The question asked most often when choosing a keyword tool is "which one is most accurate?"
It is hard to answer, and not because the tools are too close to call. It is hard because the "search volume" two tools display is often not the same quantity at all. One shows how many times the term was searched across the whole marketplace in a month; another shows how much exposure the term brought one product this week. The two can differ tenfold and neither is wrong.
So the first step in choosing a keyword tool is not comparing numbers. It is working out what each number is.
Four fields in one response that look like search volume
The traffic keyword reverse lookup returns the terms an ASIN has appeared for in the first three pages of search results. Each row carries four fields that all look like search volume:
| Field | What the docs say | Period | Scope |
|---|---|---|---|
items.searches | Monthly searches: total searches for the term on the marketplace | Calendar month | Whole marketplace |
items.searchesRank | Weekly search rank: the ABA search frequency rank, smaller is hotter | Week | Whole marketplace; a rank, not a count |
items.calculatedWeeklySearches | Estimated weekly exposure the term brought this product — not the term's total searches | Week | One product |
items.impressions | Total impressions of all ASINs on the term's results page, not one ASIN's | Calendar month | The whole results page |
Four fields, three periods, three scopes, and one of them is a rank rather than a count.
That table is the answer to "which is most accurate": first ask which row it shows. If tool A's "search volume" corresponds to searches and tool B's to calculatedWeeklySearches, no gap between them says anything about accuracy.
One more detail: the periods within a single row do not agree either. items.keyword covers terms that reached the first three pages "in the last 30 days or in a given calendar month", while searches in the same row is a calendar-month figure and searchesRank a weekly one. Before comparing anything in one table, decide which period is the baseline.
Same field name, possibly a different type
The second trap people miss is type.
searches is documented as a string in the traffic keyword reverse lookup and as an integer in the order keyword reverse lookup. Same name, same meaning, different type across two endpoints.
It does not change how you read the value, but it does change how you merge. Join the two endpoints on keyword and one column is text while the other is numeric, and sorting quietly breaks — under string ordering "9800" sorts ahead of "12000". Converting to numbers before merging is the step most often missed when building a keyword library. The full merge process is in building your own keyword library.
Match tools to tasks instead of looking for one that does everything
"Keyword tool" is really an umbrella for several kinds of tool. They answer different questions from different data, and one tool rarely does every kind well. Split by task:
| The question you need answered | The data to use | Docs |
|---|---|---|
| What is this market searching for | ABA weekly and monthly search term ranks | ABA weekly ranks |
| Which terms have I not thought of | Keyword mining and keyword expansion | Keyword miner, Keyword expansion |
| Which terms a competitor gets exposure on | Traffic keyword reverse lookup | Traffic keyword reverse lookup |
| Which of those terms actually convert | Order keyword reverse lookup | Order keyword reverse lookup |
| Whether the term is heating up off the marketplace | Search trends | Google keyword trends |
The table also explains why "one tool is best" is an unreliable claim: demand-finding tools win on coverage, reverse lookup tools win on precision for a single ASIN, and those two goals pull in opposite directions.
The practical approach is to work out which one or two of these you actually do every week and choose tools for those alone. How to evaluate reverse lookup tools item by item is covered separately in choosing a reverse ASIN lookup tool.
Three questions to ask when evaluating
Whatever tool you trial, run the same set of terms through it and ask three questions. They surface most of the differences:
1. Is this search volume for the whole marketplace or for one product? This is the biggest gap. Marketplace-wide searches and the exposure one product receives can be an order of magnitude apart. A tool labelling the latter as "search volume" is the most common source of misreading.
2. Is the period a calendar month, a rolling 30 days, or a week? Numbers with different periods cannot be compared or trended against each other. If a tool cannot tell you its period, its numbers are good for magnitude only.
3. Is it a rank or a count? ABA supplies a search frequency rank, and a rank only tells you relative position. For any number that converts a rank into a "count", ask who converted it and how.
Only a tool that can answer all three has numbers worth putting in your sheet.
Three things this data cannot do
It cannot give you your backend search term report. How many clicks and how much spend each search term drove in your own ads is authorized seller account data, and no public-side keyword tool can obtain it.
It cannot tell you what happens once a term goes into your listing. The data says how many people search a term and how much exposure competitors get on it. It cannot predict where you will rank once you use it. The only check is rank monitoring after launch.
It cannot judge whether a term is relevant to your product. However large the search volume, a term unrelated to your product brings only wasted exposure. Relevance is your call.
Common questions
Are free keyword tools usable? For magnitude and ideas, yes. But free tools rarely state their period and scope, so treat their numbers as reference only and keep them out of any sheet that needs month-over-month comparison.
Different tools give different search volumes. Which do I trust? First confirm they show the same quantity (the four-field table above). If they do, compare them; if they do not, there is nothing to compare. For ongoing use, sticking to one source matters more than chasing the most accurate one.
Can a keyword tool tell me which ad terms to bid on? It supplies competition and conversion data; the decision is yours. How to filter ad terms is in choosing ad keywords.
Will a term start converting once it reaches page one? Not necessarily. Clicks and conversions on page one often concentrate in the first few positions, which is covered in getting a keyword onto page one.