How to Cut Media Waste Without Starving Lead Volume

Any GM can cut a budget in an afternoon. Knowing which half of it was doing nothing takes measurement the store can actually trust.

A five-rooftop Harley-Davidson group was spending real money across several vendors, and every report we opened talked about clicks. Source tracking was configured differently at every store. A test lead we submitted sat overnight. Co-op claims were going in late or not going in at all.

After the rebuild, group lead volume rose roughly 14% and cost per lead fell close to 20%. The strongest rooftop cut monthly media spend 29% and cost per lead 34%, and still took more leads than the month before. Rounded figures, reconciled against the group's CRM reporting and spend logs.

Step 1: decide what the money is supposed to do

Before anyone touches a campaign, write down what the store needs the media to do. Who else is bidding in this market, how far a customer will realistically ride in to buy, and which units have to be gone before the season closes.

A campaign can look efficient by Google's standards and still be the wrong campaign. It pushes units that are already spoken for, buys clicks in counties the store never delivers to, or spends dealer money on work the factory would have reimbursed. No amount of bid tuning fixes any of that.

Step 2: repair the measurement before you optimize anything

Take apart the conversion actions, the forms, the call tracking and the rule the CRM uses to decide something counts as a lead. Then submit real test leads and follow them all the way in. Whatever the ad platform claims it produced, the dealership's own records are the tiebreaker.

Until the source data agrees with itself, nobody can say which campaigns deserve protection and which ones should die. In the five-rooftop group, unattributed-lead rates ran from under 1% at one store to more than 13% at another before we standardized the mapping.

Step 3: give every campaign one job

A campaign should have a purpose you can say out loud in a sentence: catch the people already shopping this store, create demand that did not exist, or move the category the floor is long on. Campaigns without a job start overlapping, bidding against each other and splitting one customer's budget three ways.

Google's search terms report shows what people actually typed to trigger the ad. Read it against what the campaign was built to do and against what is sitting on the floor. Performance Max needs the same supervision as everything else: exclusions, honest conversion actions, and a clear answer to what you asked it to optimize toward. Left alone it finds the cheapest conversions in the account and reports them as a win.

Step 4: separate demand capture from demand creation

Somebody typing the dealership's name into Google has already picked the store. Somebody typing "used Road Glide near me" has not. Report those two apart from each other, and resist both easy conclusions — that brand traffic is money you are lighting on fire, or that non-brand is automatically new business. Run them separately for a stretch and read what the CRM says.

The point is to learn the real mix, not to hand every sale to whichever ad the customer happened to touch last.

Step 5: walk the path the customer walks

The ad, the unit page, the form and the phone number are one chain, and the chain is worth what its weakest link is worth. Targeting that lands a qualified shopper on a generic homepage, or on a vehicle page that will not load on a phone, is a click you paid for and threw away.

Test it the way a customer would, on a phone: filter the inventory, open a unit, submit the form, call the number at nine at night and see who picks up. Fix the biggest leak before you pay to pour more traffic into it.

Step 6: protect the lead after it lands

The test lead that sat overnight during the group audit cost that store a customer no bid adjustment was ever going to buy back. The rebuild tied source mapping, routing and follow-up together across all five rooftops, so what we bought and what the store did with it got graded on the same page.

Pied Piper's 2026 Internet Lead Effectiveness study found only 47% of powersports dealers answered the question a shopper submitted through the website by email or text. That number belongs in the media review, because a lead nobody answers is worth a different amount than a lead somebody works.

Step 7: put a weekly decision on the calendar

Once a week, look at spend pacing, search terms, source accuracy and whatever the vendors still owe you. Change one thing at a time and write down what you changed and why you changed it. By month end the report should read as a record of decisions and what they did, not two columns of totals sitting next to each other.

If the store already has a marketing manager

A capable in-house manager can run this cadence, given the right access and a few weeks of training. We audit the accounts, rebuild the structure and build the weekly scorecard, then teach the manager how to read a search-terms report and when to escalate instead of guessing. After that we stay on as the technical reviewer and the store keeps the keys.

Where there is nobody to hand it to, we run the media and the operating layer ourselves. Either way leadership ends up with one issue list and one standard of proof.

Leadership checklist

• Does every campaign have one documented commercial purpose?
• Are OEM-funded and dealer-led budgets separated?
• Do conversion actions match real calls, forms and CRM leads?
• Are brand navigation and new-demand campaigns reported separately?
• Does campaign emphasis match current inventory and season?
• Has every high-value landing page and lead route been tested?
• Does the weekly review end with recorded decisions and owners?