days-to-online and aging metrics in the table.

You can also see the same AI-powered powersports CRM on our sister site, PSXDigital's AI-powered powersports CRM page.

Sources

• U.S. Bureau of Labor Statistics, Data series CMU2020000000000P: Wages and salaries as a percent of total compensation for private industry workers (Employer Costs for Employee Compensation; 70.0 percent in the second quarter of 2026, the June 2026 reference period)
• Google Analytics Help, Get started with attribution (the "Paid and organic last click" model gives 100% of the value to the last channel clicked)
• Auto Dealer Today, Measuring Return On Investments In Customer Relationship Management (2006; know your starting point, compare at 30, 60 and 90 days)
• Fullpath, Automotive CRM ROI: How to Actually Measure Success (vendor article, March 2026; activity versus value)