The Keyword Golden Ratio: Finding Keywords You Can Actually Rank For

Here's the problem with most keyword research advice: it was written for sites that already rank. "Target this 10,000-search keyword" is fine guidance if you have a decade of backlinks. If your site is six months old, it's a recipe for publishing into the void.

The Keyword Golden Ratio, developed by Doug Cunnington, is the rare method designed for the rest of us. It finds keywords where the competition is so thin that a new page can rank in days or weeks — not months or years. This is the full explanation: the math, the method, and the honest limitations.

The core insight

Search volume tells you how many people want an answer. It tells you nothing about how hard it is to provide one. KGR's insight is to compare demand against a specific, measurable form of supply: the number of pages Google considers relevant enough to put the keyword in their title.

When 500 people search a phrase each month but only 30 pages have it in their title, that's a lopsided market. Demand exists; dedicated supply doesn't. A focused new page can walk right in.

The math

The formula is simple division:

KGR = (allintitle results) ÷ (monthly search volume)

You get the allintitle number by searching Google for allintitle:"your keyword" — the count of results is pages with the exact phrase in their title. You get volume from any keyword tool (or estimates).

The thresholds:

• Under 0.25: golden. Publish immediately; these rank fast.
• 0.25 to 1.0: good. Worth targeting, expect weeks not days.
• Over 1.0: competitive. Skip unless you have authority.

A worked example

Say you run a site about home coffee brewing. Your keyword tool suggests "best pour over coffee maker for beginners" — 400 searches a month. You search allintitle:"best pour over coffee maker for beginners" and Google reports 62 results.

KGR = 62 ÷ 400 = 0.155. Under 0.25 — golden. A well-structured page targeting this phrase has a genuine shot at page one within weeks, because barely anyone has bothered to optimize for it directly.

Compare: "best coffee maker" — 40,000 searches, allintitle count of 38,000. KGR = 0.95. Technically under 1.0, but the absolute competition (38,000 optimized pages) tells you this is a different sport entirely. Which brings us to…

The honest limitations

KGR is a screening tool, not a guarantee. Four things it doesn't account for:

1. Absolute competition. A KGR of 0.5 on a 100-search keyword (50 competing pages) is a different fight than 0.5 on a 10,000-search keyword (5,000 pages). Always look at the raw allintitle number too.

2. Intent match. Low competition sometimes means low commercial value — nobody targets the keyword because nobody buys from it. Check that the intent fits your goals before you publish.

3. SERP features. If Google answers the query directly with a featured snippet, knowledge panel, or AI overview, even position one gets thin clicks. Look at the actual results page, not just the numbers.

4. Volume accuracy. Third-party volume numbers are estimates, sometimes wildly off for long-tail phrases. Treat KGR as directional — a way to sort candidates, not a prophecy.

The workflow that actually works

Here's how to use KGR in practice, start to finish:

Step 1: Mine broadly. Start with a seed topic and collect every autocomplete suggestion across Google, YouTube, and Bing. You want hundreds of raw phrases — the long tail is where KGR shines, and you can't screen what you haven't collected.

Step 2: Filter by intent. Before running any numbers, delete everything that doesn't match what you can actually serve. A cooking blog shouldn't target "best CRM software" no matter what the ratio says.

Step 3: Screen with KGR. Run the allintitle check on survivors. Sort by ratio. Your golden list (under 0.25) is your publishing queue, ordered by volume.

Step 4: Check the SERP manually. For each golden keyword, look at the actual top 10. Are they forums? Thin affiliates? Outdated posts? That's your opening. Are they Wirecutter and Forbes? Move on — the ratio lied, or the market changed.

Step 5: Publish the best answer, fast. KGR keywords reward speed and completeness, not authority. A genuinely complete page — covering the question fully, with original detail — wins these SERPs because the bar is low.

Building a KGR pipeline

Manual KGR works for a handful of keywords. For a real content operation, you need a pipeline: mine hundreds of suggestions, filter by intent, screen ratios in bulk, then manually review the golden list. The manual SERP check stays human — it's the judgment call — but everything before it should be automated. That's the difference between KGR as a party trick and KGR as a traffic strategy: the former finds five keywords, the latter feeds a publishing calendar.

Track your golden keywords in a simple sheet: keyword, volume, allintitle, KGR, SERP notes, publish date, rank check date. Revisit monthly. Some golden keywords decay as competitors discover them; others open up as SERPs shift. The sheet is a living asset — after a year, it's a map of your niche's soft spots.

When KGR finds nothing

Sometimes you run the method and every keyword comes back competitive. Three moves: go longer (add specificity — "for beginners," "under $50," "in small apartments" — each qualifier thins competition), go adjacent (your seed's neighboring topics may be softer), or go newer (emerging subtopics have no entrenched competition yet). A empty golden list usually means the seed was too broad, not that the niche is closed.

Also consider: maybe the niche genuinely is saturated for written content but thin on video, or vice versa. KGR thinking applies per-format. A keyword unwinnable in Google might be wide open on YouTube.

KGR and content clusters

Individual golden keywords are wins; clusters of them are a strategy. When your mining surfaces five related golden keywords ("pour over coffee ratio," "pour over grind size," "pour over vs drip," "best pour over dripper," "pour over troubleshooting"), don't write five disconnected posts — build a cluster: one pillar page covering the topic broadly, with each golden keyword as a dedicated sub-page linking back. The pillar accumulates authority, the sub-pages capture the specific queries, and internal linking distributes ranking power across all of them.

This is where KGR graduates from tactic to moat. Competitors can copy one keyword; they can't easily replicate a 20-page cluster built over six months. Each golden keyword you publish is a small win. Fifty of them, interlinked, is a traffic engine that compounds while you sleep.

Why this matters more in 2026

AI overviews now appear in a significant share of searches, and they disproportionately cover informational head terms — the big keywords everyone fights over. The long tail, where KGR operates, is comparatively less saturated by AI answers. Specific questions with specific intent still send clicks to specific pages.

The strategy writes itself: mine the long tail, screen with KGR, publish complete answers faster than competitors bother to. It's not glamorous. It compounds.

ContentCraft AI's KGR screener automates steps 1 and 3 — mining hundreds of suggestions and scoring them — so your time goes to steps 4 and 5, the parts that actually require judgment.

Quick reference card

Formula: allintitle count ÷ monthly volume. Under 0.25: publish now. 0.25–1.0: good target. Over 1.0: skip unless authoritative. Always check: absolute competition, intent match, SERP features, and the actual top 10 results. Mine broadly, filter by intent, screen by ratio, verify manually, publish completely. Revisit quarterly — golden keywords decay and new ones open as SERPs shift.