Why We Didn’t Optimize Google Ads Until Conversion Signals Were Rebuilt

A decision case from an international engineering B2B context

This decision case describes a real project from an international engineering B2B context and focuses on Google Ads decision logic, rather than on commercial outcomes or performance metrics.

The case is not presented as a success story or an example of campaign optimisation. Instead, it documents a diagnostic decision process: which risks were identified, which signals were analysed, and why restructuring conversion signals was a necessary step before any attempt to optimise or scale advertising.

The purpose of this case is to demonstrate how decisions are made in complex B2B environments, where traffic alone is not a result and signal quality matters more than volume.

Engineering test system illustrating why signal definition matters before optimisation: complex B2B environments require clarity on what constitutes a real decision signal before scaling Google Ads.

Table of Contents

1. Introduction: Why Google Ads Decisions Fail in Complex B2B

2. Context: Engineering B2B, Long Sales Cycles, and Decision Risk

3. Symptoms: what looked acceptable but carried risk

4. Visibility & Decision Review Framework

5. Key Findings from the Review

6. The Decision: Restructuring Conversion Signals Before Scaling

7. What Would Have Happened Without This Decision

8. Conclusion: Optimisation Starts with Signal Definition

9. Evidence box

Context: business and initial situation

The project involved a Finland-based B2B manufacturer of specialised engineering equipment used for corrosion and materials testing under HPHT conditions and according to NACE / ASTM standards.

The target audience consisted of engineers, QA laboratories, and R&D teams.

The market was international, with long sales cycles and delayed purchasing decisions.

Google Ads was considered as an always-on visibility channel and a potential source of inbound demand. From the outset, however, it was clear that in this type of B2B context traffic itself could not be treated as a meaningful outcome.

The definition of a “conversion” would determine whether the entire system measured intent or noise.

Symptoms: what looked acceptable but carried risk

At a surface level, nothing appeared broken:

• campaigns were structured correctly,

• traffic was present,

• analytics interfaces showed user activity.

A closer look revealed several risks:

• In this category, the primary user action is often not a form submission, but email contact. Engineers frequently copy an email address, save it, and return later. This behaviour does not resemble typical B2C lead generation.

• Analytics data showed clicks on email and contact elements, but if conversions were defined incorrectly, Google Ads would optimise toward weak signals.

• The initial assumption of “increasing traffic” posed a classic B2B risk: impressions and clicks could grow while qualified enquiries remained unchanged.

  
  

Visibility & decision review: what was analysed

The analysis was conducted across three layers.

1. Users and behavioural signals

We examined which contact elements were actually used as micro-signals of intent.

GA4 event data over the analysed period showed:

• email copy actions,

• clicks on email addresses,

• repeated returns to contact-related pages.

• 
  

These signals do not prove sales. They represent decision traces: what users choose to do instead of submitting a form.

2. Search: how the market formulates the problem

The search campaign structure was built around:

• engineering standards (NACE / ASTM),

• technical terminology (SCC, HIC, autoclaves, loops),

• testing methodologies rather than marketing language.

This reflected how the audience searches: by technical problem definition, not by generic product terms.

3. AI interpretation: how the company is understood

An AI visibility assessment showed that:

• the company’s identity was interpreted consistently,

• its technical role was clear,

• but recommendation precision was limited.

The limiting factor was the lack of explicit decision context:

• when this company should be selected,

• for which roles,

• and under which scenarios.

The overall visibility score was 9/12, based on four criteria: identity, audience clarity, problem–solution alignment, and recommendation context.

Decision framework illustrating why signal definition precedes optimisation in engineering B2B Google Ads.

Key decision risks identified

• The primary risk was not in ad copy or keywords, but in the optimisation logic itself.

• The system risked learning from the wrong signals.

• In this B2B context, the most honest early-stage signal was email intent, not form submissions.

• AI systems could identify what the company does, but lacked decision artefacts explaining why and when it should be recommended.

Decision taken: Restructure conversion signals

One decision was taken before any optimisation or budget scaling: to rebuild the conversion and signal framework first.

Decision artefacts

The following actions were fixed as part of the decision:

• Primary conversions were defined as email-based intent signals:

  • copy_email

  • click_email

• The form conversion (generate_lead) was removed as a primary goal because:

  • it occurred rarely,

  • it did not reflect the typical decision path,

  • and it would train the system on a weak signal.

• Success measurement was defined in terms of:

  • CPA / CPQL,

  • lead relevance and quality,

  • not traffic growth or click volume.

Counterfactual: what would have happened otherwise

Had campaigns been launched or scaled without rebuilding the signal framework:

• optimisation would likely favour users who click easily but are not engineers or buyers;

• reports would show apparent improvements (CTR, clicks) without increasing real enquiries;

• any conclusions about performance would remain questionable, as the primary metric would not represent genuine intent.

Conclusion: a decision rule

In international engineering B2B, Google Ads should not be optimised before it is clear which user actions represent a real decision signal.

Practical rule:

If real communication starts with email, email actions must become the primary learning signal. Otherwise, advertising spend purchases traffic, not intent.

Evidence box

This case is based on real project materials and analytical observations from work with an international engineering B2B company.

The analysis draws on:

• project briefs and strategic documentation defining business context and KPI logic;

• GA4 click and event data interpreted as intent signals (copy_email, click_email);

• search campaign structures aligned with engineering standards and methodologies;

• an AI Visibility Scorecard assessing identity clarity, audience definition, problem–solution alignment, and recommendation context.

Commercial performance metrics are intentionally excluded, as this case focuses on decision principles, not performance reporting.