Most keyword research frameworks start and end with two numbers: search volume and keyword difficulty. That’s like choosing which house to buy based only on square footage and price per square foot. You’d miss the neighborhood, the commute, the school district, and whether the foundation has cracks.
A keyword research framework is a structured methodology for identifying, evaluating, and prioritizing search terms based on multiple factors including business relevance, searcher intent, competitive gap, content fit, and conversion potential. Volume and difficulty are inputs. They’re not the framework.
“We stopped leading with volume in our keyword research three years ago,” says Hardik Shah, Founder of ScaleGrowth.Digital. “When we started scoring keywords on business relevance first, the traffic we generated was smaller in volume but 4x more likely to convert. The CFO noticed that.”
What’s wrong with the standard keyword research approach?
The standard approach looks like this: open Ahrefs or Semrush, enter seed keywords, export a spreadsheet, sort by volume descending, filter by keyword difficulty, pick the top 50, create a content calendar. Done.
The problem isn’t any individual step. The problem is what’s missing.
Missing: Business relevance scoring. A B2B SaaS company targeting “what is project management” (volume: 12,000) will get traffic from students writing essays, not from IT directors evaluating software. High volume, zero pipeline impact.
Missing: Intent segmentation. The keyword “CRM” could mean someone wants a definition, a comparison, a free tool, or a purchase. Each intent requires a different page type. Treating all four as the same keyword is how you end up with one page trying to serve four audiences and satisfying none of them.
Missing: Competitive gap analysis. Are you finding keywords your competitors rank for that you don’t? Are you finding keywords where the current top results are weak and beatable? Standard volume/difficulty filtering catches neither. For a complete process on competitive gaps, see our competitor content gap analysis guide.
Missing: Content format mapping. Some keywords demand video. Some demand tools. Some demand 4,000-word guides. Some demand 500-word definitions. If your format doesn’t match what Google rewards for that query, difficulty scores become meaningless.
The six-dimension keyword research framework
Here’s the framework we use at ScaleGrowth.Digital. It evaluates every keyword across six dimensions before it enters our production pipeline.
Dimension 1: Business relevance (weight: 30%)
Can you draw a line from this keyword to revenue in two steps or fewer?
Step 1: Someone searches this keyword and lands on your page. Step 2: They take an action that moves them toward a purchase (signs up, requests a demo, downloads a resource, visits a pricing page).
If you need more than two steps, the keyword is top-of-funnel awareness content. That’s fine, but it should be scored differently than a keyword where the searcher is actively evaluating your category.
We use a 1-5 scale:
| Score | Criteria | Example (for an SEO agency) |
|---|---|---|
| 5 | Direct purchase/hire intent | “SEO agency Mumbai” |
| 4 | Comparison/evaluation intent | “best SEO tools for enterprise” |
| 3 | Solution-seeking intent | “how to improve organic traffic” |
| 2 | Problem identification | “why is my website traffic dropping” |
| 1 | Pure information | “what is a meta description” |
The scoring forces a conversation that most keyword research skips. Is this keyword actually connected to what we sell? An honest answer eliminates 30-40% of keyword lists.
Dimension 2: Search intent clarity (weight: 20%)
How clear is the dominant search intent for this keyword? Look at the top 10 results. If all 10 are how-to guides, the intent is clear. If results include a mix of product pages, blog posts, tools, and videos, the intent is fragmented.
Why does intent clarity matter? Because when intent is fragmented, Google is still figuring out what users want. Your page needs to match the dominant intent perfectly, or Google will test it briefly and then demote it.
- Score 5: 9-10 of the top 10 results share the same format and intent. Crystal clear.
- Score 4: 7-8 results share intent. Some variance but a clear winner.
- Score 3: 5-6 results share intent. Google is split.
- Score 2: Mixed intent across SERP. Multiple page types competing.
- Score 1: Completely fragmented. Local packs, knowledge panels, videos, and articles competing.
Keywords with score 3 or below are harder to rank for than their keyword difficulty number suggests, because even if you beat the competition, Google might prefer a different content format next month.
Dimension 3: Competitive opportunity (weight: 20%)
This goes beyond keyword difficulty scores. KD is a useful starting point but it’s a blunt instrument. It doesn’t tell you if the top results are actually good.
For each keyword, we answer three questions:
Are the top results beatable on content quality? Open the top 3 results. Read them. Are they comprehensive? Are they current? Do they actually answer the query well? If the #1 result is a 2019 article with outdated data, that’s an opportunity regardless of what the KD score says.
Is there a content gap in the top results? Do all the top results cover the same subtopics and miss others? If every article about “keyword research” covers tools and process but none cover how to score keywords by business relevance, that’s your angle.
What’s the domain authority gap? If the top 5 results all have DR 80+ and you’re at DR 30, keyword difficulty alone doesn’t capture how hard it will be. But if position 6-10 includes sites with DR 30-40, there’s room.
Dimension 4: Search volume and trend (weight: 15%)
Yes, volume matters. But it’s the fourth dimension, not the first.
Raw monthly volume tells you the ceiling. But you also need the trend. A keyword with 5,000 monthly searches and declining 20% year-over-year is worth less than a keyword with 2,000 monthly searches and growing 30% year-over-year.
Google Trends is free and underused for this. Ahrefs and Semrush both show volume trends in their keyword tools. Look at the 2-year trend, not just the snapshot.
Also consider: does this keyword trigger SERP features that reduce organic CTR? A keyword with 10,000 monthly searches but an AI Overview, featured snippet, People Also Ask, and shopping results might only deliver 1,500 clicks to organic results. Ahrefs’ “clicks” metric helps here. It shows estimated actual clicks, which can be dramatically lower than search volume.
Dimension 5: Content effort (weight: 10%)
What will it take to create content that ranks for this keyword?
- Score 5 (minimal effort): You have expertise in-house, data readily available, simple content format. 3-5 hours total.
- Score 4: Standard content piece, some research needed. 5-10 hours.
- Score 3: In-depth piece requiring data collection, expert interviews, or original research. 10-20 hours.
- Score 2: Major content project with design, interactive elements, or tool building. 20-40 hours.
- Score 1 (maximum effort): Multi-contributor project with original data, video, and extensive design. 40+ hours.
This dimension prevents you from filling your queue with ambitious projects that never get finished. A mix of quick wins (score 4-5) and longer-term investments (score 1-2) keeps the pipeline productive.
Dimension 6: Topical authority fit (weight: 5%)
How well does this keyword fit within your site’s existing topical coverage? A keyword that extends a topic cluster you’ve already built is easier to rank for than a keyword in a completely new topic area.
Google’s understanding of topical authority means that sites with depth in a subject area rank more easily for related terms. If you’ve published 15 articles about content marketing, a new article about “content marketing ROI” fits naturally. An article about “supply chain logistics” does not, even if your business is somehow related.
We score this based on existing content coverage:
- Score 5: You have 5+ published pages on closely related topics.
- Score 4: You have 2-4 related pages.
- Score 3: You have 1 related page.
- Score 2: No related content, but it fits your brand expertise.
- Score 1: No related content, and it’s a stretch from your core expertise.
How to apply the framework in practice
Here’s the workflow we follow for every keyword research cycle:
Step 1: Generate the initial keyword list. Use 3-5 seed keywords in Ahrefs Keywords Explorer. Export all keyword ideas with volume above 50. Also run a content gap analysis against your top 5 competitors. This typically produces 500-2,000 raw keywords.
Step 2: De-duplicate and cluster. Group keywords by topic. “Keyword research tools,” “best keyword research tool,” and “keyword research tool free” all belong to the same cluster. Pick the highest-volume variant as the primary keyword for each cluster. This reduces your 500-2,000 keywords to 100-300 topic clusters.
Step 3: Score each cluster across all six dimensions. This is the time-intensive step. For 100 clusters, budget 4-6 hours. Most of the time goes into SERP analysis (dimensions 2 and 3). The other dimensions can be scored quickly once you have a calibrated scale.
Step 4: Calculate weighted scores and rank. Multiply each dimension score by its weight, sum them, and sort. Your priority queue emerges.
Step 5: Apply the create-versus-optimize filter. For each top-priority cluster, check if you have existing content. If yes, decide whether to optimize or rewrite based on current performance and content quality. Our SEO prioritization framework covers this decision in detail.
Step 6: Feed into the content brief process. The top 15-20 keywords become content briefs for the next production cycle. Brief creation is where the keyword research becomes actionable. See our guide on what a great SEO content brief looks like.
What most keyword research tools miss
Ahrefs, Semrush, and Moz are excellent tools. But they all have the same limitation: they’re optimized for volume and difficulty analysis, not for business relevance or intent scoring. Those dimensions require human judgment.
Specifically, tools miss:
Conversion potential. No tool can tell you which keywords drive revenue for your specific business. That data comes from your analytics, your CRM, and your sales team’s input.
Content quality of top results. A KD score of 45 could mean “the top results are mediocre sites that rank because nobody better has tried” or “the top results are excellent and exactly on-intent.” Those are completely different competitive situations. Only manual SERP review reveals which one you’re facing.
Brand relevance. Tools can’t judge whether a keyword aligns with your brand positioning. A keyword might be highly relevant to your industry but wrong for how you position yourself within it.
The framework compensates for these gaps by building human judgment into the scoring process. Tools provide the raw data. The framework provides the analysis layer.
How does AI change keyword research?
Two significant shifts since 2024:
AI Overviews reduce organic CTR for informational queries. Google’s AI Overviews now appear on roughly 30% of US queries, according to data from various SEO industry analyses in early 2025. For pure informational queries (“what is X”), AI Overviews often answer the question directly, reducing clicks to organic results by an estimated 25-40%.
This means your keyword research framework needs to account for AI Overview presence. A keyword with 5,000 monthly searches and an AI Overview might only deliver 2,000-3,000 organic clicks. Ahrefs’ “organic clicks” metric is becoming more important than raw search volume.
AI chatbots create new keyword opportunities. People are asking questions in ChatGPT, Claude, Gemini, and Perplexity that they used to search on Google. These queries tend to be longer, more conversational, and more specific. Traditional keyword tools don’t capture this demand yet. But if your content answers these queries well, AI tools will cite you, driving referral traffic from a completely new channel.
“We now run every keyword list through two filters,” says Hardik Shah, Founder of ScaleGrowth.Digital. “The traditional filter: can we rank for this in Google? And the AI filter: can we get cited for this in ChatGPT and Perplexity? Keywords that score well on both are top priority.”
Building your keyword research framework
Start with the six dimensions. Score 20 keywords manually to calibrate your scale. Then build a spreadsheet or Notion database that makes scoring fast for the remaining keywords.
After two quarterly research cycles, you’ll have enough performance data to validate and adjust your weights. Keywords you scored highly should be performing well. If they aren’t, your scoring needs recalibration.
If you want a team that runs this process for your brand as part of a broader growth system, reach out. Our Organic Growth Engine includes keyword research as a continuous input, not a one-time project, and every cycle’s research builds on the performance data from previous cycles.
Related Service