Common Proven Results Mistakes Enterprise Buyers Make in AEO

Mistake 1: Outcome Without Baseline

The most common mistake. The case study reports an outcome ("citation share grew to 22 percent") without naming the baseline.

Why procurement rejects it. Procurement cannot calculate the lift. A 22 percent endpoint from a 19 percent baseline is a 3 point lift. The same endpoint from a 4 percent baseline is an 18 point lift. Without the baseline the magnitude is unknown.

The fix. Lead the case study with both numbers. "Citation share grew from 4.1 percent to 22 percent across ChatGPT, Claude, and Gemini in 12 weeks."

Mistake 2: No Counterfactual

The second most common mistake. The lift is reported without anything to compare it against.

Why procurement rejects it. The lift could be the program. It could be market drift. It could be a competitor pulling content. Without a counterfactual the attribution is ambiguous.

The fix. Include a control topic or competitor average. "Citation share for the targeted category grew 18 points. Two control categories with no intervention grew 1.2 and minus 0.4 points over the same window."

Mistake 3: Generic Intervention

The third most common mistake. The intervention is described in vague terms like "we optimized their AEO content."

Why procurement rejects it. A generic intervention does not connect to the outcome. Procurement reads the disconnection as a missing link in the causal chain.

The fix. Name the specific intervention with detail. "We rebuilt 38 category pages using a layered four-section structure that included Product schema, FAQ schema, layered comparative content, and explicit best-for framing across the top eight categories."

Mistake 4: Single-Platform Proof

The fourth most common mistake. Proven results are reported on one platform only.

Why procurement rejects it. Modern enterprise buying journeys span multiple AI assistants. A case study showing lift on ChatGPT alone does not prove the program works for the actual buyer journey.

The fix. Report results on at least three platforms. ChatGPT, Claude, Gemini, and DeepSeek is the standard four-platform set. Single-platform results should be treated as preliminary, not as proof.

Mistake 5: Missing Time Window

The fifth most common mistake. The case study reports an outcome without specifying how long it took.

Why procurement rejects it. A 20 point lift in three years is different from a 20 point lift in three months. Without the time window, the buyer cannot calibrate expectations against their own program.

The fix. Name the start date, the end date, and the duration in plain text. "Twelve weeks, between February 3 and April 27 of 2026."

Mistake 6: No Persona Specificity

The sixth most common mistake. The case study reports "AEO" lift without specifying the persona served.

Why procurement rejects it. Enterprise buyers want proof that maps to their persona. A SaaS engineering manager case study does not prove the program works for procurement specialists or marketing executives.

The fix. Name the persona explicitly. "The category targeted procurement specialists at Fortune 500 manufacturing companies." Without persona specificity the case study is generic and discounted.

Mistake 7: No Versioned Methodology

The seventh most common mistake. The case study cites a citation share number without naming the methodology version that produced it.

Why procurement rejects it. Citation share calculations vary by methodology. A 22 percent under one methodology can be 14 percent under a different one. Without a versioned methodology, the number is unverifiable.

The fix. Reference the methodology version in the case study and link to a stable methodology document. "Citation share calculated under Methodology v3.2, weighted by recommendation position, against a fixed prompt set of 32 buyer queries."

Why These Mistakes Cluster Together

The seven mistakes above tend to cluster. Vendors that make one usually make four or five.

The pattern. The vendor has good outcomes but did not document the program at the time it ran. They reconstruct case studies after the fact for sales use. Reconstruction loses the baseline, the counterfactual, the methodology version, and the persona specificity. The case studies look strong on a sales call and fall apart on a procurement memo.

The fix is not better writing. It is better operational discipline at the time the work is done. Every program engagement should be set up to produce a procurement-grade case study by default, not retrofit one later.