Updated 2026-08-02
How to read a Cerebro export: every column explained
A Cerebro export (June 2026 format) has 15 columns: Keyword Phrase, Keyword Sales, Cerebro IQ Score, Search Volume, Search Volume Trend, three H10 PPC bid columns, Sponsored ASINs, Competing Products, CPR, Organic, Title Density, Smart Complete, and Amazon Recommended. Here is what each one means for niche research — and the formatting traps that corrupt spreadsheet analysis before you even start reading.
What does each column mean?
| Column | What it is | How to read it for niche research |
|---|---|---|
| Keyword Phrase | The search as typed by buyers | Your raw material — never translated, never "cleaned" |
| Keyword Sales | Estimated monthly units sold attributed to the keyword | Demand that actually converts (complements volume) |
| Cerebro IQ Score | Demand-to-supply ratio (volume vs competing products) | Quick flag, not a decision |
| Search Volume | Estimated monthly searches | The demand axis — read it at family level, not per phrasing |
| Search Volume Trend | 30-day change in search volume | Recent momentum only — not seasonality |
| H10 PPC Sugg. Bid (min/max) | Suggested PPC bid range | The price of attention on this phrase |
| Sponsored ASINs | Count of sponsored products seen for the keyword | How paid-contested the phrase is |
| Competing Products | Number of products matching the phrase | Supply axis — beware the floors (see below) |
| CPR | Helium 10's estimate of units needed in ~8 days to reach page 1 | Cost of entry for launch, family minimum is the door |
| Organic | Numeric signal from organic ranking data | Mostly useful as provenance (see below) |
| Title Density | Page-1 titles containing the exact phrase | The titling axis — full guide here |
| Smart Complete | Keyword surfaced via Smart Complete | Provenance flag — tends to be on-topic |
| Amazon Recommended | Keyword from Amazon's recommendations | Provenance flag — tends to be precise long-tail |
The three provenance columns deserve their own guide: Smart Complete, Organic, Amazon Recommended — the relevance prior nobody reads.
Which formatting traps break spreadsheet analysis?
We have parsed dozens of real Cerebro exports. These five traps appear in essentially all of them:
- Floors disguised as numbers. Competing Products can read
>10,000— a floor, not a value. A spreadsheet either coerces it to text (excluded from every numeric filter) or you strip the>and quietly turn "at least 10,000" into "exactly 10,000". - Thousands separators as text.
1,482imports as text in many locales; sums silently ignore those cells. - Empty markers.
-andN/Amean "unknown", which is not zero. Averaging them as 0 drags every family metric down and can rank an unknown above a measured value. - Invisible characters. Real exports contain BOM markers and invisible Unicode (we have seen U+FFFC leading dozens of phrases in a single export). Two visually identical keywords then count as different — and your deduplication misses them.
- The same demand, many spellings. Not a formatting trap but the structural one: plurals and word order fragment one demand across many rows. Collapsing them into families is step one of any honest reading.
What should you actually do with the export?
Read it in this order: collapse into families → provenance triage → demand-vs-Title-Density outliers → verify on page 1. The full sequence is in You ran Cerebro. Now what?
Frequently asked questions
Why does my Cerebro export show ">10,000" instead of a number?
Some columns are floors, not measurements: ">10,000" competing products means "at least 10,000". If your spreadsheet coerces it to text or drops the row, every average or filter built on that column is silently wrong. Treat floors as their own category.
What is a good Cerebro IQ Score?
IQ Score is Helium 10's demand-to-supply ratio — search volume relative to competing products. Higher means more demand per competitor at a glance. Use it as a first-pass flag, never as a decision: it inherits every flaw of both inputs, including competing-products floors.
Which columns matter most for niche research?
For finding niches (rather than optimizing PPC): Search Volume and Keyword Sales for demand, Title Density for how contested the exact phrase is, Competing Products for supply, and the three provenance columns — Smart Complete, Organic, Amazon Recommended — as a relevance filter.