Amazon keyword tools compared — and when you still need a human

Amazon keyword tools compared: prices and reverse ASIN

Almost every comparison you will read says pricing "varies by plan". Here are the actual numbers, each read off the vendor's own pricing page in September 2026. Where a vendor does not publish a figure, this says so rather than guessing.

Two things in that table are worth more than the rest of this article.

The first is Sonar. It was the free keyword tool everybody recommended, it came from Sellics, and after Perpetua absorbed Sellics the page now carries a live notice reading "We’re sorry, but this feature is no longer available". It is still on page one of several "best free Amazon keyword tool" listicles. If a guide is recommending it in 2026, that guide has not been checked since it was written.

The second is Perpetua, which most comparisons still file under keyword tools. It is now an ad management platform starting at $695 a month with automated keyword harvesting as a feature of that. Useful, but not comparable to a $39 research tool, and putting the two side by side without saying so misleads people about what they are choosing between.

The free tools, including the one you are already paying for

Before you buy anything, use what Amazon gives you. Three data sources sit inside Seller Central and are better than most paid estimates, because they are not estimates.

Brand Analytics Top Search Terms gives you the actual search frequency rank for a term, plus the three most-clicked ASINs and their share of clicks and conversions. That last part is the bit no third-party tool can replicate: it tells you not just that a term has volume, but whether that volume converts and to whom. It requires Brand Registry.

Search Query Performance shows, for your own ASINs, the queries that produced impressions, clicks, adds to basket and purchases, with your share of each. This is the report that tells you whether a keyword change worked, and it is the single most underused thing in Seller Central.

The advertising search term report shows exactly what shoppers typed before clicking your ad. It is the only source in the list that reflects real spoken demand rather than a normalised index. Every Sponsored Products account generates it, and most sellers never open it.

Between them these cover most of what a paid keyword tool sells you. What they do not do is show you someone else’s keywords, which is the one job that genuinely requires a third-party tool and the reason Cerebro and its equivalents sell.

There is a fourth free source most people forget: Amazon’s own search bar. Type the first two words of your category and read what autocomplete suggests, then repeat with each letter of the alphabet appended. It is tedious and it is real demand data straight from the shopping interface, unmodelled and current. The paid tools scrape exactly this and sell it back to you in a table. For a single product launch, twenty minutes of this beats a month of subscription.

The same goes for the related searches strip and the "customers frequently viewed" carousels on a competitor’s detail page. Both are Amazon telling you, in public, which products it considers substitutes for each other — which is the substitution set your keyword list has to cover.

The genuinely free third-party options

Helium 10 has a free tier and a free keyword tool page. SellerApp publishes a free reverse ASIN tool with no price on the page. Beyond that, "free" usually means a capped trial. That is fine for a one-off check and useless for ongoing work, which is the honest reason most sellers end up paying for something.

Helium 10 or Jungle Scout?

This is the actual decision most people are making, and the answer depends less on features than on what you are for.

Helium 10 is the deeper keyword instrument. Cerebro separates organic rank from sponsored rank on the same keyword, shows title density, and compares up to ten ASINs at once. Those three things together are why agencies default to it: the organic-versus-sponsored split tells you whether a competitor earned a position or bought it, which changes whether the keyword is worth chasing at all. Platinum is $129 a month billed monthly, or $99 billed yearly. Note that Helium 10's own Cerebro page and its pricing page currently show different figures for the same plans — check which you are being charged before you assume.

Jungle Scout is broader and cheaper at the entry point. It reorganised in 2025–26 into Catalyst for sellers under roughly $1M and Cobalt for enterprise, and Keyword Scout is gated by plan: 50 searches a month on Starter, 250 on Growth Accelerator, unlimited on Brand Owner. If you are still choosing products as well as keywords, Jungle Scout does both jobs in one subscription and Helium 10 does not do the first one as well.

Both now ship an MCP integration that exposes account data to AI assistants, which is a real convenience and not a reason to pick either.

The two specialists worth knowing

Data Dive at $39 is the tool agencies use when they want the full keyword matrix across a set of competitor ASINs rather than a pretty report. It calls the operation a "Dive" rather than a reverse ASIN lookup, which is why it is missed in comparisons, but ASINs go in and a master keyword list comes out.

SmartScout at $49 is not really a keyword tool. It is brand and seller-level data — who owns what, which brands are growing, where the category's revenue actually sits. If your question is "which competitor should I be studying", it answers that better than any keyword tool, and then you take those ASINs to Cerebro.

What these tools actually measure

Every one of them shows you five or six columns, most sellers sort by the first one, and the first one is usually the least useful. It is worth knowing what each column is claiming.

Search volume. An estimate of how many times a phrase was searched in a period. Modelled, not measured — see the next section. Useful for ranking candidates against each other, dangerous as an absolute.

Organic rank versus sponsored rank. Where a given ASIN sits in results for that keyword, separated into the position it earned and the position it bought. This is the most decision-relevant column in any keyword tool and the one sellers ignore most often. A competitor sitting at organic position 3 has a listing and a sales history you would have to out-compete. A competitor at sponsored position 1 and organic position 40 is renting the spot, and the keyword is more open than it looks.

Title density. How many of the products on page one carry that exact phrase in their title. Low title density on a term with real volume is the clearest opportunity signal in the whole discipline: demand exists and nobody has claimed the words. High density means you are late, and winning the term will take more than a listing edit.

Competing products. How many listings return for the phrase. On its own it is close to meaningless — a term with 40,000 competing products and no title density is easier than a term with 800 and full density. Read it against the other two columns or not at all.

Relevancy or opportunity scores. Each vendor has one, each calculates it differently, and none will tell you how. They are reasonable first-pass sorts and a poor basis for a decision you have to defend to a client or a finance director. Treat a proprietary score as the tool’s opinion, not as data.

Indexed or not indexed. Whether a given ASIN currently appears in results for the phrase at all. This is the fastest way to check whether a backend keyword change took, and it is the column to reach for after you edit a listing rather than before.

If you read only two of these, read organic-versus-sponsored rank and title density together. They answer the only question that matters at the research stage: is this term winnable by someone with my resources, or am I about to spend three months losing to a brand that got there first.

How accurate is any of this?

Less than the interface implies, and the direction of error matters more than the size.

Amazon does not publish keyword search volume to third parties. Every number you see in Helium 10, Jungle Scout or anywhere else is modelled — inferred from sales rank movement, clickstream panels, advertising data and sampling. Different vendors model it differently, which is why the same keyword can show materially different volumes in two tools open side by side on the same screen.

The practical consequences:

• Relative volume is reliable, absolute volume is not. If tool A says a term gets four times the searches of another term, trust the ratio and ignore the raw figures.
• Accuracy degrades outside the US. Panels are thinner in smaller marketplaces, so a term with a confident-looking number on amazon.com may be close to guesswork on amazon.co.uk or amazon.de.
• Long-tail estimates are the least trustworthy and, awkwardly, the most commercially interesting. A term modelled at 200 searches a month might be 40 or 600.
• Brand Analytics beats all of it for terms it covers, because search frequency rank is Amazon's own measurement rather than somebody's model of it.

The right use of a keyword tool is to build a candidate list and rank it. The right use of Amazon's own reports is to decide which candidates were correct. Sellers who skip the second step end up with beautifully researched listings optimised for demand that does not exist in their category.

A cheap way to sanity-check any tool

Pick three keywords you already rank well for and already have data on from Search Query Performance. Look them up in whichever tool you are evaluating and compare. You are not checking whether the numbers match — they will not. You are checking whether the tool puts them in the same order as your real impressions data, and whether it gets the gaps roughly right. A tool that inverts the order of three terms you already understand will do the same on the terms you do not, and no amount of feature list makes up for that.

Run the same three through a second tool before you commit to an annual plan. Where two independent models agree, you can act with reasonable confidence. Where they disagree by a factor of five, treat the term as unknown and let advertising settle it — a fortnight of exact-match spend on an uncertain keyword costs less than a year of the wrong subscription and produces better data than either vendor.

When you still need a human

Four judgements sit outside every tool in the table, and all four decide whether the research was worth doing.

Relevance, which is not the same as volume

A tool will hand you "waterproof phone case" for a rugged tablet sleeve because competitors rank for it and the string overlaps. Index for it and you buy clicks from people who wanted something else, which depresses your conversion rate on that term and, because conversion rate is a ranking input, quietly damages your position on the terms you do want. Irrelevant traffic is not free. Nothing in any tool flags this, because relevance is a claim about the product, and the tool has never seen the product.

What the words actually mean to a buyer

"Heavy duty" and "industrial" can be near-identical in a keyword export and describe two different shoppers with different price expectations. So can "organic" and "natural", which are also a compliance question rather than a copy question in several categories. A tool sorts by volume. A person asks which word the person with the credit card would use.

Where the keyword goes, now that titles are 75 characters

This is newly hard. Amazon cut product titles to 75 characters in all categories except media on 27 July 2026, and introduced a 125-character Item Highlights field that began displaying beneath the title on 10 August. Amazon states both are inputs for search and neither is prioritised over the other.

That turns keyword placement from a listing exercise into a genuine allocation problem: you now have far less prime real estate and a second field whose weighting nobody outside Amazon can measure. No tool has an opinion on which of your twelve priority terms earns one of the 75 characters. That is a judgement about your category, your brand, and what a shopper scanning a results page needs to see in the first four words.

Knowing when to stop

Keyword research has a point of diminishing return that arrives much earlier than the tool's export suggests. A thousand-row keyword list is a research artefact, not a plan. Most listings are carried by between eight and twenty terms, and the work that matters is deciding which, then measuring whether you were right. Our keyword research service ends with a mapped set per ASIN rather than an export, for exactly this reason.

Which one should you actually buy?

If you want an outside read on which of those rows you are actually in, our free listing checker scores a live ASIN in about a minute, and a written audit comes back within 48 hours with the keyword gaps named rather than scored.

Questions people also ask