Outerank Brain
Your daily SEO strategist — reads every module, picks the 3–5 actions worth doing today, then measures whether they actually worked.
What it is
Outerank Brain is your in-app SEO strategist. Every night at 04:00 UTC it takes a snapshot of every project — audit score, keywords in top 10, GEO citation rate, backlink count, internal-link health, content shipped — and compares to yesterday + niche-prior baselines. It then runs 74 rules (7 delta-based, 67 state-based) across 8 categories to surface the 3–5 highest-leverage actions for today: 'Audit score dropped 8 pts — open audit', 'You have 12 keywords ranking 11–20, push them to top-10', '50 orphan pages — fix the orphans', etc. Each action carries a predicted-impact score, a 95% confidence interval, a one-click CTA to the right module, and a 'why this works' rationale. Where the math is genuinely grounded — keyword-position and Search-Console-mined actions — the rec also carries a PROJECTED CLICKS forecast ('+120 organic clicks/month within 30 days'), and the day's plan rolls up into a trajectory strip ('act on these 3 → ~+380 clicks/mo'). When you mark it Done, Brain snapshots the metric baseline and re-measures at +7 / +14 / +30 days — including grading the projected clicks against your real clicks — so you can see whether the action actually moved the number.
Why it matters
Most SEO tools dump dashboards. Brain decides. Without it, every morning starts the same way: 'where do I even begin today?' With it, the answer is on screen the moment you log in — ranked by impact, gated by confidence, linked to the module that fixes the thing. And because Brain remembers what you marked done vs dismissed, it learns your style: rules you keep dismissing get down-weighted, rules where you got real lift get up-weighted. It also blends in cross-customer 'niche priors' so day-1 customers get day-300 wisdom for free. The combination is what makes the loop compound: do action → measure outcome → outcome refines the next rec → next rec is sharper.
How to use it (step-by-step)
Follow these in order. Each step is 30 seconds to 2 minutes.
Step 1
Open /modules/brain (or click the strip on Overview)
Brain has its own page with three tabs: Today, Scorecard, History. The compact strip on the Overview page is a doorway to the same recs.
Step 2
Read the Today tab
3–5 ranked actions for your active project. Each card has a title, body, predicted impact / 10 + confidence %, and a primary CTA to the module that fixes it. The 'why' button reveals the rationale (why this rule fired, why it works).
Step 3
Read the projected trajectory
When a rec is grounded in real numbers (keyword positions, search volume, or Search-Console query data), it shows a green '+N clicks/mo' forecast, and the day's plan rolls into a trajectory strip near the top: 'act on these moves → +X clicks/month within 30 days, from ~Y to ~Z today'. It's a conservative estimate — Brain grades it against your real clicks at +30 days.
Step 4
Take action
Click the primary CTA to open the module. Do the fix. Come back to Brain and click ✓ Done — that snapshots the metric so Brain can measure the outcome later.
Step 5
Use Snooze / Dismiss honestly
Snooze 3d hides a rec for 3 days then it comes back if still relevant. Dismiss removes it AND tells Brain you don't want similar recs (your personal learning weight for that rule drops by ~25%).
Step 6
Watch the Scorecard tab
After a few mark-dones + 7 days, the Scorecard shows your completion rate, your avg +7d / +14d / +30d lift, top measured wins, and which categories pay off most for YOUR project.
Step 7
Use the History tab for accountability
Filter by status (Done / Dismissed / Snoozed / Expired). Click any row to expand — you'll see the rationale + the measured outcome (baseline → +7d → +14d → +30d delta with %).
Step 8
On the Agency plan, open the Agency view
If you're on Agency, the 'Agency view →' button on the Brain page header opens a multi-client digest: every project on one scroll, critical-first ordered, top-3 open recs per project.
Every module feeds Brain.
Audit findings, rank deltas, backlink swings, GEO citation rate, content publishing — every module emits signals into Brain. The agent reads the full picture, not isolated charts.
- Audit + Tracker + Backlinks + GEO + Content all stream signals
- User survey shapes per-account weights (e.g. ecommerce → speed×1.4)
- Project + competitor data unified
- Free — no extra credit cost to feed Brain
Signals collected · today
Priority-ordered actions, not alerts.
Brain's recommendations are ACTIONS, not warnings. "Add Article schema to /blog/* (high ROI)". "Fix the 3 critical CWV issues on /products". Each one has a category, a target metric, and a believed impact %.
- Priority order — top 3 worth doing this week
- Category + target metric per rec
- Believed-impact % (an estimate, surfaced honestly)
- Snooze / dismiss / complete — your call
Brain · today's top 3
P1 · Fix 3 critical CWV issues on /products — +5pts Speed score.
P2 · Add Article schema to /blog/* — +12% GEO citation rate.
P3 · Internal-link /lookbook from 5 hubs — fix orphan.
Outcomes measured — accuracy graded.
When you mark a recommendation "done", Brain measures the actual metric movement 7, 14, 30 days later. The Scorecard tab shows Brain's accuracy across your real outcomes — a number you can show clients.
- Real outcome measurement (not self-reported)
- 7d / 14d / 30d lift % per recommendation
- Per-category accuracy (which areas Brain nails)
- Top wins surfaced for the case-study screenshot
Brain accuracy · per category
The recommendation engine learns.
Recommendations that consistently land high lift get prioritized for similar projects. Recommendations that flop drop in weight. The next time Brain runs, it's slightly smarter because of YOUR outcomes.
- Per-user weights shaped by survey + outcomes
- Per-niche calibration (ecommerce vs SaaS vs local)
- Cross-project pattern detection (compound knowledge)
- Transparent: each rec's weight is auditable
Compound knowledge · niche signals
Why Outerank
What other "AI SEO tools" can't prove worked.
Brain measures its own accuracy on your real outcomes. The Scorecard tab IS the proof.
Every module is a signal source
Audit + Tracker + Backlinks + GEO + Content all feed Brain. The agent sees the whole board, not isolated charts.
Actions, not alerts
Recommendations are tasks — not red badges. "Add Article schema to /blog/*" is something you DO.
Outcomes measured for real
When you mark a rec done, Brain checks the actual metric 7/14/30 days later. The Scorecard tab is your accuracy.
Survey + outcomes shape weights
Ecommerce sites get speed weighted up. SaaS gets backlinks weighted up. The weights are auditable, not a black box.
Niche-wide compound knowledge
What works for similar projects feeds your recs. The platform gets smarter as more customers ship outcomes.
Honest about uncertainty
Brain shows believed-impact %, not promises. When confidence is low, it says so. No oversold predictions.
What costs credits
No surprises. Browsing and viewing saved results is always free.
Pro tips
- →Brain's morning email lands at 05:15 UTC for any project with a CRITICAL anomaly in the last 24h — don't ignore those. Critical means cohort-significant, not just 'slightly worse'.
- →Recs auto-expire when the underlying rule no longer fires (you fixed it). You don't need to manually clear stale recs.
- →The 95% confidence interval is honest — wide intervals mean 'we're not sure how big the lift will be'. Treat them as ranges, not predictions.
- →If a rule_key seems wrong for your business, dismiss it consistently — your personal-learning weight will push that rule below 0.5×, effectively turning it off for you while keeping it on for users it helps.
- →The 'algorithm-shift' rec (rare; fires only when cohort-wide |Δposition| spikes 2.0× baseline) means a likely Google update — don't panic-edit your site that week. Observe for 48h.
Plain-English glossary
Every bit of jargon on this page, explained.
- Rule key
- The internal name of a Brain rule (e.g. 'audit_low_score', 'kw_close_to_top10'). Visible in the rec's evidence panel.
- Predicted impact
- Brain's estimate of how big the win will be on a 0–10 scale, shrunk toward your niche prior via Bayesian update.
- Confidence interval (95% CI)
- The range Brain is 95% sure the lift will land in. Wide CIs = Brain isn't certain how big the win will be.
- Outcome ledger
- The database row created when you mark a rec done. Stores baseline + future +7d / +14d / +30d values + computed deltas — and, for forecasted recs, the projected vs realized clicks.
- Projected trajectory
- A conservative projected-clicks-per-month forecast, shown only on recs where the math is grounded (your tracked positions × search volume, or live Search Console impressions × the CTR-by-position curve). Never an AI guess. Brain grades it against your real clicks at +30 days — so the promise is accountable, not marketing.
- Counterfactual lift
- Actor-cohort lift MINUS dismiss-cohort lift — the part of the lift we can credibly attribute to YOUR action (not the SEO weather).
- Personal learning weight
- A multiplier between 0.5× and 1.5× applied per-user per rule, based on your done-rate over the last 90 days.
- Algorithm-shift signal
- When |Δposition| pools 2.0× over the 7-day baseline across 30%+ of customers, Brain fires a critical 'observation week' rec.
Questions, answered
Site Audit FAQs
Will Brain ever pop the same rec twice?
Why are some Overview cards showing 'all clear'?
How accurate are the predicted impacts?
Does Brain ever auto-publish or auto-fix things?
Can I see Brain's full math on /trust?
What's the difference between Brain and Forge?
How we compare
Honest comparison — where we win, where the big tools win.
vs. Ahrefs / Semrush dashboards
They show data. Brain shows DECISIONS. Both are fine — the difference is whether you spend the first 30 mins of your day reading numbers or doing work.
vs. ChatGPT with custom GPTs
A custom GPT can advise IF you paste your data. Brain reads your modules directly + tracks the outcome — so the next rec is informed by whether the last one worked.
vs. Hiring an SEO consultant
Consultants are smarter, but they're $200/hr and don't see your daily data. Brain is cheaper, faster, and never tired.
Ready to try Outerank Brain?
Open it on one of your projects.
Come back here whenever you need a refresher.