umarattique.com
← All posts

Keyword Difficulty Is Lying to You: Why I Research SERPs, Not Scores

July 18, 2026 · Umar Attique Shah · Strategy

Every week I see the same advice: “Check the KD before you target a keyword. Under 30, go for it. Over 50, forget it.”

I think this advice quietly ruins more SEO campaigns than any algorithm update. Here’s why — and what I do instead.

What KD actually measures

Keyword difficulty is a tool’s estimate, and every tool calculates it differently — mostly from the backlink profiles of the pages currently ranking. That means KD is really answering one narrow question: how many links do the current top results have?

It is not answering the questions that decide whether you can rank:

The SERP is the ground truth

So before I trust any score, I open the results page and read it like a competitor. Three signals tell me almost everything:

1. Who’s ranking. If page one is Quora, Reddit, and template blogs, that’s not competition — that’s Google begging for a better answer. A “KD 40” query with a weak SERP is easier than a “KD 10” query where three funded brands are fighting.

2. How well the intent is served. Read the top three results as if you searched the query yourself. Did you get your answer? If you’d hit the back button, so will everyone else — and that’s your opening.

3. Coverage depth. Are the ranking pages complete resources, or 600-word posts that answer a third of the question? Gaps in coverage are gaps you can walk through.

Where the low-KD gold hides

In genuinely competitive industries — finance, SaaS, health — the head terms are a warzone. But insiders don’t search like outsiders. Doing research for a finance client recently, I found evergreen process keywords and institutional jargon terms with single-digit KD, sitting untouched next to KD-70 head terms. The pattern repeats everywhere:

My working rule

KD tells you how hard the keywords you thought of are. The SERP tells you how hard the niche actually is.

Use KD as a rough first filter if you like. But never let a number talk you out of a keyword — or into one — before you’ve read the page it’s describing. The tools are guessing. The SERP isn’t.

I put this approach to the test on a brand-new site and documented the results — read the case study here.

Want results like this on your site? Work with me →