What an AEO Grader Actually Tells You, And What It Doesn't

Why I ran an AEO grader on my own brand

On May 6, 2026, I ran the HubSpot AI Search Grader on Everyday Advantage Studio, the main EAS catalog site. I had two reasons. First, I wanted a baseline before making any GEO changes so I'd have real before/after data. Second, the EAS GEO vertical is built around recommending GEO tools, and I wasn't going to do that without running the same diagnostic I'd recommend to anyone else.

What came back was instructive, not because the scores were particularly good or bad, but because the report contained information that most people reading it would completely ignore. The competitor buckets were wrong on two of the three engines. Understanding why that matters, and what to do about it, is the entire point of this post.

I'm publishing the full data here because transparency about real results is the EAS standard. If you're going to trust a recommendation about GEO tools, you should be able to see what happens when the same person runs those tools on their own brand.

The baseline scores: what three engines saw

The HubSpot AEO Grader tests your brand across ChatGPT (powered by GPT-5.2 with a knowledge cutoff of August 31, 2025), Perplexity (real-time retrieval, no cutoff), and Gemini (Gemini 3 Flash Preview, cutoff January 2025). Here are the overall scores for Everyday Advantage Studio:

The natural reaction to seeing scores in the 30s is "these are bad scores." That reaction is not wrong, but it's also not useful. The scores tell you the magnitude of the problem. They don't tell you which problem you have. For that, you need to look at the six sub-metrics and, most importantly, the competitor buckets.

What each metric actually measures

The HubSpot grader breaks the overall score into six dimensions. Understanding what each one is actually measuring changes how you prioritize remediation.

Two things jump out immediately from the EAS data. First, sentiment scores are relatively strong, 23/40 on ChatGPT, 27/40 on Gemini. The language the engines use about the brand is neutral-to-positive. There's no reputation damage to repair. Second, Share of Voice is zero across all three engines. The brand doesn't surface in any category prompts at all.

That combination, decent sentiment, zero share of voice, tells you something specific. The engines have absorbed some information about the brand, and they don't dislike what they've found. But they haven't categorized it clearly enough to surface it when someone asks about tools in the space. That's a category clarity problem, not a reputation problem.

The competitor bucket: the most underused signal in any AEO report

Every AEO grader shows you which competitors it grouped your brand alongside. Most people scroll past this section. That's a mistake. The competitor bucket is the most valuable data on the page.

Here's why: AI engines don't evaluate your brand in isolation. They evaluate it in relation to other brands they've already categorized. When a user asks "what's the best email tool for creators," the engine doesn't search all brands it knows about, it pulls from its mental model of the creator email tool category. If your brand isn't in that mental model, you won't appear regardless of how optimized your content is.

The competitor bucket tells you which mental model the engine has filed your brand in. And for EAS, the results across three engines revealed three completely different, and mostly wrong, answers.

Gemini got it right. EAS is a curated catalog of operator-grade tools, that's exactly what Kit, LTK, Benable, and Lasso do for their respective audiences. ChatGPT put EAS in the affiliate network category, alongside the platforms that run affiliate programs for brands, not the sites that recommend tools to individual operators. Perplexity put EAS in the affiliate tracking SaaS category, alongside software tools that help marketers measure campaign performance.

Both ChatGPT and Perplexity are wrong. And the wrongness matters because when someone asks ChatGPT "what's the best affiliate tool catalog for indie founders," it pulls from its mental model of creator curation sites, and EAS isn't in that model. It's in the affiliate network model, which means it surfaces when people ask about running an affiliate program, not when they ask about finding curated tools.

Why three engines gave three different reads

Each engine processes information differently, which is why the same brand can get completely different categorizations from three different systems.

Why ChatGPT put EAS in the affiliate network category

ChatGPT's knowledge cutoff for this grader run was August 31, 2025. It relies primarily on its training data rather than real-time web retrieval. EAS launched in 2024, which means it had a relatively short window to appear in training data before the cutoff. The word "affiliate" appears prominently in affiliate marketing content, and the pattern of language used on affiliate publisher sites often resembles the language used by affiliate networks. Without enough distinctive content establishing the "curator" identity, ChatGPT defaulted to the nearest recognizable affiliate category in its training data.

Why Perplexity put EAS in the tracking SaaS category

Perplexity uses real-time web retrieval, which means it found the live site. But the live site at the time of the grader run hadn't yet deployed the structured data block or the tagline system that anchors the brand in the creator curation category. Without JSON-LD declaring the organization type and knowsAbout topics, Perplexity had to infer the category from the text content, and the text content mentioned affiliate marketing tools heavily. It landed in the affiliate tools category, which is closer than ChatGPT's read but still wrong.

Why Gemini got it right

Gemini's categorization was correct, and the grader report explains why. The grader noted that Gemini "pulled digital footprints across social media, creator communities, and marketing resource aggregators." EAS has an active Pinterest presence with significant pin volume. Pinterest is a creator-curation platform by nature, and the pins pattern-matched to how creator curation sites present tools. Gemini's heavier weighting of social signals gave it the right read that ChatGPT and Perplexity missed.

This reveals an important tactical point: Gemini is the engine most responsive to social signal optimization. If your brand has a Pinterest, LinkedIn, or YouTube presence, those signals carry meaningful weight in Gemini's categorization. ChatGPT and Perplexity respond more directly to site content, structured data, and third-party citations.

What the sentiment data tells you, and what it doesn't

EAS scored 23/40 on ChatGPT sentiment, 14/40 on Perplexity, and 27/40 on Gemini. The Perplexity sentiment score being lower than the others is worth examining.

Perplexity's grader report noted "limited public mentions and online presence" and "absence of press coverage or media attention." This is not a statement about brand reputation. It's a statement about brand visibility. Perplexity uses real-time retrieval, so when it sees limited third-party mentions, that directly depresses its sentiment analysis, not because anyone is saying negative things about the brand, but because there aren't enough sources for it to draw a strong positive signal from.

The polarization scores, 22 on ChatGPT, 15 on Perplexity, 25 on Gemini, are all low. Low polarization can mean two things: either the brand is well-liked without strong detractors, or the brand doesn't have enough visibility to generate strong opinions either way. Given the low Brand Recognition scores, this is almost certainly the latter. The brand hasn't generated enough public discourse to create divided opinions.

The five remediation levers, ranked by impact

The AEO grader output points at the problem. It doesn't tell you what to do about it in priority order. Based on the EAS diagnostic and the broader research on what actually moves AEO scores, here are the five levers ranked by leverage for a brand starting near zero recognition.

Lever 1: Get cited on third-party sites engines crawl

This is the single highest-leverage action for any brand with low recognition scores. Directories like G2, Capterra, Product Hunt, AlternativeTo, and Indie Hackers; tool roundup posts on established blogs; podcast transcripts; and guest articles on relevant publications all create the third-party citations that retrieval engines weight heavily. Self-published content matters less than how other authoritative sources describe you.

Lever 2: Publish category-defining content under a consistent brand line

A locked tagline on every page (homepage, footer, masthead, blog posts, newsletter) gives engines a consistent phrase to associate with your entity URL. Engines need repetition of the same phrase tied to the same domain to build the category association. For EAS, this meant deploying "The Operator's Edge" consistently across all touchpoints rather than leaving it defined only in an internal brand bible.

Lever 3: Deploy JSON-LD structured data

Structured data replaces engine guesswork with declared facts. A properly constructed Organization schema with knowsAbout topics, a named founder, and sameAs cross-platform references tells the engine exactly what category to place your brand in. Without it, the engine infers, and as the EAS data shows, inference produces wrong answers. The JSON-LD implementation is covered in detail in the companion post: JSON-LD for Indie Operator Brands: The Minimum Viable Schema.

Lever 4: Publish income reports, case studies, and data transparency content

Real numbers generate third-party citations because other publications quote them. An income report showing $X in affiliate commissions from a specific tool gives other bloggers a data point to reference, which generates inbound citations that feed back into lever 1. This is the mechanism by which transparency content compounds. It's not just content; it's a citation-generation machine.

Lever 5: Get the named operator on niche podcasts

Podcast transcripts are indexed by both training corpora and Perplexity's real-time retrieval. A guest appearance on a relevant podcast with a transcribed audio track creates a third-party citation with the named operator's expertise associated with a specific topic. One quality podcast appearance per month for six months creates meaningful movement in personal brand recognition signals, particularly relevant for Perplexity's source-quality scoring.

The 90-day re-measurement plan

Running an AEO grader once produces a baseline. Running it again 90 days later, against the same prompts and the same engines, produces the data that matters: whether the remediation worked and whether the engines updated their categorization.

The EAS re-measurement is scheduled for August 6, 2026. The specific metrics to watch:

The single most important metric to track is whether ChatGPT and Perplexity update their competitor buckets to include creator-curation brands instead of affiliate networks and tracking SaaS. Score deltas without bucket shifts mean the engines are noticing the brand more but still miscategorizing it. Bucket shifts confirm the repositioning has actually landed.

This post will be updated with the August data when the re-measurement runs. That's the EAS standard, not a one-time report, but a recurring diagnostic tied to specific targets.

What to do with your own AEO grader report

If you've run an AEO grader or are about to, here's the sequence that produces the most useful read:

1. Skip the overall score for now. It's a summary of six different problems compressed into one number. It's not a useful starting point.
2. Go straight to the competitor bucket. Write down which brands each engine grouped you with. Then ask: is that the category I want to be in? If not, that's your primary problem.
3. Check Brand Recognition next. Scores below 5/20 mean the engine doesn't know you exist. Scores above 10/20 mean it knows you but may have you in the wrong category. These require different remediation.
4. Look at Presence Quality. This is the most actionable metric in the short term, it reflects your own published content and structured data, which you control directly.
5. Note Perplexity separately. Perplexity uses real-time retrieval, which means changes to your site and third-party citations show up much faster than in ChatGPT or Gemini. If you're going to prioritize one engine for early movement, prioritize Perplexity.
6. Build a 90-day measurement plan before you do anything. Define the specific metrics and bucket changes you're targeting. Otherwise you'll run the grader again in three months and not know whether what you did worked.

The HubSpot AI Search Grader is free. It takes about 60 seconds to run. If you haven't run it on your own brand yet, that's the starting point, not this post, not any tool, not any strategy. Run the grader first, then come back and apply the diagnostic framework above to your own results.