Keyword Difficulty Scores: What They Measure and Why Tools Disagree
- Filed November 10, 2026
- Keyword research, guide 5 of 6
A keyword difficulty score is a vendor’s estimate of how hard it would be to rank on the first page of Google for a query, expressed on a 0 to 100 scale. It is not a Google metric. Each tool builds the score from its own data with its own formula, which is why the same keyword can be “easy” in one tool and “hard” in another. Ahrefs says these tools “all use the same 0-100 scale, but each one calculates it differently” (Ahrefs on keyword difficulty). Used well, difficulty helps compare keywords within one tool. Used as a fact about Google, it misleads.
What the scores are built from
Vendors describe their own scores, in their own words:
| Tool | How the vendor describes it |
|---|---|
| Ahrefs (KD) | "We pull the top 10 ranking pages for your keyword and look up how many websites link to each of them"; "the more links the top-ranking pages for your keyword have, the higher its KD score" |
| Semrush (KD%) | "A metric that measures the SEO effort it would take for your content to rank organically in the top 10 results", calculated from "a variety of factors such as the median number of referring domains pointing to the URLs that are ranking" |
Sources for the table: Ahrefs on keyword difficulty and Semrush on keyword difficulty.
Both lean heavily on backlink data about the pages that currently rank. That makes difficulty a proxy for one dimension of competition, and it inherits the tool’s link index: a tool that has discovered fewer links for a page will rate the keyword easier.
Google’s guidance on third-party SEO tools reminds readers that such tools “don’t have access to our internal ranking data” (Google on third-party SEO tools). A difficulty score is a model of competition built outside Google.
Why tools disagree
- Different inputs. One tool counts referring domains to the top 10; another uses a median; another adds other factors.
- Different link indexes. Each vendor crawls the web separately and discovers different links.
- Different SERP snapshots. The top 10 for a query changes with location, device and date; each tool samples at different times and places.
- Different scaling. Scores are normalized differently and the scales need not be linear, so “40” does not mean the same distance from “20” in two tools.
- What is left out. Link-based scores do not capture content quality, search intent match, brand strength in the niche, or SERP features that take clicks.
What difficulty does not measure
- Intent fit. A query with low difficulty is no opportunity if your site cannot serve its intent.
- Topical relevance of your site. A site with depth in a topic may compete where the score suggests otherwise, and the reverse.
- Results page features. A low-difficulty query dominated by a local pack, shopping units or an AI Overview may send few organic clicks.
- Your own site’s strength. The scores described above rate the keyword, not your site’s chances; some vendors offer separate personalized metrics, which follow their own methods.
A multi-tool interpretation example
Suppose an accounting software company compares three keywords across two tools. All figures are illustrative.
| Keyword | Tool A KD | Tool B KD% | Results page notes | Reading |
|---|---|---|---|---|
| invoice template | 72 | 58 | Template sites and large brands; shopping-style template galleries | Hard in both tools; strong competitors; template intent |
| how to invoice as a freelancer | 18 | 41 | Guides from small blogs and two large brands | Tools disagree; results show smaller sites ranking, so worth testing |
| net 30 payment terms meaning | 9 | 22 | Glossary pages; an AI Overview appears | Easy by links; AI Overview may take clicks; small but relevant |
How to read it:
- Use each tool to rank keywords against each other, not against another tool’s numbers. Within Tool A, “net 30” is easier than “freelancer invoicing”, which is easier than “invoice template”. Tool B agrees on the order.
- Where tools disagree on a keyword’s level, look at the results page. If smaller, less-linked sites rank with useful content, the keyword may be more open than a link-based score suggests.
- Combine difficulty with intent and value. The freelancer query matches the product and the audience, which may justify the effort even at moderate difficulty.
A better question than “how difficult?”
Ask instead: “Can we produce the best result for this query, given who ranks now and what we know?” Check:
- the intent of the results, and whether you can serve it;
- what the ranking pages offer, and what yours would add;
- how many of the ranking sites are similar to yours in size and focus;
- which features occupy the results page;
- how relevant the query is to your business.
Difficulty helps triage a long list. These five checks decide.
Practical rules
- Pick one tool for difficulty and stay consistent.
- Record the tool and date with every score.
- Do not report difficulty as a Google measure.
- Re-check difficulty when results pages change, for example after a core update.
Difficulty scores compress one view of competition into a number. They are useful for sorting and misleading when treated as a forecast.