We Asked Five Etsy Keyword Tools the Same Question. They Disagreed by 24,000x

We may earn a commission when you buy through links on this page, at no extra cost to you. Tools can’t pay for placement or scores. How we make money · How we test

Data collected July 16, July 22 and July 30, 2026 · Five tools, fresh free accounts, the same seed keywords, every number transcribed off the screen. Raw data and method are in full below, and the dataset is free to cite.

Every Etsy seller is told to check the search volume of a keyword before making anything. So we asked five keyword tools the same question on the same day: how many people search Etsy for “personalized dog mom gift” each month?

eRank said fewer than 20. Alura said 481,900. That is not a rounding difference. It is the same question, the same day, and a 24,000x gap between two products that thousands of sellers pay for.

We then did the thing nobody seems to do: we checked whether any of these numbers can be corroborated. One number in these tools can be independently verified, and it turns out to be the one they all agree on. Everything they disagree about is the part no one can check.

The headline: 1.68x versus 24,000x

Etsy keyword tools report two numbers side by side, in the same font, with equal confidence.

Competing listings is a count. Somewhere there is a real number of live listings matching a phrase, and anyone can go and count it. Search volume is a model. Etsy does not publish search counts to anybody, so every vendor estimates it privately and never says how.

Put the two side by side across our five same-day keywords and the difference is not subtle:

Read the bottom-right cell again. On the keyword where the tools agreed most closely, they were still 59x apart. Their best case is worse than the countable metric’s worst case by a factor of thirty-five.

The same five keywords, three tools, one day

Reported monthly Etsy searches. Note the axis: each step is 10x.

Every search volume we recorded

Five keywords run through four tools on the same day, July 16, 2026, on free accounts. EHunt is in the table because its answer is itself a result.

“none” means the tool returned no exact-match figure for the keyword we searched. EHunt answers a different question by design: it returns a database of related keywords rather than statistics for your phrase, so on three attempts our own search term never appeared as a row. We count that as a structural finding rather than a missing data point.

Two tools later ran the full ten-keyword set. Here they are head to head, and this is where it stops being about scale and starts being about shape:

Look at the bottom row. eRank measured “dad joke mug” at under 20 searches a month. Sale Samurai reported 30,875, its single highest figure across all ten keywords, higher than “custom name necklace” and “custom pet portrait.” A novelty phrase came out ahead of a genuine head term.

Note also what eRank’s column does that Sale Samurai’s does not. eRank spans “under 20” to 5,298, a range of more than 265x. Sale Samurai spans 10,398 to 32,922, a range of 3.2x, with nothing below ten thousand. Real search demand is power-law distributed. One of those columns has the shape of real demand and one does not.

The control: competing listings

Now the same five keywords, same tools, same day, on the metric that can actually be counted.

Every keyword within 1.68x, most within 1.4x, and the ordering identical in all three tools. Sale Samurai, tested two weeks later, lands in the same band: its counts sit 1.07x to 1.71x above eRank’s with the ordering unchanged, and part of that gap is simply listings added in the fourteen days between tests.

This is the whole study in one comparison. Hand four vendors a question with a checkable answer and they all get roughly the same result. Hand them a question nobody can check and they return numbers three to four orders of magnitude apart, formatted identically, with no error bars and no methodology note anywhere in any of the four products.

The advice that fails: “just compare within one tool”

The standard defence of these tools, and one we have made ourselves on this site, is that the absolute numbers do not matter. Pick one tool, stay inside it, and use the volumes to rank keywords against each other. Relative comparison, not absolute truth.

Our ten-keyword dataset lets us test that claim directly, and it only half holds.

Here is where each tool ranks the same five keywords, 1 being the highest volume that tool reported:

eRank returned “under 20” for three of these five, so those three are tied at the bottom rather than separately ranked.

The top of the list is trustworthy. The rest is not. All four tools put “custom name necklace” first or second and “digital planner goodnotes” second or third. If you only need to know which of your candidate keywords is the biggest, any of these tools will tell you, and they will agree.

Then look at “personalized dog mom gift.” It is Alura’s number one keyword out of five. It is Sale Samurai’s last. EverBee puts it in the middle. eRank says it cannot measure it at all. The same phrase, on the same platform, ranked first and last by two tools a seller might reasonably choose between.

So the safe version of the advice is narrower than the one everybody repeats, including us until today: use one tool to identify your strongest keyword, and do not trust any of them to rank the weak ones. The weak end is exactly where the tools stop agreeing, and unfortunately the weak end is where most new sellers are looking, because that is where competition is low enough to enter.

Why the numbers are made: the permutation tell

There is a visible artifact that shows these volumes are generated rather than measured, and we found it independently in three of the five tools.

Shuffle the words in a phrase. Real search queries are distinct strings, so “custom name necklace” and “name custom necklace” are different queries with different real demand, and a measured number would show that. Here is what the tools show instead.

Sale Samurai, on the three words custom, name and necklace:

One reordering comes back byte-identical to the seed. Another, “name necklace custom,” comes back 41% lower. Word order either matters to this model or it does not, and here it does both at once. Elsewhere in the same product, “pdf crochet pattern” reports 33,624 against 10,398 for “crochet pattern pdf,” a 3.2x gap for a phrase no human types.

Alura shows the same artifact more baldly: “personalized dog gifts for mom” and “personalized gifts for dog moms” both report exactly 458,954.

EverBee shows it with a twist that gives the game away. For “personalized dog mom gift” its related list offers “mom gift dog person,” “dog person gift mom” and “dog mom person gifts,” all sitting at roughly 4,206 to 4,218 volume. Nearly identical volumes. But their competing-listing counts range from 15,858 to 151,938, a tenfold spread across the same words reordered.

That last one is the cleanest evidence in this study. Within a single product, on a single screen, the modeled number stays flat across permutations while the counted number varies by 10x. The counted number is responding to reality. The modeled one is responding to a bag of words.

Who is actually searching

There is a reason to think the low numbers are closer to the truth than the high ones, and it comes from the one panel in this category that reports something checkable about demand: eRank’s Searchers by Country breakdown.

Etsy keyword tools cannot tell a buyer typing “personalized dog mom gift” from a print-on-demand seller in another country researching what to list. Both are searches. Both inflate volume. Geography is the closest available proxy for telling them apart, because genuine buyer demand for US-market giftware concentrates in the US, UK, Canada and Australia.

Now notice which keyword this lands on hardest. “personalized dog mom gift,” the phrase at the centre of the 24,000x disagreement, is 38.5% Vietnam and 23.1% US. The tool reporting 481,900 monthly searches for it is, on this evidence, largely counting other sellers doing exactly what you are doing.

And “gothic home decor” is the counter-example that makes the point. It has the highest competition of anything we tested at 468,689 listings, which on any other reading says avoid. But 68.6% of its searchers are in the US, and its trend line ramps cleanly into September and Halloween. High competition with real buyers behind it is a completely different proposition from low competition with none.

What to do instead

None of this means the tools are useless. It means you should use the parts that survive scrutiny.

• Trust the competing-listing count. It is a real count, the three tools we ran on the same day agree on it within 1.68x with a fourth landing in the same band two weeks later, and it answers the question that actually matters when you are entering a market: how many people are already selling this?
• Use search volume only to find your single strongest candidate, and only within one tool. Do not use it to rank your weak or mid candidates, because that is precisely where the tools contradict each other.
• Never quote an absolute volume as fact, in a blog post, a course, or a decision to spend six weeks making something. There is no version of “481,900 people search this” that is defensible.
• Treat a tool that admits ignorance as more credible, not less. eRank returning “under 20” reads like a failure and is the opposite. On the same keyword, on the same day, three other tools produced a confident number instead.
• Check who is searching if your tool will tell you. A keyword whose searchers are mostly outside your selling market is other sellers researching, not demand.
• Sanity-check any volume against the phrase’s own subsets. If a three-word phrase reports more searches than the two-word phrase inside it, as “dad joke mug” did against “dad joke,” the model is not measuring queries.

Method, limits and what would change our mind

What we did. Fresh free-tier accounts on each tool, created for testing, no paid plans. The same seed keywords entered into each product and every figure transcribed from the screen rather than from vendor marketing. eRank, EverBee, Alura and EHunt were run on July 16, 2026, within the same session. eRank’s second batch ran July 22 because its free plan allows five lookups a day. Sale Samurai ran July 30, through its Chrome extension on live Etsy, because its web app research tools are paywalled on the free tier.

Limits, stated plainly.

• Five keywords are common to four tools, not ten. Only eRank and Sale Samurai have the full ten-keyword set, because free-tier caps are real and we do not pay for tools.
• Sale Samurai’s data is fourteen days later than the rest. Search volumes and listing counts both drift. That is why we hold it apart in the same-day tables and note it wherever it appears.
• We tested free tiers. It is possible paid tiers use better data, though no vendor claims this and the extension and web numbers we could compare did not differ.
• “Under 20” is a floor, not a value. Every spread involving it is a lower bound, so 24,000x is the smallest the true gap could be.
• We cannot prove which tool is right. Nobody outside Etsy can. This study shows the tools cannot all be right, which is a weaker claim and a fully supported one.

What would change our mind. A vendor publishing its methodology and error bars, or Etsy releasing real query data. If any tool in this table wants to show its working, we will run the test again and publish the result whichever way it falls. We will also re-run this study as the tools change; the date on every figure is there so you can tell when it goes stale.

Cite or reuse this data

This dataset is free to use, including commercially, with a link back to this page. No permission needed. If you are writing about Etsy tool accuracy, take the tables.

Suggested citation: SellerTested (2026). We Asked Five Etsy Keyword Tools the Same Question. They Disagreed by 24,000x. https://sellertested.com/etsy-search-volume-study/

If you spot an error, or you work at one of these companies and think we have measured something wrong, write to hello@sellertested.com. Corrections get priority and are published with a dated note.

FAQ

Which Etsy keyword tool has the most accurate search volume?

Nobody can prove accuracy from outside, because Etsy does not publish search data. What we can say is which tool behaves most like a measurement. eRank reports “under 20” when demand is too low to measure instead of generating a number, and its volumes span more than 265x across our ten keywords, which is the shape real demand has. Sale Samurai’s ten volumes fit inside a 3.2x band with nothing under 10,000, which is not. Alura’s are the largest by far and look Google-scale rather than Etsy-scale.

Why do Etsy keyword tools show such different search volumes?

Because none of them can measure Etsy searches directly. Etsy does not publish query counts, so each vendor builds a private model from its own mix of sources and none disclose the method. Competing-listing counts, by contrast, can be counted directly, which is why all four tools land within 1.68x of each other on that metric and three to four orders of magnitude apart on volume.

Is it safe to compare keywords within a single tool?

Partly. In our data all four tools agreed on the strongest keyword, putting “custom name necklace” first or second. They fell apart on everything below that: “personalized dog mom gift” was ranked first by Alura, third by EverBee, last by Sale Samurai, and unmeasurable by eRank. Use one tool to pick your best candidate, and do not rely on any of them to rank weak or mid keywords against each other.

What does a search volume under 20 mean in eRank?

That eRank does not have enough data to give a number, and says so rather than inventing one. Four of our ten keywords came back that way. It reads like a broken tool and is the opposite: on the same keywords on the same day, other tools returned confident figures in the thousands and hundreds of thousands.

Why do reordered keyword phrases show different volumes?

Because the related-keyword panels are generated from combinations of words rather than measured queries. We saw the artifact in three separate products. Sale Samurai returned exactly 18,799 for both “custom name necklace” and “name custom necklace” while giving a third arrangement 11,030. Alura returned exactly 458,954 for two different phrasings. EverBee held volume nearly flat across three reorderings while their competing-listing counts ranged from 15,858 to 151,938, which is the model and reality disagreeing on the same screen.

Are high search volumes on Etsy real buyer demand?

Often not. Keyword tools cannot distinguish a buyer from a seller researching what to list, and both register as searches. eRank’s country breakdown is the only place we found this exposed: “crochet pattern pdf” was roughly 79% Pakistan and Morocco, and “personalized dog mom gift” was 38.5% Vietnam against 23.1% US. By contrast “gothic home decor” was 68.6% US, which reads as genuine buyer demand despite having the highest competition we measured.

Should I stop using Etsy keyword tools?

No. Use the competing-listing count, which is real and which the tools agree on, and use volume only to find your single strongest candidate inside one tool. The free tiers are enough for that. Our six-tool comparison covers which ones are worth paying for once you outgrow the free plans.

Data collected July 16, July 22 and July 30, 2026, on free-tier accounts. Full per-tool findings live in our reviews of eRank, EverBee, Alura, EHunt and Sale Samurai. Method: how we test.