What we walked into
This commercial HVAC contractor came to us with a clear constraint: they wanted leads from commercial buyers only — property managers, facilities directors, building owners — and had no interest in residential service calls. The business had the crew capacity and licensing to scale commercial work, but their Google Ads account wasn't reflecting that focus. Most of their budget was being consumed by residential search traffic, and the leads coming through were the wrong kind.
The structural problem is baked into the HVAC category. Search volume skews heavily residential. Terms like 'AC repair near me' or 'HVAC installation' pull enormous traffic from homeowners, and Google's broad and phrase match behavior drags commercial-intent campaigns into that pool constantly. The client's prior account had no systematic negative-keyword framework, which meant impression share on genuinely commercial queries was low and cost per lead was high — at $290 per lead on roughly $6,400 in monthly spend, the unit economics weren't bad, but volume at 22 leads per month was too thin to support growth.
The core tension we had to resolve was this: tightening targeting aggressively enough to exclude residential intent without collapsing impression share on the commercial terms that actually mattered. If we over-excluded, we'd starve the campaigns of traffic. If we under-excluded, we'd keep burning spend on homeowners. Getting that filter right — and then building a structure around it — was the whole problem.
What we changed
- We built a commercial-only keyword list from scratch, mapping terms by buyer type and job size, then paired it with an exhaustive negative-keyword list covering residential modifiers, DIY intent, and homeowner-specific phrasing across all match types.
- We split campaigns into individual SKAGs organized by service location so each geographic market had its own budget lever and we could read performance signals cleanly without one city's volume distorting another's.
- Once we identified that the per-location SKAG structure was inflating costs — Smart Bidding was fragmenting learnings across too many low-volume campaigns — we consolidated everything into a single campaign with location targeting layered inside it.
- We applied bid modifiers at the location level within the consolidated campaign to preserve the geographic prioritization we had built into the SKAG structure, so budget still weighted toward the client's highest-value service areas.
- We used ad customizers to dynamically insert city and service-type references into ad copy, maintaining location-specific messaging for users without needing separate campaigns to carry that personalization.
What happened next
The first thing that moved after consolidation was impression share. Fragmented into per-location SKAGs, the account had been sitting at 22% impression share on commercial terms — the campaigns simply didn't have enough conversion data individually to let Smart Bidding bid confidently. Once we consolidated, the single campaign accumulated signal faster, and impression share climbed to 58%. We were showing up on more of the searches that mattered without loosening the keyword or negative structure.
Cost per lead dropped from $290 to $165 as the bid strategy stopped over-bidding to compensate for uncertainty. With more conversion history feeding into a single campaign, the algorithm got more accurate about which auctions were worth entering and at what price. The negative-keyword framework meant the impressions we were winning were higher-intent to begin with, so conversion rates on clicks improved in parallel.
Monthly leads went from 22 to 71 — a 3x increase — while monthly spend moved from $6,400 to $12,000. That spend increase is meaningful: we weren't just getting more leads at the same cost, we were scaling into a larger budget and still cutting cost per lead nearly in half. The account today runs a single consolidated campaign with location bid modifiers and ad customizers handling the specificity that the SKAG structure used to provide structurally.
The takeaway
In verticals where one buyer segment dominates search volume, the negative-keyword list is as important as the keyword list — and campaign consolidation only works if you've done that filtering work first. Running one campaign with clean targeting and a full conversion history outperforms multiple fragmented campaigns with precise structure but thin data. Location specificity in ad copy and bids can be preserved through customizers and modifiers without splitting campaigns in ways that starve Smart Bidding of the signal it needs to perform.