AEO Analytics

Run an analytics loop, find an opportunity, and defend your diagnosis. This certification tests whether you can configure an environment, analyze the data, and translate what you find into an actionable recommendation.

Target Audience

Analysts and AEO Practitioners

Skills

Deep Analysis, Opportunity Discovery

Submission Type

Project Based Assignment

About

This certification tests whether you can run Answer Engine Insights like an analyst.

You will open Profound, set up a clean config, run the full diagnostic loop from the Setup and Analyze sections of this course, find a real opportunity in your data, and present it on camera with a root cause and a specific recommendation.

There is no template. There is no right answer. Every config is different, every competitive landscape is different, and the opportunity you find will be unique to your data.

What we evaluate is the rigor of your process. Did you clean the lens before you looked through it. Did you drill past the aggregate. Did you trace the supply chain. Did you arrive at a diagnosis that tells you exactly what to do next.

The Loop

The analytical loop follows the progression taught in Lessons 11 through 17. It is not a checklist. It is a sequence of increasingly focused moves.

1. Filter your platforms. Clean the lens before you look through it. Select the platforms that represent real usage for your market. The course teaches ChatGPT, AI Overviews, AI Mode, and Gemini as the default big four. Every number downstream depends on this filter being right.

2. Scan ranks at the topic level. Look for big movement and low rank. Movement means something changed. Low rank means opportunity space. Use rank as your primary signal, not visibility score. Rank is relative and truthful. Visibility score tracks momentum but hides your competitive position.

3. Drill into prompts. The topic aggregate is always lying. Two prompts in the same topic can move in opposite directions. The story is at the prompt level. Always.

4. Read the citation supply chain. Citations are the root of everything. The domain mix tells you the playbook: companies mean owned content can work, institutions mean you need partners, earned media means you need to earn or buy your way in. The page slugs tell you the format that is winning. The visibility-citation gap tells you whether your brand is in the answer but your content is not the source.

5. Isolate by platform when warranted. If the aggregate looks uniform, stay there. But the same topic can have completely different supply chains on ChatGPT versus AI Overviews. When it does, you need a different fix for each.

6. Arrive at a root cause and recommendation. Not "we dropped." Not "visibility is down." A full diagnosis: which prompts moved, on which platforms, what changed in the citation supply chain, and what specifically you should do about it. Create owned content. Improve existing content. Pursue earned media. Or accept a structural limitation and reallocate.

Config Hygiene

A dirty config produces dirty data. We evaluate your config as part of the certification.

Tags are required. At minimum, apply intent tags — commercial versus informational — to your prompts. Without them, informational prompts silently wreck your aggregate visibility score and you cannot build clean views. Initiative tags are strongly encouraged. The system we teach: INTENT, CAMPAIGN, STRATEGY, EXPERIMENT. Use it or present a justified alternative.

Competitor list must be curated. Your competitors should reflect who actually shows up in your answer landscape. Irrelevant competitors muddy the view. Missing competitors create blind spots. Fix both in Settings.

Citation categories should be reviewed. If you have PR placements, earned media partners, or affiliate relationships that are miscategorized, fix them before you analyze. The supply chain labels need to reflect your actual strategy.

Name matching should be checked. Brand variants cause ghost mentions — your brand is referenced but not counted. This is especially common with hyphenated names, abbreviations, and brand-plus-product patterns. Fix them.

Extra Dimensionality

The baseline requirement is a clean tag structure and a complete analytical loop. The strongest submissions go further. There is no formal extra credit, but additional analytical dimensionality will be viewed favorably.

Personas. If you have personas configured, show how visibility or citations shift by persona. This demonstrates that you understand context-dependent answers. The Persona Loop — Observe, Build, Inject, Compare — is a direct application of Lesson 7.

Regions. If your business operates across geographies, filter by region. Show how the competitive landscape or citation supply chain differs.

Tag-filtered views. Using AND/OR tag filtering to isolate a specific slice — for example, INTENT: Commercial AND STRATEGY: Historical Data — and running the loop on that slice shows operational maturity.

Sentiment layer. If your opportunity involves narrative, pull sentiment themes and trace them to source citations. This shows you can distinguish a perception problem from a reality problem.

Platform matrix. Use the platforms page matrix view — competitors by platform, topics by platform, tags by platform — to identify structural patterns across the performance landscape.

What You Submit

One deliverable.

A Loom video walkthrough, 10 to 15 minutes. Record yourself running the analytical loop live inside Answer Engine Insights. Start by showing your config: tags, competitors, citation categories. Then run the loop. Filter platforms, scan ranks, drill into prompts, read citations, isolate by platform if needed, arrive at a root cause and recommendation.

We want to hear the thinking, not just the clicking.

What Passing Sounds Like

A failing walkthrough: "Our visibility is 45% and it went down. Here are the citations. I think we should make more content."

A passing walkthrough: "I filtered to the big four platforms and scanned ranks. Historical Data is our weakest topic at Rank 4 with 1.9 million volume, so I drilled in. Two prompts are trending in opposite directions — 'best historical data tools' is climbing at 32% while 'how to track historical AI data' dropped 18 points. The aggregate hid both. I went to citations on the climbing prompt. The domain mix is 60% company blogs, which means owned content can work here. The top cited page is a competitor listicle on semrush.com. Our domain is not in the top 20. The visibility-citation gap is massive: 32% visibility, 0.4% citation share. Our brand is in the answer but our content is not the source. I filtered to ChatGPT specifically and the supply chain is even more fragmented — Reddit and Wikipedia dominate. My recommendation: create an owned listicle-format page targeting this exact prompt, structured to be extractable by engines, and pursue a mention on the SE Ranking post that already cites our competitors."

That level of specificity is the standard. Every claim traces to data. Every recommendation traces to the supply chain.

Submit for certification

Rubric

Submissions are evaluated across five dimensions, each scored on a 1–5 scale

Category	Weight	What we're evaluating	3 - Excellent	2 - Meets expectations	1 - Needs improvement
Config Hygiene & Tagging Architecture	20%	Can you set up a clean analytical environment before you start analyzing?	Full tagging system (intent + initiative tags) with AND/OR filtering demonstrated live. Competitor list is intentionally scoped + competitors are manually tagged in the citation category view. Citation categories include at least one manual reclassification. Name matching has been reviewed. The walkthrough treats config as the foundation of analytical credibility.	Intent tags (commercial/informational) are applied consistently. Competitor list has been curated for the config. Citation categories have been spot-checked. The walkthrough shows the config and explains the tagging logic before beginning analysis.	No tags applied, or tags are inconsistent and incomplete. Competitor list is the default with no curation. Citation categories have not been reviewed. The config is not ready for rigorous analysis and the walkthrough starts analyzing without addressing it.
The Analytical Loop	20%	Can you execute the full diagnostic progression: platforms, ranks, prompts, citations, and platform isolation?	Full loop executed with platform-specific citation isolation revealing a supply chain split the aggregate masked. The walkthrough narrates the thinking at each step ("I looked here because..."). Earned media sources are identified. The visibility-citation gap is diagnosed as structural or actionable. The loop feels like a practiced drill, not a first attempt.	Platforms are filtered to a justified set. Rank scanning identifies a real topic. Prompt-level drill-down reveals the story the aggregate hid. Citation supply chain is read: domain mix is interpreted, page formats are identified, and the visibility-citation gap is acknowledged. The loop is complete.	Analysis stays at the topic level without drilling into prompts. Citations are glanced at but not read as a supply chain. No platform filtering applied. The walkthrough describes numbers without interpreting them.
Dimensionality & Diagnostic Depth	10%	Can you add analytical layers (personas, regions, tags, sentiment, platform matrix) that sharpen the diagnosis beyond the baseline loop?	Multiple dimensions layered: tag-filtered slices, persona-based visibility comparison, platform matrix patterns, or sentiment themes traced to source citations. Each added layer is justified by what it reveals. The diagnosis could not have been reached without the extra dimensionality.	At least one additional dimension is used meaningfully (tag-filtered view, persona comparison, or platform matrix). The added layer sharpens or changes the diagnosis. The walkthrough explains why that dimension was worth adding.	Analysis uses only the default view with no tag filtering, no persona comparison, no regional slicing, and no sentiment. The diagnosis is one-dimensional: a single number and a generic recommendation.
Opportunity Articulation & Walkthrough Quality	50%	Can you translate data into a root cause statement and a specific next action, and can you present it clearly on camera?	The root cause statement is precise enough to hand to a content team as a brief. The recommendation specifies format (based on what is winning in citations), target (based on supply chain gaps), and priority (based on volume and business importance). The walkthrough presents the analysis as a story with a beginning (config), middle (the loop), and end (the diagnosis). The presenter speaks as an analyst delivering a finding, not a student completing an assignment.	The opportunity is specific: names the topic, the prompt, the citation gap, and the recommended action. The root cause traces from rank to prompt to supply chain. The walkthrough demonstrates the analytical thinking, not just the navigation. A live run through real data is shown.	The opportunity is vague ("we should make more content") with no link to specific citation data or supply chain analysis. The walkthrough is surface-level, describing what each screen shows without explaining what it means. No root cause statement.