Cut paid spend 18%, appointments still grew
Multi-state production homebuilder
Took over a fragmented Google Ads and Meta program. Rebuilt over two hundred campaigns, redirected budget out of an unproven channel, and held appointment volume on lower paid spend.
Situation
A multi-state production homebuilder, running paid media across many metro markets. Active campaigns spanned Google Search, Google Display, and Meta. Roughly half of the Google budget was going to Display. The program had been built up over years by a mix of internal team members and an external agency, each working with their own naming conventions, bidding strategies, and conversion goal selections.
A former coworker had landed there in a marketing role, knew the team was paying an agency too much for not enough discipline, and asked for an outside audit. The audit became the engagement, which started in late 2025.
What was broken
Six things needed work before anything else made sense.
- Over two hundred campaigns with inconsistent naming. Hard to pull a clean report. Harder to teach a learning algorithm what good looks like.
- Half of Google spend going to Display. A channel the team trusted, with platform-self-reported conversion numbers nobody had pressure-tested.
- Conversion goals miscategorized on a meaningful share of campaigns. Smart Bidding was optimizing toward the wrong outcomes.
- Bidding strategies inconsistent across campaign types. Some campaigns on Max Clicks, others on portfolio configurations that had stopped working.
- UTM parameters partial or wrong. GA4 could not reliably attribute paid sessions to the campaign that drove them.
- Headlines and descriptions not fully populated on a large share of ads, capping Quality Score and click-through rate.
Half of the Google budget was going to a channel the client trusted but had never been pressure-tested.
The work
The first months were rebuild work.
Every active campaign was rewritten with full headline and description coverage. Conversion goals were corrected. Bidding strategies were standardized by campaign type. UTM parameters were normalized so GA4 could trust the source data again.
The account structure started as a hybrid: division at the top, neighborhood inside it. Portfolio bidding inside each division held the Search program together while everything else was being repaired. As performance and inventory data accumulated, that structure migrated again, this time to individual daily budgets at the neighborhood level for finer-grained control.
In parallel, Display was challenged. Rather than trust the platform-reported conversion numbers, the engagement designed a geo-based incrementality test to measure what Display was actually contributing. That test is in flight.
The result
18% less paid spend, 55% fewer clicks, and appointments still grew.
Comparing the engagement window to the same months a year prior:
- Paid media spend down 18%. Google down 20%, Meta down 14%.
- Paid clicks down 55%. Most of the lost volume was Display impressions that were never going to convert.
- Google Ads click-through rate up from 1.4% to 6%. Direct result of the Display-to-Search shift, plus tighter ad copy on Search.
- Engaged session rate up from 33% to 58%. Better-quality visitors landing on the site.
- Captured weekly leads roughly doubled in GA4, under a stricter event configuration that counts less than the prior setup.
- CRM appointments up 3.4% year over year across the engagement window.
Less money. Fewer clicks. More closed conversations.
What’s still true
The engagement is ongoing. Per-neighborhood individual budget management has been rolled out across all divisions, replacing the portfolio-bidding scaffolding from the early months. The Display incrementality test is live and will produce a verdict in summer 2026. Reporting infrastructure built during the engagement continues to feed weekly pacing decisions and division-level performance reviews.
