Brand Health

Summary#

Brand Health runs six brand-measurement tools against one brand at the same time: how often the AI engines mention you, how often the web does, what the sentiment is, which stories are driving it, what the press is saying and how the social signal looks. It costs 10 credits.

Purpose#

Brand measurement is scattered across places that do not talk to each other — mentions in one tool, sentiment in another, AI answers in a third. That makes it impossible to answer the question an executive actually asks: is our brand in better or worse shape than last quarter? This workflow puts all six signals in one document so the answer is a single read rather than a reconciliation exercise.

Overview#

You give it a brand domain. Six tools fire:

The AI Visibility step is genuinely live: it writes a set of category questions that never name your brand, asks every AI engine that has a key configured, and counts how often you come back in the answer. The others draw on measured provider data. The run card is explicit about the mix: Brand Monitor + Media + Social via DataForSEO · AI Visibility queries 4 LLMs live · Perception & Narrative Drivers are AI estimates.

Benefits#

• One number for AI visibility, measured rather than assumed.
• Owned, earned and AI-generated coverage in one document.
• A repeatable quarterly artefact — the same six sections every time.
• The AI Visibility step also feeds Growth Actions, so a Brand Health run unlocks a ranked action queue.
• A failed step does not lose the run; the rest still render.

Use Cases#

• Quarterly brand review. One export that answers "how are we doing" across search, AI and press.
• Post-launch check. You shipped a campaign and want to see whether the narrative moved.
• AI-visibility baseline. You need to know whether the assistants your buyers use actually name you before you invest in changing that.
• Crisis awareness. Media Monitoring surfaces coverage and crisis flags early.
• Competitive positioning. Run it on a rival to see the reputation gap you can position against.

Requirements#

• A signed-in account on a paid plan. None of these six tools runs on Free.
• At least 10 credits of monthly allowance available.
• A brand that has some public footprint. Very small or brand-new names return honest empty states rather than invented numbers.
• The AI Visibility step needs at least one AI engine configured on the platform. Which engines are live is a platform setting, not an account setting — see AI models used across the platform.

Permissions#

Cost#

10 credits per run. The Quota Cost card reads 10 / 10 credits per run, and the run banner says: Running this workflow uses 10 reports (6 tools: AI Visibility + Brand Monitor + Perception + Narrative Driver + Media Monitoring + Social Tracker).

That is 1 + 3 + 1 + 1 + 1 + 3, the same as running the six tools individually. Cached steps still charge. See example costs nothing. See How credits work.

Navigation Path#

Dashboard → Get Started → Brand Health

The sidebar entry carries a red WORKFLOW badge.

Inputs#

The topbar country picker is sent with the run and shapes the measured steps. Press Enter to run — Tip: hit Enter to fire. See example fills nike.com when the field is blank and renders stored samples free.

Step-by-Step Guide#

1. Open Dashboard → Get Started → Brand Health.
2. Set the country picker if the brand's market is not worldwide.
3. Type the brand domain into brand domain (e.g. nike.com).
4. Press Run Brand Health Scan · 10 reports, or press Enter.
5. Watch the six rows in Analysis Progress. Expect one to three minutes — the AI Visibility step is querying live models and is usually the slowest.
6. Read the hero, then the Strategic Insights and Action Plan.
7. Open the per-tool blocks for the evidence.
8. Export or share the report from the hero buttons.

Reading the Results#

The hero. The brand, VERIFIED FROM DATAFORSEO, US · Desktop, the title Brand Health Report, and N of 6 analyses completed. Quick-stat tiles appear for whichever figures came back above zero — most often Mentions from Brand Monitor.

AI Visibility — the block to read first. This is the only measurement in Metric Vault that queries the assistants directly.

• Visibility score is the share of chances the brand was actually named: appearances divided by prompts multiplied by engines, as a percentage. Ten prompts across four engines is forty chances; being named in ten of them is 25%. Under 10% means the assistants effectively do not know you in this category. Between 10% and 40% means you appear when the question is close to your niche and vanish when it widens. Above 60% is category leadership.
• Platforms breaks the score down per engine. A high ChatGPT score with a low Perplexity score usually means you have brand authority but weak citable, current web sources — Perplexity leans on live retrieval. Fix that with well-structured, factual pages rather than more marketing copy.
• Top prompts lists the questions where you did appear; the opportunity is always in the ones where you did not. Each miss is a content brief.
• Sample excerpts show the actual sentences the models produced — a win, a mixed mention and a miss. Read these. They tell you how you are characterised, which no score can.
• Sentiment by platform flags an engine that describes you worse than the others. That usually traces to one dominant source.
• Citation sources lists the domains the engines drew on. This is the single most actionable list in the report: those are the pages to be present on.
• If an engine has no key configured on the platform, the insight text names it, so a missing platform row is an explained absence rather than a zero.

Brand Monitor. Mention volume, reach, sentiment score out of 100, share of voice, channel split and a mention trend. A sentiment score under 25 is a genuine reputation problem; 30% or more negative mentions is the same signal seen from a different angle. Share of voice matters more than raw mentions — 800 mentions is strong in a quiet category and invisible in a loud one.

Perception and Narrative Drivers. Perception characterises how the brand is talked about; Narrative Drivers ranks the stories doing the pulling, with a strength label and the gaps you should own but do not. Treat both as AI-written interpretation of measured coverage — the run card says so plainly. They are excellent for framing a communications plan and are not a substitute for the mention counts above.

Media Monitoring. Mentions over 30 days, reach, sentiment breakdown, outlet types, top outlets, mention trend, competitor coverage, press opportunities and crisis alerts. Read crisis alerts first, then press opportunities — the latter is a ready-made pitch list for Press Pitch Assistant.

Social Tracker. Followers, engagement rate, posting cadence and top posts on the platforms Metric Vault can measure: Facebook, Pinterest and Reddit, plus news and content signals. Instagram, X, YouTube and TikTok are not estimated — they need platform-specific API access, and the block says so rather than inventing follower counts. An empty Social Tracker block on a brand that is clearly active on Instagram is expected behavior, not a failure.

Strategic Insights and Recommended Action Plan. Above the blocks, a short insight set badged REAL · NO FABRICATION and a numbered plan, derived from the six payloads.

What good and bad look like together. Strong mentions with weak AI visibility means the web knows you and the models have not caught up — a structured-content and citation problem. Weak mentions with strong AI visibility usually means a narrow, well-defined niche where you are the obvious answer. Weak on both is a starting line; strong on both with negative sentiment is the most urgent case in this report.

Examples#

Screenshots#

Tips#

• Run this before Growth Actions. AI Visibility is one of the source tools the growth queue reads, so a Brand Health run unlocks a sharper queue.
• Read the sample excerpts before the score. The words are more useful than the percentage.
• Save the citation-sources list; it is your placement target list for the quarter.
• Re-run on a rival to see the same six measures for them.
• Quarterly is the right cadence. Brand signals move slowly, and re-running in the same week often returns cached figures for the full 10 credits.

Best Practices#

• Use a domain, not a common word. A brand name that is also an ordinary noun inflates the mention count with false positives.
• Track the AI visibility score over time rather than judging a single reading. The absolute number depends on the questions generated for that run.
• Pair a falling sentiment score with Narrative Drivers to find out why before responding.
• Treat Perception and Narrative Drivers as briefing material for a communications team, and Brand Monitor and Media Monitoring as the evidence base.
• Where the report finds a reputation gap in a competitor, take it to Competitor Intelligence for the structural picture around it.

Common Mistakes#

• Reading a missing engine row as a zero score. If an engine has no key configured, it did not answer; the insight text names it.
• Expecting Instagram, X, YouTube or TikTok follower data. They are excluded on purpose rather than estimated.
• Comparing visibility scores across different runs of different brands without noting that each run generates its own prompt set.
• Running it on a brand name with no web presence and reading the empty states as failures.
• Using Perception as a measurement. It is an interpretation; the numbers are in Brand Monitor and Media Monitoring.

Limitations#

• Six fixed tools, one brand per run.
• AI Visibility asks ten generated questions per run. It is a well-designed sample, not an exhaustive audit of everything an assistant might say.
• The engines queried depend on which providers are configured on the platform.
• Social coverage is limited to Facebook, Pinterest and Reddit plus news and content signals.
• Perception and Narrative Drivers are AI-written interpretations, labelled as such in the run card.
• Mentions come from an indexed content corpus, not from every site on the web.
• Each step is capped at 120 seconds; slower steps are dropped so the rest renders.
• Saved results are kept for 90 days.

Troubleshooting#

FAQs#

Which AI engines does it actually ask? ChatGPT, Gemini, Perplexity and Claude, plus Google's AI Overview, for whichever of those has a key configured on the platform. The KPI card reads LLMs Queried 4 — GPT · Gemini · Perplexity · Claude. Any engine that did not run is named in the result text.

Does it ask the models about my brand by name? No, and that is the point. The questions are category questions that never mention your brand, so appearing in the answer is a genuine signal rather than a prompted one.

Why did my visibility score change when nothing changed? Each run generates its own set of questions, and the models are non-deterministic. Read the trend across several runs, not a single number.

Is this the same as the Visibility Overview tool? Visibility Overview is one of the six steps. Running it alone costs 1 credit and gives you only the AI part. See Visibility Overview.

Can I schedule this? The workflow itself is run on demand. AI-visibility tracking with drop alerts is handled separately — see AI prompt tracking and Notifications and alerts hub.

Why is social coverage so limited? Because the alternative is fabricating follower counts. Metric Vault reports the platforms it can measure and states which ones it cannot rather than filling the gap with estimates.