Everything so far starts from one idea: let traffic in and remove the bad stuff. There are businesses where that idea is not enough, because the bad stuff is infinite and the good stuff is a short list. For them there is the inverse approach: the white list. You define the vocabulary you do want and everything else is negativised automatically.
How it works
- You define a white list of acceptable words (roots): your product, its synonyms, its variants, the buying modifiers.
- Every incoming search term is compared against the list: if all its words (or the significant ones) are on the list, it passes; if it contains any word that is not, the term is negativised (usually in exact, so only that term is cut).
- The negativised terms are reviewed periodically: the ones that did make sense reveal words missing from the white list.
It is a fine sieve: at first it blocks too much; with each review, the white list grows and the sieve settles down.
When it is the right choice
| Situation | White list? |
|---|---|
| A very specific product with a closed vocabulary (one model, a regulated service, a spare part) | Yes: the good stuff can be enumerated |
| Broad match with Smart Bidding and a lot of noise | Yes: it tames broad without giving it up |
| Sectors with ambiguous searches (words with several meanings) | Yes: only the combinations with your meaning get through |
| A wide, fast-changing catalogue (general ecommerce) | No: the white list cannot be maintained |
| A new account where you still do not know how people search for you | Not yet: learn with classic negatives first |
| You want to discover new terms | No: by definition, the white list blocks the unknown |
How the white list is built
- From your conversions: every word from the terms that have converted in the last 12 months. It is the seed, and it is objective.
- From your keywords and their natural variants.
- Buying modifiers: buy, price, quote, hire, online, near me, urgent, 24 hour...
- Geography: your cities and areas.
- Function words: of, for, with, the, in... (either ignored in the comparison or included as standard).
- Your brand and the brands you sell.
A sensible white list has between 100 and 500 words. If it needs 5,000, the business is not a white-list business.
Risks and how to mitigate them
- Blocking real demand you did not know about: the main risk. Mitigation: review the blocked terms with high impressions every week; if they make sense, add their vocabulary to the white list and remove the negative.
- Vocabulary that changes (a new product, a trending term): the white list is updated with every catalogue change.
- Overconfidence: a white list does not fix a bad landing page or broken measurement; it only guarantees that the traffic is on topic.
- Conflicts: an exact negative generated by the white list can clash with a new keyword you add later. Periodic cleaning of conflicting negatives is compulsory.
Combining it with scenarios
The two systems complement each other: the white list guarantees the topic; the performance scenarios (previous lesson) clean up, within the topic, whatever spends without converting. Order of application: white list first (it filters intent), scenarios afterwards (they filter performance), and the usual safeguards: whatever has converted is not touched, the brand is protected.
💡 Ninja trick: the Suite's Negative Keywords (White List) is this method: a white list in its sheet, a nightly comparison of every term, exact negatives in shared lists, a "blocked with impressions" tab for reviewing and rescuing, and a Phase 2 clean-up that removes negatives conflicting with keywords. And the account pattern: Negative Keywords for the topic, SQONS for the performance, both at the same time.
What you should remember
- White list = you define what is allowed; everything else is negativised.
- For closed vocabularies, broad with Smart Bidding and ambiguous sectors. Not for wide catalogues or accounts that are still learning.
- The seed: the words from your conversions; a healthy size is 100-500.
- Main risk: blocking new demand → review the blocked terms every week.
- It combines with the scenarios: topic first, performance afterwards.