Keyword difficulty is a 0 to 100 score that estimates how hard it is to reach Google’s first page, based mostly on backlinks to the pages already ranking. It ignores on-page quality, intent and who is searching, so treat it as a first filter, not a verdict.
By Vlad Ivanov · Last updated: October 2026Key Takeaways
- Ahrefs calculates KD from the referring domains of the top 10 ranking pages, on a 0 to 100 scale that is not linear (Ahrefs).
- Ahrefs states that KD does not take into account any on-page factors, and that it estimates top-10 chances, not top 3.
- Semrush uses six bands, from 0-14 “very easy” to 85-100 “very hard”, and blends referring domains, link ratios, Authority Score and SERP features (Semrush).
- In 20 keyword SERPs we checked on 7 October 2026, 9 had a Reddit thread in the top 3. KD does not tell you that.
- “ahrefs alternatives” has a KD of 2 (DataForSEO) and a Reddit thread at position 2.
- The term gets about 320 US Google searches a month, and DataForSEO estimates 261 monthly mentions in AI assistant queries.
What is keyword difficulty?
Keyword difficulty, usually shortened to KD, is a score from 0 to 100 that SEO tools attach to a keyword to estimate how hard it will be to get a page onto Google’s first page. Every major tool has one: Ahrefs KD, Semrush KD%, Moz Difficulty, and the KD figure in DataForSEO data. It sits next to search volume in almost every keyword report, which is why so many people pick keywords by it. If you are new to keyword research, start with what keyword research is.
How is keyword difficulty calculated?
Ahrefs looks at the top 10 ranking pages and counts how many referring domains they have. More referring domains across the top 10 means a higher score. The scale is not linear, and Ahrefs is explicit that KD “doesn’t take into account any on-page factors” and estimates the chance of reaching the top 10, not the top 3 (Ahrefs KD page).
Semrush blends the median referring domains to ranking URLs, the median dofollow/nofollow ratio, the median Authority Score of ranking domains and SERP features, then places the result in six bands (Semrush KD).
| Semrush KD% | Label |
|---|---|
| 0-14 | Very easy |
| 15-29 | Easy |
| 30-49 | Possible |
| 50-69 | Difficult |
| 70-84 | Hard |
| 85-100 | Very hard |
Both formulas are mostly about links. That is reasonable: links still matter. But it means KD answers one narrow question, “how many links do the current winners have?”, and people read it as “can I rank?”.
Why does KD mislead?
It ignores who is ranking. A Reddit thread or a thin forum page can sit in the top 3 with almost no links. Low KD and a beatable SERP often go together, but so do low KD and three huge brands with old pages. It ignores on-page quality, by Ahrefs’ own definition. It ignores the SERP layout. An AI Overview or a video carousel above the results takes clicks no matter what your KD is; see the AI Overviews CTR statistics. It ignores who is searching. A KD 5 keyword full of students is still worthless to a B2B product.
What did we find when we checked KD against real SERPs?
On 7 October 2026 we pulled DataForSEO KD and the live US top 10 for 20 keyword-research keywords. Here are the ones with a KD score, next to who actually ranks in the top 3:
| Keyword | US volume | KD | Top 3 domains | Reddit in top 3 |
|---|---|---|---|---|
| ahrefs alternatives | 590 | 2 | geotoolbox.ai, reddit.com, ustechautomations.com | Yes |
| semrush alternatives | 1300 | 10 | eesel.ai, searchlogistics.com, reddit.com | Yes |
| mangools alternative | 140 | 0 | mangools.com, softwareadvice.com.au, reddit.com | Yes |
| answerthepublic alternative | 110 | 0 | blog.answersocrates.com, reddit.com, answerthepublic.com | Yes |
| ahrefs vs semrush | 880 | 21 | searchatlas.com, g2.com, linkedin.com | No |
| b2b keyword research | 140 | 0 | neilpatel.com, konstructdigital.com, designatwork.com | No |
| best keyword research tool for affiliate marketing | 40 | 0 | competitors.app, databox.com, onclickdigitalmarketing.com | No |
| reddit keyword research tool | 50 | 0 | reddit.com, devta.so, reddit.com | Yes |
| buyer intent keywords | 90 | 5 | marketbold.com, robbierichards.com, wsiworld.com | No |
| keyword difficulty | 320 | 29 | seo.com, eesel.ai, writesonic.com | No |
| youtube comment scraper | 170 | 1 | medium.com, youtube.com, reddit.com | Yes |
Nine of the 20 SERPs had a Reddit thread in the top 3. For a new site, that is the signal that matters: a forum thread ranking means Google has no great page yet. KD shows none of it. It also flags the opposite case: “keyword difficulty” itself has KD 29, yet its top 3 are seo.com, eesel.ai and writesonic.com, none of them giant brands.
What should you check instead of a KD score?
Use KD to throw out the obvious no-go keywords, then open the SERP and check four things:
- Weak results in the top 10: forums, Reddit, Quora, thin pages, outdated posts. One or two of these is your opening. The method is in how to find low competition keywords.
- Authority of the top 3, not the top 10: the clicks are in the top 3, and KD averages them away. A quick way to size them up is in the SEO competitor analysis guide.
- SERP features above the results: AI Overviews, video packs and ads push organic results down.
- Who the searcher is: does this keyword come from your buyer or from someone who will never buy? See buyer intent keywords.
This is the logic I built into ICP Keywords’ Opportunity score: it combines difficulty with the live domain rating of the top 3, intent, SERP features (video results and above-the-fold AI Overviews lower the score) and how well the keyword fits your Ideal Customer Profile, with weights you set yourself. Tools like LowFruits take a similar approach for weak spots, which I cover in the Semrush alternatives list.
Does KD matter for new sites?
More than for established sites, but in the opposite direction from how most people use it. A brand-new site should ignore volume-first lists and look for SERPs where weak pages already rank, whatever the KD. I tested this in keyword research for zero-DR sites, and how many links you really need is covered in how many backlinks you need to rank.
ICP Keywords’ Opportunity score adds live top-3 authority, SERP features and buyer fit to difficulty. 100 free credits, 7 days, no card.
Frequently asked questions
What is keyword difficulty?
Keyword difficulty (KD) is a 0 to 100 score SEO tools give a keyword to estimate how hard it is to reach the first page of Google. Ahrefs bases it on the number of referring domains pointing to the top 10 ranking pages.
How is keyword difficulty calculated?
It depends on the tool. Ahrefs uses the referring domains of the top 10 ranking pages and says the scale is not linear. Semrush blends the median referring domains, the dofollow/nofollow ratio, the median Authority Score of ranking domains and SERP features.
What is a good keyword difficulty score?
For a new site, KD under 10 to 15 is a reasonable starting filter. But check the live SERP: a KD 2 keyword can still have strong sites in the top 3, and a KD 30 keyword can have a forum thread you can beat.
Is keyword difficulty accurate?
It is accurate about what it measures, which is backlinks to the current top 10. It does not measure on-page quality, search intent, SERP features or who is searching, and Ahrefs says itself that KD does not account for on-page factors.
Why do Ahrefs and Semrush show different KD for the same keyword?
They use different formulas and different backlink indexes. A score of 20 in one tool is not the same as 20 in the other, so never compare KD across tools.
What should I use instead of keyword difficulty?
Use KD as a first filter, then check the live top 10: who ranks (forums, thin pages, big brands), whether the pages match the intent, whether an AI Overview or video box sits above the results, and whether the people searching are your buyers.
Does keyword difficulty matter for AI search?
Less than for Google. AI assistants cite pages that answer the question directly, and in our tests they often cited small sites. KD tells you nothing about whether ChatGPT will cite you.
Vlad Ivanov runs the Words At Scale YouTube channel (26,000+ subscribers) and nine Search Console-verified sites, and built ICP Keywords, a keyword research tool for B2B and niche sites. LinkedIn
