← Resources

Playbooks

Keyword analysis in seo

An end-to-end keyword analysis workflow for SEO - intent classification, difficulty filtering, SERP forensics, clustering, and the prioritization math that separates ranked content from buried content.

Updated

Before you start

  • A working SEO tool: Ahrefs, SEMrush, SpyFu, Mangools, or Ubersuggest
  • Google Search Console connected to your domain (free, mandatory)
  • Your domain's Domain Rating / Authority Score so you can filter realistic keywords
  • 2-3 hours for the first full analysis; 30 minutes monthly to refresh

The playbook

8 steps

0/8

  1. Start with seed keywords from your existing demand

    Don't brainstorm in a vacuum. Pull the top 50 queries Search Console already shows impressions for, plus the top 20 keywords from your best-converting landing pages in GA4. These are your seeds - real demand Google has already validated for your domain. Brainstorming from scratch leads to keyword lists that look great in a tool and produce zero ranking.

    # Search Console -> Performance -> Queries
# Sort by impressions desc, export top 50
# GA4 -> Reports -> Landing page -> filter by Organic Search
# Cross-reference to find keywords already pulling weight

    Expected outcome

    A 50-100 seed keyword list grounded in actual impression data, not guesswork.

  2. Expand seeds with the tool's keyword explorer

    Drop each seed into Ahrefs Keywords Explorer or SEMrush Keyword Magic Tool. Use the Matching Terms, Questions, and Also Rank For reports - they expand 1 seed into 200-1,000 related queries. Stop once you have 2,000-5,000 raw candidates; more becomes unmanageable.

    TipThe Questions report often surfaces high-intent long-tail with low difficulty - the lowest-hanging fruit in most analyses.

    Expected outcome

    A raw candidate pool of 2,000-5,000 keywords from your seeds.

  3. Classify every keyword by search intent

    Tag each keyword as informational, commercial investigation, transactional, or navigational. The SERP tells you which: top-10 dominated by blog posts = informational; review/comparison/best-of pages = commercial; product/pricing pages = transactional; brand names = navigational. Match your content type to the dominant SERP intent or you won't rank no matter how good the content is.

    # Quick intent heuristics:
# 'how to X', 'what is X' -> informational
# 'best X', 'X vs Y', 'X review' -> commercial
# 'buy X', 'X pricing', 'X coupon' -> transactional
# 'brand name', 'brand login' -> navigational

    Expected outcome

    Every candidate keyword tagged with intent; mismatched intents flagged for removal.

  4. Filter by realistic keyword difficulty

    Your domain can only rank for keywords within a difficulty band tied to your Domain Rating. Rough rule: KD < DR × 1.2 is winnable with good content; KD between DR × 1.2 and DR × 1.5 needs links; KD > DR × 1.5 is aspirational. Filter your candidate list aggressively. A DR 25 site chasing KD 60 keywords is wasting content investment.

    Expected outcome

    Candidate pool cut to keywords within your domain's realistic ranking band.

  5. Run SERP forensics on the survivors

    For your top 50 surviving candidates, manually open the SERP. Note: dominant content type (listicle / how-to / product / comparison), word count of #1-3 results (paste URLs into Ahrefs to see), schema markup used, SERP features present (featured snippet, PAA, AI Overview, video carousel). The pattern across the top 3 is the structural template you have to match.

    TipIf AI Overviews already answer the query, expect 30-50% lower click-through even if you rank #1. Discount those keywords or skip them.

    Expected outcome

    Per-keyword SERP profile documenting the content format, length, and SERP features required to compete.

  6. Cluster keywords into topic groups

    Multiple keywords usually map to one page, not many. Group semantically similar keywords (same intent, same SERP overlap >40%) into clusters. Each cluster = one article. Tools that automate this: Keyword Insights, Surfer, or Ahrefs' Parent Topic field. Skipping clustering produces cannibalized articles competing against each other.

    # Manual clustering signal:
# If keywords A and B share 4+ top-10 results, cluster them.
# If they share <2, they're separate clusters.
# Ahrefs Parent Topic does this automatically.

    Expected outcome

    50-200 keyword clusters, each representing one piece of content to ship.

  7. Score and prioritize each cluster

    Score = (sum of cluster search volume × commercial intent multiplier) / (KD × content-cost-to-produce). The clusters with the best score-per-effort go first. Don't ship in alphabetical order or whatever Ahrefs returns - prioritize ruthlessly by ROI per article.

    Expected outcome

    Ranked content roadmap - top 20 clusters become your next quarter's editorial calendar.

  8. Set up tracking before you publish anything

    Add your prioritized clusters to your rank tracker before the articles ship. Tracking from publish date gives you 0-30-60-90 day ranking curves that inform future analysis. Tracking only after rankings stabilize means you never learn what your actual ranking velocity looks like.

    Expected outcome

    Rank tracking dashboard primed with the new keyword set; baseline captured at publish.

Shuttergen

Keyword analysis ranks pages. Shuttergen ranks ads.

Once your keyword analysis surfaces the winners, Shuttergen helps you generate paid creative for those same intent groups - so your organic + paid coverage compounds instead of competing.

Generate creative freeAudit framework →

Pitfalls

What goes wrong

  • Confusing keyword volume with traffic potential

    Search Volume = total searches per month. Your traffic from ranking #1 = volume × CTR for position 1 × (1 - SERP feature dilution). A 10K-volume keyword with an AI Overview and a featured snippet may deliver 500 clicks; a 2K-volume clean SERP may deliver 800. Model traffic, not volume.

  • Skipping the intent check

    If you publish a how-to article on a transactional keyword, you won't rank. Google has decided what intent that query has - matching it is non-negotiable. Always validate intent via the SERP before writing.

  • Targeting KD > DR × 1.5 without a link plan

    Difficulty scores aren't suggestions. Keywords above your authority band don't rank from content quality alone; they need explicit link-building. Either fund the link campaign or pick easier keywords.

  • Building keyword lists, not content roadmaps

    A spreadsheet of 800 keywords helps nobody ship content. Cluster, score, and convert into an ordered editorial calendar - that's the output that actually moves rankings.

  • Treating keyword analysis as one-and-done

    Search demand shifts. Your monthly refresh (30 minutes) catches new question patterns, new SERP features eroding old keywords, and seasonal demand you'd otherwise miss.

Limits

When this playbook won't work

  • Your domain has no ranking history and you're targeting KD > 30 keywords - build authority on long-tail first
  • Your category is hyper-saturated and the top 10 are owned by DR 80+ domains with thousands of links each
  • AI Overviews already answer 80%+ of your category's queries - the addressable click market is collapsing regardless of analysis quality
  • Your buying cycle doesn't involve search at all - some B2B and product-led categories move through other channels

The difference between keyword research and keyword analysis

Keyword research = finding queries. Keyword analysis = deciding which ones are worth your content investment, what intent they signal, what content shape they require, and how they cluster into pages. Most teams do the first and skip the second, which is why most teams' content roadmaps look like spreadsheets and not strategies.

Analysis is where the leverage lives. Two teams with the same 5,000-keyword research output will produce wildly different traffic outcomes based on how they classify intent, filter difficulty, cluster topics, and prioritize. The analysis layer is where SEO maturity actually shows up.

Treat keyword analysis as a recurring system, not a one-time project. Demand shifts. SERP features shift. Your domain authority shifts. Run the full analysis annually; refresh the priority cluster scoring monthly. The discipline compounds over 12-18 months in a way ad-hoc analysis never does.

Keyword analysis ranks pages. Shuttergen ranks ads. Once your keyword analysis surfaces the winners, Shuttergen helps you generate paid creative for those same intent groups - so your organic + paid coverage compounds instead of competing.

Generate creative free

Tool stack for serious keyword analysis

Ahrefs for keyword explorer, difficulty scoring, Parent Topic clustering, and SERP analysis. The Standard plan ($249/mo) is the typical entry point for serious work. The Keywords Explorer alone justifies the spend for most SEO-driven businesses.

SEMrush is comparable on the core analysis features and stronger on PPC integration. Pick SEMrush if you want one platform for paid and organic keyword analysis.

Search Console is free, mandatory, and underused. It's the only source of ground-truth impression/click data for your domain. Connect it to whichever paid tool you pick.

Mangools, Ubersuggest, Serpstat are the budget tier - usable for early-stage analysis but they lack the depth needed for clustering and SERP forensics at scale.

Internal: find-seo-competitors, analytics-keywords-report, keyword-monitoring.

FAQ

Frequently asked

What is keyword analysis in SEO?
The process of evaluating keyword candidates by intent, difficulty, SERP shape, and traffic potential, then clustering them into prioritized content groups. Research finds the candidates; analysis decides which ones you write about and in what order.
How is keyword analysis different from keyword research?
Research generates a candidate list. Analysis evaluates that list - filtering by difficulty, classifying intent, modeling traffic potential, clustering into topic groups, and prioritizing by ROI per article. Analysis is where most of the value is.
What's the best tool for keyword analysis?
Ahrefs Keywords Explorer for depth and Parent Topic clustering; SEMrush Keyword Magic Tool as a comparable alternative. Search Console is mandatory free input. Mangools and Ubersuggest are budget options that work for early-stage analysis.
How do I know which keywords are realistic for my domain?
Rough rule: KD < your Domain Rating × 1.2 is winnable with quality content alone. KD between DR × 1.2 and DR × 1.5 needs links. KD > DR × 1.5 is aspirational without a major link or PR campaign.
How many keywords should I analyze?
Start with 2,000-5,000 raw candidates; filter to 200-500 by intent + difficulty; cluster to 50-200 content groups; ship the top 20 by ROI score first. The funnel matters more than the starting number.
How often should I redo keyword analysis?
Full analysis annually. Monthly 30-minute refresh on priority clusters and seasonal demand. Quarterly check for SERP-feature shifts (AI Overviews, featured snippets) that change traffic models for existing rankings.
Does keyword analysis still matter with AI Overviews?
More than ever. AI Overviews reshape which keywords have clickable traffic. Analysis is how you identify the keywords still worth ranking for and the ones where ranking #1 will produce minimal clicks.

Related

Keep reading

Resource

Analytics keywords report

Analytics-side keyword reporting.

Resource

Keyword monitoring

Ongoing tracking after analysis.

Resource

Seo analysis of competition

Competitor lens on the same data.

Resource

Keyword ranking api

Programmatic rank tracking.

Research

Anatomy Of Good Meta Ad Library

Adjacent competitive audit framework.

Keyword analysis ranks pages. Shuttergen ranks ads.

Once your keyword analysis surfaces the winners, Shuttergen helps you generate paid creative for those same intent groups - so your organic + paid coverage compounds instead of competing.

Generate creative freeAudit framework →