Six weeks later, every dollar was earning two. Here's how — and why we hand the client the query-classification logic to own.
Weekly blended ROAS, pulled directly from the ad account. The account crossed break-even in week three and kept compounding as the governance system fed cleaner data into bidding.
$80K–$150K/month against the world's largest travel aggregators, with ROAS at 1.02 — broad-match waste compounding daily and no negative-keyword governance.
Full rebuild around high-intent booking queries, a custom AI system classifying every search term daily, and value-over-cost bidding with CPA floors.
ROAS doubled, CPA down 41%, and a 1,800-term negative list plus a 47-page change log handed over — portable if they ever leave.
FlightSearchDirect was competing head-to-head with the largest travel aggregators in the world on Google Search. Despite $80K–$150K/month in budget, ROAS had fallen to 1.02 — the account was barely breaking even on every dollar spent.
The root cause was targeting quality. Campaigns triggered on huge volumes of irrelevant, low-intent queries — non-flight terms, informational searches, competitor brand terms with no commercial intent. With no negative-keyword governance and no system to continuously classify incoming search terms, wasted spend compounded every single day.
No bigger budget. We fixed the foundation, automated the governance, and pointed bidding at profit instead of volume.
Every campaign restructured into tightly themed ad groups by route type, airline brand and booking intent. Broad-match waste cut entirely — the account stopped paying to appear on queries that could never convert.
An automation layer classified every incoming search term daily — auto-promoting converters to keywords, pushing waste to negatives, and tagging each action “Added by AI” for a full auditable trail. Negatives grew from 200 to 1,800+ in six weeks.
Switched to conversion-value bidding with hard CPA floors, so the algorithm optimised toward profitable bookings rather than raw conversion volume — finally working with clean data instead of against noise.
Before any of it, conversion tracking was rebuilt to count completed bookings rather than clicks — so every downstream optimisation rested on a number that reflected real revenue.
Every figure measured from the live ad account at the start of week one and the end of week six.
Discovery → measurement → execution → compounding.
Pulled 90 days of search-term data. Identified 4,200+ wasted query types. Rebuilt conversion tracking to count bookings, not clicks.
Killed the broad-match-only campaign. Built tightly themed ad groups by route type and booking intent. First 600 negatives shipped.
Query Governance system deployed. Daily classification of every incoming search term. Every change tagged “Added by AI” with an auditable log.
Value-over-cost bidding with CPA floor thresholds. The algorithm finally had clean data to optimise against — and found the profitable queries fast.
From 1.02 to 2.08 ROAS. CPA down 41%. The 1,800+ negative list keeps growing — auditable, owned by the client, portable if they ever leave us.
Most agencies keep their logic opaque on purpose. We hand it over — documented, portable, and owned by the client.
An account spending $150K/month at 1.02 ROAS is a burning building. Manual management simply cannot keep pace with the volume of new search queries at that scale.
The Query Governance system was what made recovery possible. Building an automated layer that auditably classifies and acts on every search term — daily — is the difference between an account that drifts and one that improves continuously.
The “Added by AI” tagging means the client owns the logic. If they ever leave us, the system goes with them. That's the part most agencies won't do — their bid logic is opaque on purpose. Ours is open by design.
They handed us a 47-page log of every change they'd made. No other agency has ever done that. The query-governance system is now part of how we think about Search.
Book a free 20-minute Revenue Leak Audit. We'll review your campaigns, identify the biggest leak, and write you a plan — same week.