Semrush is $139 a month and Ahrefs is $129. Here is what you actually lose by not paying, what you do not, and how to close the gap.
There is no free source of real monthly search volume. Not one. Every free tool showing you "1,900 searches/month" is doing one of three things:
Even the paid tools are estimating. Their numbers disagree with each other on the same keyword, often by a factor of two or more. Volume is directional everywhere, and precise nowhere.
Once you accept that volume is an estimate even when you pay for it, the question changes from "how do I get the numbers" to "what decisions actually need them".
| Capability | Free equivalent? |
|---|---|
| Estimated search volume | No. This is the real gap. |
| Keyword difficulty score | No — but you can look at who ranks, which is the underlying signal. |
| Competitor keyword lists | No. |
| Backlink data | No. |
| Rank tracking over time | Partly — Search Console shows your own positions free. |
| Keyword discovery | Yes. Autocomplete is the same underlying behaviour. |
| Question mining | Yes. Autocomplete plus People Also Ask. |
| Seasonality | Mostly. Google Trends is free. |
Read that column honestly: the paid tools' moat is volume, difficulty and competitor data. Discovery — finding the queries in the first place — is not part of the moat. That is why a free autocomplete tool can return several hundred real keywords in two seconds.
If your site has any traffic at all, this is the single most valuable keyword source you have, and it is the only one showing real impressions and clicks rather than estimates.
The move most people miss: filter to queries with impressions but a position between 8 and 25. Those are queries Google already thinks you are relevant for and where you are nearly ranking. Improving an existing page for those is far cheaper than writing a new one.
Covered in detail in the long tail guide. In short: expand one seed with alphabet soup, question prefixes and comparison suffixes, then rank by how many separate queries surfaced each result. Our keyword generator automates it across Google, YouTube and Amazon.
Trends will not tell you a keyword's volume, but it will tell you whether interest is rising, falling or seasonal, and how two terms compare relatively. For deciding which of two phrasings to use, that is often the only question that matters.
Free, ignored, and the most direct competition signal available. Search your target query and read:
Amazon autocomplete is pure buying intent. Reddit and niche forum search show the language real people use, which is frequently not the language the industry uses — and that mismatch is where uncontested keywords live.
You need volume less often than the tools imply. Three substitutes:
Not never. Pay when:
For a single site in its first year, the free stack plus Search Console covers the decisions you will actually make. Spending the subscription money on writing more pages usually returns more.
| Week | Do this |
|---|---|
| 1 | Expand 5 seeds. Filter to questions. Pick 20 with real answers you can write. |
| 2 | Search each one. Drop the ones where ten brands rank. Expect to lose half. |
| 3 | Write the survivors properly — one page per query, actually answering it. |
| 4 | Submit to Search Console. Then wait. Eight weeks minimum before judging anything. |
That plan costs nothing but time, and it produces real position data you own — which is the thing no subscription can sell you.
Expand your first seed across Google, YouTube and Amazon autocomplete.
Open the keyword generator →The same logic applied to competitor catalogues, pricing and launches.
Shopify research guide →