Early Bird pricing ends January 13. Every ticket is $595 until then. Get your ticket

A three-person panel on the DMSC 2026 main stage
All Topics

Topic 04 of 08

Measurement & Attribution

Proving which marketing dollars sold cars, and cutting the ones that didn't.

5 sessions in 2026 · 11 speakers · 16 partners

Why It Matters Now

Measurement means tying marketing spend to cars sold and repair orders written, not to clicks or leads alone. Franchised dealers spent a record $9.96 billion on advertising in 2025, most of it on digital channels, and each vendor reports results its own way, so the same buyer can be credited more than once. Without a shared yardstick, budget decisions default to whoever sends the most convincing monthly report.

The questions dealers ask are blunt. Which vendor actually sold the car? Is last-click misleading us? Can TV be measured now? What should a sale cost? The answers rest on a few habits: match sold deals in the DMS back to marketing, test big channels with holdouts, own your analytics and ad accounts, and treat any single attribution report as one view rather than the final word.

What You May Hear on Stage

The ideas and changes 2027 sessions on Measurement & Attribution are likely to dig into.

  1. Judging Sources by Cars Sold

    Dealers on industry forums in 2026 are ranking lead sources by cost per sold and shown appointments rather than lead volume. Expect vendors to be asked for deal-level proof, not lead counts.

  2. AI Answers Hide the First Touch

    19% of 2025 buyers used AI sites or AI search overviews, and Google users click fewer links when an AI summary appears. Fewer website sessions no longer mean less demand, so measure leads and sales.

  3. Holdout Tests Check Vendor Credit

    Even large advertisers doubt their attribution: 60-75% of advanced-measurement users told IAB it falls short. Pausing a channel or holding out a group is the plainest way to see what it really adds.

  4. Streaming TV Joins the Sales Report

    Streaming now takes 49.0% of US TV viewing time, and streaming ads can be matched to households and then to DMS sales. Expect sharper questions about consent behind smart-TV viewing data after Texas sued five TV makers in December 2025.

Questions Dealers Are Asking

Asked on dealer forums, in search and on panels this year, answered plainly with the sources.

Questions about the event? See the FAQ
How do we figure out which marketing dollars actually sold cars?

Start from sold deals, not leads. Each month, match DMS deals back to lead sources, tracked calls and ad exposure using a fixed lookback window, then divide each channel's spend by its matched sales. Buyers often touch several sources, so expect overlap and compare the total of vendor claims with your real unit count. Your CRM can also send closed sales back to GA4 and Google Ads when your forms capture a shared ID, so those tools learn from sales instead of form fills.

Sources: support.google.com, support.google.com

A vendor says its ads influenced a big share of our sales. How do we check that?

Ask what counts as influence (an impression, a click or a household on a matched list), the lookback window and the match method. A match shows that a buyer was exposed, not that the ad caused the sale, since many of those buyers were already in market. The cleaner test is a holdout: withhold ads from a comparable group, or pause the channel for a set period, then compare sales.

Sources: support.google.com, iab.com

Should we trust last-click or multi-touch attribution?

Use both and know each one's bias. Last click gives all credit to the final tracked step, which favors search and the last lead source, and a CRM usually records only where the lead came in. GA4 retired its first-click, linear, time-decay and position-based models in November 2023; data-driven attribution remains, with a 90-day default lookback for most key events. Neither proves cause, so check your biggest channels with holdout tests.

Sources: support.google.com, support.google.com

What should we set up in GA4 so the numbers mean something?

Mark real lead actions as key events: form submissions, calls, chats, credit applications and service bookings. Many dealer website providers follow the Automotive Standards Council's shared GA4 event names, which makes data comparable across vendors. Set event data retention to 14 months, the longest option for standard properties, and keep junk signals such as accidental click-to-calls from mobile apps out of the goals your ads optimize toward.

Sources: support.google.com, support.google.com, automotivestandardscouncil.com

Can streaming TV ads really be tied to car sales?

Partly. Streaming took 49.0% of US TV viewing time in July 2026, and streaming ads can be matched to households, usually by IP address, and then to sold deals in the DMS. That shows overlap between exposed homes and buyers, not that the ad caused the sale, so ask for an unexposed holdout group. Smart-TV viewing data is also under legal pressure: Texas sued five TV makers in December 2025 over how they capture what people watch.

Sources: nielsen.com, adexchanger.com, pcworld.com

What's a reasonable advertising cost per car sold?

NADA's 2025 figures give a benchmark: franchised dealers spent $739 on advertising per new vehicle sold, and an estimated $586,246 per dealership for the year. Use it as a sanity check, not a target, since brand, market and used-car volume move the number. For individual sources, track cost per sold and cost per shown appointment, so a weak source is not confused with weak lead handling.

Sources: nada.org, news.dealershipguy.com

Website traffic is down but sales are steady. Should we worry?

Not by itself. Shoppers now get answers before they click: 19% of 2025 buyers used AI sites or AI search overviews, and Pew found Google users clicked a regular result on 8% of visits with an AI summary versus 15% without one. Judge the site on leads, calls, appointments and matched sales rather than sessions, and watch whether searches for your dealership name hold up.

Sources: coxautoinc.com, pewresearch.org

What should a monthly vendor report include, and who should own the accounts?

Ask for spend, leads by type, matched sales with the method and lookback window stated, cost per sold, and a row-level export you can check against the DMS. Have search vendors report spend on your own dealership name separately, since many of those shoppers were already looking for you. Keep the dealership as owner of the Google Ads, Meta and GA4 accounts, with more than one admin, so the history stays when vendors change.

Sources: support.google.com, c-4analytics.com

Take It Back
to the Store.

Ask Your Vendor

  1. When you claim a sale, what counts as your touch (an impression, a click or a lead), and what lookback window and match method do you use?

  2. Will you send the row-level list of matched deals each month so we can check it against our DMS and other vendors' claims?

  3. Will you run a holdout or pause test with us to measure the sales we would not have gotten without you?

Try This Monday

0 of 3 done

Terms Worth Knowing

Last-click attribution
A model that gives all credit for a lead or sale to the final tracked touch before it.
Data-driven attribution
GA4's model that splits credit across touchpoints based on patterns in your own conversion data.
Sales matchback
Matching sold deals in the DMS to the people or households your marketing reached, or who became leads, within a set time window.
Incrementality
The sales that happened only because of an ad, measured by comparing an exposed group with a similar group that was held out.
Cost per sold
Marketing spend divided by vehicles sold over the same period, ideally tracked for each source.
Key event
A GA4 action you mark as important, such as a lead form or click-to-call, which GA4 used to call a conversion.

Go Deeper

Research, tools and conversations from Pasch Group, and the reporting worth your time.

From Pasch Group

Further Reading

Bring Your Questions.

Three days in one room with the dealers solving them and the executives who build the tools.