Autocomplete demand signals: what they tell you and what they do not
Autocomplete is the list of suggestions that appears as you type into a search box. Every major marketplace and search engine has one, and it is built from what real people search. That makes it one of the most useful free signals a seller has - as long as you know what it can and cannot tell you.
What autocomplete tells you
- People search this phrase. Suggestions are built from real searches, so a phrase that appears is being typed.
- People search it in this marketplace. Amazon.co.uk and Amazon.de suggest different things, because their shoppers search differently.
- How shoppers describe the product. Suggestions show the words buyers use - often not the words a supplier's catalogue uses.
- Where interest is broad. A phrase suggested on Amazon, Google, YouTube and TikTok at once has interest well beyond one site.
What autocomplete does not tell you
- How many people search. A suggestion's position is not a count. Two phrases next to each other might differ a hundredfold in searches.
- That people buy. Search is interest; a sale needs a product, a price and a listing that convinces.
- That the interest lasts. Suggestions change with seasons, news and trends.
Any tool that shows a "search volume" built from autocomplete is guessing. When you see one, ask where the number comes from. The only real search counts for Amazon come from Amazon itself, in the Brand Analytics reports available to registered brand owners.
How to use it well
- Collect phrases across several sources and marketplaces, and note where each one appears.
- Watch them over weeks. A phrase that keeps appearing is a market; one that appears once is noise.
- Confirm with other evidence: the sales rank of the products already selling, their reviews, and how many sellers compete.
- Use buyers' own words in your title and bullet points - they are, quite literally, what people type.