The scenario
This playbook applies to a B2B technology company, say one selling infrastructure monitoring software to mid-sized IT teams, that relies on Google Ads for most of its inbound demo requests. The account reports a healthy cost per lead. The problem surfaces in the CRM rather than the ads platform: many leads are students, job seekers, existing customers looking for support, or companies far too small to buy.
A typical trigger is marketing and sales disagreeing about whether paid search is working at all. Axoria is brought in to run the PPC programme and settle that disagreement with data.
The challenge
The account is optimising for the wrong signal. Smart Bidding has been trained on form submissions, so it has learned to find cheap form fills wherever they come from. Loose broad match, a single generic landing page and no feedback from the CRM mean the algorithm cannot tell a demo request from a hobbyist apart from one from a head of infrastructure. The real objective is not a lower cost per lead but a lower cost per sales-accepted opportunity, measured in the CRM.
Research and analysis
We export a year of search-term data and join it to CRM lead status by click ID. In this scenario support-intent queries, competitor-brand queries and “free” or “open source” modifiers tend to account for much of the spend and almost none of the pipeline, while product-category queries with a company-size qualifier convert less often on the form but far more often in the CRM.
A tracking audit runs alongside. The common findings are duplicate conversion actions double-counting submissions, and a conversion window too short to capture the lag between first click and demo request for a considered B2B purchase.
Strategy
One principle: give the bidding algorithm the outcome the business cares about, then structure everything else so that outcome is easy to find.
Offline conversion import
Push sales-accepted and opportunity-created stages from the CRM back into Google Ads via GCLID, so Smart Bidding can optimise toward qualified outcomes rather than form fills.
Intent-led structure
Rebuild campaigns around intent themes with tightly matched ad groups, controlled use of broad match, and a negative keyword architecture shared across the account.
Landing pages by intent
Replace the single generic page with a small set of pages matched to query intent, each with a qualification question on the form to filter obvious non-buyers before they reach the CRM.
We agree the trade-off in advance: form-fill volume will fall and cost per raw lead will rise. Both numbers go on the dashboard from the start so nobody is surprised.
Execution
The first month goes on tracking: deduplicating conversion actions, extending the window, capturing GCLID into hidden form fields and scheduling daily CRM uploads. We switch the bidding target only once the offline import has accumulated enough qualified conversions, because moving Smart Bidding to a sparse signal too early starves it.
The restructure follows. Campaigns are rebuilt around intent themes with a shared negative list covering support, careers, education and competitor-brand terms. New landing pages go live at the same time so message match improves alongside targeting. A deliberately small competitor campaign is kept with its own budget cap and a qualification-heavy page.
Optimization
With CRM outcomes flowing back into the platform, weekly optimisation comes down to one question: which search terms, audiences and pages are producing accepted opportunities? Search-term hygiene becomes a scheduled discipline. Bid targets are adjusted gradually as the signal matures, landing pages are tested on qualified rather than raw submission rate, and budget moves toward the intent themes with the shortest path to pipeline, which our lead generation reporting tracks through to closed-won.
How results are measured
The CRM is the source of truth. Every figure is measured there, on an attribution basis agreed by marketing and sales before the programme starts, and compared against the months before the engagement. Google Ads reporting is used for diagnosis, not for the scorecard.
- Cost per qualified lead: spend divided by sales-accepted leads, reported next to cost per raw form fill. A programme on track shows the qualified figure falling even while the raw figure rises, because spend stops going to queries that never produce pipeline.
- Lead quality: the share of paid-search leads accepted by sales, taken from CRM stage changes. Success looks like that share rising to the point where sales asks for more budget in the channel rather than less.
- Pipeline contribution: opportunities and pipeline value sourced from paid search as a share of all new opportunities in the CRM. In a working programme paid search becomes one of the largest sources of new opportunities on the agreed attribution basis.
What we have learned
- Fix the signal before the structure. Restructuring an account that optimises for the wrong outcome just finds it more efficiently.
- Expect volume to drop, and say so. Putting raw and qualified lead counts side by side on the dashboard prevents a panic in the second month.
- Qualification questions are a feature. Adding a company-size field reduces form completions and improves everything downstream. This applies broadly across technology and software buyers.
- The CRM is the source of truth. Google Ads reporting is only as honest as the conversion data you feed it. Read more on how we approach Google Ads management.