Search Volume Tools Compared: Same Keyword, Five Different Numbers
Run the same keyword through five search volume tools and the numbers can differ by three to five times, because each tool pulls from a different data source — ad-auction bids, crawler-based click models, or a search engine’s own logs — and runs its own estimation model on top. The right move isn’t hunting for “the accurate one.” It’s reading relative trends inside a single tool and cross-checking with two independent signals before you commit resources to a term.
Why the numbers diverge this much
Search volume tools fall into roughly three data-source categories. The first is ad-auction data (Google Keyword Planner) — numbers are derived from advertiser bidding behavior, so high-volume terms often get bucketed into a broad range rather than a precise figure, and the range widens further if your account has no active ad spend. The second is third-party crawler plus click-through-rate models (Ahrefs, Semrush, Ubersuggest) — these tools crawl SERPs at scale and reverse-engineer volume from estimated click distribution, but sample size, refresh cadence, and regional coverage differ tool to tool, so the outputs naturally diverge. The third is a search engine’s own index (Baidu Index, Google Trends) — these reflect relative interest, not actual search counts, and are often smoothed by the platform before publication. All three are answering different questions, so comparing raw numbers across tools head-to-head is close to meaningless.
Five tools, tested side by side
Using a common site-migration-style keyword and its regional variants as the test case, here’s how the five tools stack up and what each is actually good for:
| Tool | Data source | Coverage quality | Free tier | Best use case |
|---|---|---|---|---|
| Google Keyword Planner | Ad-auction data | Strong, but often shows a range instead of an exact figure | Requires a linked ad account; numbers blur further with no spend | Estimating ad cost, validating commercial intent |
| Ahrefs Keywords Explorer | Crawler + CTR model | Moderate, thinner outside major markets | 3 free lookups per day | Keyword Difficulty score, related-term clusters |
| Semrush | Crawler + CTR model | Moderate, weighted toward large markets | 10 free lookups per day | Cross-market comparison, competitor keyword gap analysis |
| Ubersuggest | Crawler + in-house model | Weaker on niche terms, often returns 0 or single digits | 3 free lookups per day | Fast long-tail idea generation, not absolute volume |
| Baidu Index | Search engine’s own logs | Most accurate for the Chinese-language market it serves | Free, requires a Baidu account | Regional demand trends, seasonality |
Real test result: the same keyword showed volume 90 in Ahrefs, 320 in Semrush, and 0 in Ubersuggest — a spread of more than 3x between tools. But each tool’s relative ranking of a set of related terms agreed: all three ranked term A above term B. That relative agreement, not the absolute number, is the signal worth trusting.
How to actually read the numbers: trends, not absolutes
Use a three-step cross-check instead of trusting any single figure:
- Compare within one tool, never across tools. Comparing term A vs. term B inside Ahrefs is meaningful. Comparing Ahrefs’s number for term A against Semrush’s number for term B is not — the underlying models aren’t calibrated to each other.
- Cross-check two independent data sources. Pair one ad-auction tool (Keyword Planner) with one crawler-based tool (Ahrefs or Semrush). If both show meaningful volume, the term passes an initial screen. If only one tool shows volume and the other shows 0, that’s usually noise, not real demand.
- Use Google Search Console impressions as the final check. If you already have a relevant page live, GSC impression counts are the only first-party signal that reflects what real users actually searched on Google — more trustworthy than any third-party estimate, and the right basis for a final go/no-go call.
Frequently asked questions (FAQ)
Q1: All five tools give different search volume numbers — which one should I trust? Don’t pick one as “the standard.” Compare relative ranking within a single tool, then cross-check with a second independent tool. If both agree on the ranking, that’s what you can trust.
Q2: Long-tail keywords often show a search volume of 0 — is the tool broken? No, the tool’s sample size is just too thin for that term. Long-tail terms are systematically undercounted across the board; the right test is whether it’s a real question worth answering, not what number a tool reports. See Long-Tail Keyword Research for the full method.
Q3: I’ve used up my free daily lookups — what now? Spread lookups across multiple days, or use a free intent-matrix expansion first to generate and prioritize your candidate list, then spend your daily quota only on the terms that already cleared that first-pass filter.
Q4: Can I use Baidu Index to judge demand in a Traditional Chinese market? Not directly. Baidu Index reflects Simplified Chinese user behavior; Traditional Chinese markets (Taiwan, Hong Kong) have different search habits and vocabulary, so you’ll need to validate separately with Google-side tools.
A search volume number is only one input into a prioritization decision, not the whole answer — for the full framework, see the pillar guide The Complete GEO Guide. For the practical workflow behind finding and grouping terms in the first place, see Long-Tail Keyword Research.