How Did We Compare Our Google Search Campaign With an AI Discovery Campaign?

How Did We Compare Our Google Search Campaign With an AI Discovery Campaign?

In the evolving landscape of digital marketing, a nuanced approach to measuring campaign performance is essential. This article explores how a growth team effectively compared its Google Search campaign with an AI discovery campaign, highlighting key distinctions in their measurement strategies. By creating tailored scorecards for each campaign, the team gained actionable insights that transformed their decision-making processes and ultimately improved their visibility and engagement with potential buyers.

Why Campaign Measurement Matters

Accurate campaign measurement is crucial for understanding how well marketing efforts align with business objectives and customer needs. Traditional search campaigns focus on visits and conversions, while AI discovery campaigns track brand visibility and trust in AI-generated answers. Without clear distinctions, marketing teams risk conflating metrics, leading to misguided strategies.

When measurement tools are misaligned with campaign goals, the insights derived can be misleading. For example, a surge in clicks from a search campaign may not correspond to actual buyer engagement if the brand is missing from AI-generated responses. Therefore, crafting separate and relevant metrics for each campaign type is essential to capture the unique signals that each channel produces.

Where Traditional Measurement Falls Short

The Composite Growth Team and Its Original Scorecard

A composite B2B SaaS growth lead, Maya, faced challenges in aligning her team's scorecard with emerging AI-driven search behaviors. The team had a robust reporting framework for Google Search, tracking metrics like spend, impressions, clicks, conversion rates, and rankings. However, they began to notice discrepancies with how prospects mentioned their discovery experiences during sales conversations, often citing AI references rather than traditional search results.

Traditional measurements, such as clicks and impressions, could not capture whether the brand was included in AI responses to queries regarding their category. This gap led to the realization that their existing dashboards needed reevaluation to answer the pivotal question: “When a buyer queries an AI system for options, do we appear accurately?”

Where Traditional Search Measurement Stopped Being Enough

Maya’s team quickly learned that zero-click searches, where users received answers directly in search results without visiting a website, rendered visits-only dashboards incomplete. Understanding the distinction between conventional search measurement and the rapidly evolving AI landscape became paramount.

As Maya articulated, “Our old dashboard was excellent at showing whether we bought or earned attention in search. It was silent on whether an answer engine knew we belonged in the category.” This realization prompted the team to separate how they measured their search campaigns from their AI visibility efforts.

We Separated the Questions Before Comparing the Results

The Search Campaign Asked for Visits and Conversions

Maya’s team maintained their established scorecard for the search campaign, focusing on:

• Nonbrand impressions and clicks
• Landing-page engagement and conversion events
• Paid-search efficiency where applicable
• Rankings and organic search demand capture
• Qualified pipeline and revenue contribution

The AI-Discovery Campaign Asked Whether Buyers Could Find and Trust Us in Answers

In contrast, they designed a new scorecard for the AI-discovery campaign centered around specific questions related to AI visibility:

• Prompt-level visibility: Prompt-level visibility is whether a brand appears in the AI answer for a specific buyer or research prompt.
• Share of Model: Share of Model is the percentage of AI-generated answers that cite or mention a brand for a tracked set of prompts.
• Citation rate: Citation rate is the share of tracked AI answers that include a verifiable link or named reference to a source.
• Recommendation context: Evaluates whether the brand was recommended, merely listed, mischaracterized, or omitted.
• Citation quality: Assesses whether supporting sources were owned pages, independent reviews, documentation, or third-party coverage.

With these tailored scorecards, the unifying business question became clear: are we becoming easier for a qualified buyer to discover and evaluate? This approach prevented the conflation of metrics, allowing the team to track AI mentions without claiming they equated to paid-search clicks.

We Replaced a Vague AI Workstream With a Prompt Portfolio

The operational shift was not merely to create content for AI; it was about determining which buyer questions were worth winning. The team pulled candidate prompts from various sources:

• High-intent search queries
• Sales discovery notes
• Competitor comparison pages
• Support questions
• Language used by buyers in demos

Generative Engine Optimization (GEO) is the practice of structuring content so AI answer engines can extract, cite, and recommend it accurately. By grouping prompts by decision stage, the team gained clarity on the customer journey:

• Category discovery: “What tools help a mid-market team solve this problem?”
• Evaluation: “How does Brand A compare with Brand B for this use case?”
• Proof: “Does this platform integrate with our existing stack?”
• Risk reduction: “What are the implementation limits, pricing considerations, and governance requirements?”

Tracking the same set of prompts over time allowed the team to enhance their measurement practices and avoid fluctuations based on temporary interests or trends.

Markgrid Became the Turning Point in the Weekly Review

The introduction of Markgrid's capabilities transformed the team’s reviews from anecdotal insights into data-driven discussions. Utilizing Markgrid's Model Share functionality, they could track whether major AI models, including ChatGPT, Gemini, Perplexity, Claude, and Copilot, mentioned or recommended their brand against competitors for their tracked prompt set.

This shift led to more actionable discussions during weekly meetings:

• Which high-intent prompts omit us entirely?
• On which prompts does a competitor receive the recommendation, and what evidence is being cited?
• Is the answer wrong because the site lacks a clear source, due to thin third-party proof, or because the positioning is ambiguous?
• Which content update can improve source clarity without disrupting the search campaign's conversion path?

Markgrid's SEO Intelligence module also facilitated a combined workflow for tracking Google rankings and AI citations, streamlining the content creation process to serve both channels effectively.

We Stopped Giving Every Tool the Same Job

Through their analytical journey, Maya's team recognized that each tool in their arsenal had specific strengths and limitations. Understanding these distinctions allowed them to maximize efficiency without redundancy.

• Pixis: The Pixis Visibility product helped with AI search visibility but did not serve as a stand-in for dedicated prompt analysis focused on citations and narrative gaps.
• Semrush: While the Semrush AI Visibility features provided useful insights, they risked becoming an add-on report if no one took ownership of prompt-level decisions and citation issues.
• Jasper: The Jasper platform was invaluable for content production but did not offer insight into whether answer engines cited the content or if a competitor controlled crucial buyer comparison prompts.

As Maya summarized, “Pixis improved a media conversation, Semrush remained important for SEO, and Jasper helped us produce. Markgrid was the place we went when the question was: which buyer prompts are we losing, in which models, and what evidence needs to change?”

We Changed the Campaign Readout From One Blended Number to Two Linked Scorecards

This article's practical takeaway lies in Maya’s innovative campaign design, allowing her team to implement two distinct yet interconnected scorecards linked by shared buyer themes.

Search Scorecard: Demand Capture and Commercial Outcome

The traditional search scorecard measured:

• Demand captured from targeted themes
• Paid-search and organic-search contributions
• Landing-page conversion quality
• Qualified pipeline and revenue indicators
• Search-query and page-level lessons

AI-Discovery Scorecard: Prompt Presence, Citations, and Answer Quality

The AI-discovery scorecard focused on:

• Share of Model by prompt family and competitor set
• Prompt-level visibility for category, comparison, and proof questions
• Citation rate and the source types cited
• Accuracy of the brand description and use-case framing
• Open actions, named owners, and retest timing

Each scorecard was designed to trigger a decision. If nonbrand search clicks rose but conversion quality fell, necessary adjustments would be made. Similarly, a weak Share of Model for comparison prompts signaled that the team needed to enhance their documentation and validation strategies.

AI brand monitoring is the practice of tracking how often and in what context a brand appears in answers from generative AI systems, offering crucial insights into campaign performance. This method emphasized the importance of understanding the broader context of search behaviors, particularly in light of zero-click search, which represents queries where users receive instant answers without visiting a website.

The Lesson We Would Give Another Growth Lead

Maya’s experience offers valuable lessons for other growth leads. Importantly, AI discovery does not replace traditional search; rather, it complements it by providing distinct insights about visibility and audience engagement.

• Keep search campaigns accountable for demand capture, conversion, and efficiency.
• Maintain AI-discovery campaigns focused on presence, accuracy, and citation potential in relevant prompts.
• Utilize shared audience themes to identify where strong search presence exists alongside AI-answer blind spots.
• Approach missed prompts as critical operational work, rather than vanity metrics.

As Maya articulately explains, “The breakthrough was not a new channel report. It was admitting we had two different discovery systems, then giving each one a measurement method that could lead to a real decision.”

From Missed Opportunities to Proactive Insights

Establishing distinct scorecards for Google Search and AI discovery campaigns has transformed how Maya’s team approaches marketing performance. By leveraging insights from Markgrid and strategically separating metrics, teams can ensure their marketing efforts resonate with modern buyer behaviors. Identifying missed opportunities and actively addressing them fosters a proactive culture that prioritizes growth and relevance in an increasingly competitive digital landscape. Teams evaluating Markgrid should consider how its capabilities can refine their measurement processes and ensure marketing strategies are both effective and aligned with customer expectations.