Keyword difficulty (KD) is a 0-to-100 score that estimates how hard it is to rank a page on the first page of Google for a given keyword. Most tools calculate it from the backlink profiles of the pages already ranking in the top 10: the more referring domains those pages have, the higher the score. It is a relative competition signal, not a verdict, and it ignores your own site's authority.
What is keyword difficulty measuring?
KD answers one narrow question: how much link authority do the pages currently winning this keyword have? It is a snapshot of the competition, not of you.
Here is the part most "what is KD" pages skip. The score is built almost entirely from one input: the referring domains pointing at the pages in the top 10. A keyword where the leaders have thousands of linking sites scores high. A keyword where the leaders have a handful scores low. That is the whole engine under the hood for most tools. Note "referring domains," not "backlinks": fifty links from one site count as one domain, which is why a page with a huge raw link count can still sit on a low-KD keyword if those links all come from the same few places.
What KD does not measure is just as important:
- Your domain's authority. KD describes the competition. It has no idea how strong your site is. A KD 30 keyword is easy for an established brand and hard for a site launched last month.
- Search intent. KD does not check whether the ranking pages match what the searcher wants. A "low difficulty" keyword you can't satisfy is still a loss.
- Content quality or topical relevance. Two pages with identical backlink profiles can rank very differently based on how well they answer the query. KD is blind to that.
- SERP features. AI Overviews, featured snippets, "People Also Ask," and ad blocks all eat clicks. A KD 10 keyword can still send you almost no traffic if an AI Overview answers it above the fold.
KD is a filter, not a strategy. Treat it as the first question you ask about a keyword, never the last.
How keyword difficulty is calculated
The mechanics differ by tool, which is exactly why the same keyword can read three different scores. Knowing the formula tells you how much to trust each number.
Ahrefs
Ahrefs pulls the top 10 ranking pages for a keyword, counts the referring domains pointing at each, and plots that against a logarithmic 0-to-100 scale. More linking domains across the leaders means a higher score. It is deliberately the simplest model of the major tools: it considers backlinks and effectively nothing else. The logarithmic scale matters more than people realize. Because the curve is steep at the top, the jump from KD 70 to KD 80 represents far more links than the jump from KD 10 to KD 20, so the difference between two high-KD keywords is much larger than the raw numbers suggest. Ahrefs estimates that a KD 40 keyword needs roughly 56 referring domains to crack the top 10. Useful as a rough rung on a ladder, not a promise.
Semrush
Semrush runs a heavier formula. It blends the median referring domains of the top-ranking pages with the dofollow-to-nofollow link ratio, the median Authority Score of those pages, search volume, and the presence of SERP features. More inputs, so the scores tend to skew higher and read more conservatively than Ahrefs on the same term. That conservatism cuts both ways: it flags crowded SERPs that a pure link count would miss, but it can also scare you off keywords that are softer than the number implies.
Moz
Moz scores difficulty from the Page Authority and Domain Authority of the pages in the top 10, on a 1-to-100 scale. It is leaning on its own authority metrics rather than a raw link count, which is a different lens on the same question.
The takeaway: KD is an estimate produced by a model, and the models disagree on purpose. None of them is "the" difficulty. They are educated guesses about how crowded the field is.
How to read a KD score
The number on its own is close to useless. KD becomes a real targeting tool only when you read it next to two other things: search intent and your own domain authority.
Step 1: Sanity-check the score against your own DA
A KD score is a competition reading, so you have to weigh it against your own strength. If your Domain Authority is modest, low-KD keywords are where you win in the near term. High-KD head terms are a 12-to-24-month authority play, not a launch fight. Picking KD 70 keywords on a six-month-old site is how agencies burn a year of budget and call it "building authority." If you're starting from scratch, the smarter sequence is the one we lay out in SEO for a new website: win the small fights first, earn the links, then climb.
Step 2: Cross-reference search intent
Low KD plus the wrong search intent is a trap. If the keyword is informational but you only have a product page to offer, you will not rank, no matter how soft the KD looks. Always confirm that the pages currently ranking match the kind of page you can credibly make. Match the page type to what's already winning, not to what you wish were winning.
Step 3: Open the actual SERP
Numbers lie by omission. Before you commit, look at the live SERP. A KD 15 keyword owned by an AI Overview, three ads, and a giant brand is harder than the score admits. A KD 45 keyword where the top results are thin, dated, and off-intent is more winnable than the score admits. While you're in there, scan the referring-domain counts on the top results and ask the honest question: can I realistically earn links in that range? That number, not the KD label, is the real entry fee, and closing the gap is what a link building program is for. The SERP is the ground truth; KD is the weather forecast.
A simple working rule
Stack the three signals: low-to-moderate KD, intent you can satisfy, and a SERP without an unbeatable incumbent or a click-stealing AI answer. When all three line up, you have a target. When only the KD looks good, you have a guess.
Why the same keyword shows different KD in different tools
Because the tools measure different things. Ahrefs is almost pure referring domains. Semrush folds in authority scores, link quality, volume, and SERP features. Moz uses its own PA and DA. Same keyword can read KD 23 in one tool, KD 58 in another, and "medium" in a third. None is wrong; they are answering slightly different questions. Pick one tool, learn how its scale behaves on keywords you know, and stay consistent. Comparing raw KD numbers across tools is comparing apples to a different orchard.
Where keyword difficulty fits in real SEO work
KD is a triage tool at the top of the keyword research funnel. It tells you, fast, which fights are worth having from where you stand today. It does not tell you whether a keyword is worth winning (that's intent and commercial value), and it does not tell you how to win it (that's content, links, and technical execution). Good SEO strategy uses KD to build a ladder: a base of winnable low-KD terms that earn early authority, which then makes the high-KD head terms reachable later.
In practice, that ladder is rarely a list of single keywords. It is a set of topic clusters, where a handful of low-KD supporting pages feed authority to the competitive money page at the center. KD tells you which rung to start on; the cluster is how you climb it. The score is the map. It is not the trip.
The short version
KD tells you who you're up against. It does not tell you whether you can win, what you'll win, or how. Use it as a filter at the top of keyword research, weigh it against your own authority and the live SERP, and never let a single number make the call for you.
If you'd rather not spend your week squinting at KD columns, that's our job. We build SEO programs that target the keywords you can win now and ladder toward the ones worth winning later, with the strategy explained in plain English the whole way (and the pricing explained the same way). Get in touch and we'll tell you straight which keywords are worth your time. No obligation, no runaround.
Browse more terms in the MoonSauce SEO glossary.
