Most people use “keyword research” and “keyword analysis” interchangeably, and the conflation causes real problems. Research is generative — it produces a long list of terms people search. Analysis is evaluative — it decides which of those terms are worth your time. Skipping the second step is why so many content plans consist of high-volume keywords the site will never rank for, alongside long-tail terms that will rank and never produce a customer.
The Distinction
Keyword research answers: what do people search for in this space? Output is a list, often thousands of rows.
Keyword analysis answers: which of these should we act on, in what order, and with what kind of page? Output is a prioritised, much shorter plan.
Research is largely a tooling exercise — plug in seeds, export results. Analysis is judgment, and it is where the value is. Our guide to how to do keyword research covers the generative half; this covers what to do with the list afterwards.
The Four Dimensions
Every keyword should be assessed on four axes. Weakness on any one can disqualify a term regardless of how good the others look.
1. Volume — How Many People Search It
The obvious metric, and the one most over-weighted.
What to check: average monthly searches, plus the trend. A term at 2,000/month declining 30% year on year is worth less than one at 800/month growing.
Cautions: volume figures are estimates and vary between tools. Seasonal terms show misleading annual averages — a term that gets 12,000 searches in November and nothing else averages 1,000/month and is not a steady traffic source.
Do not disqualify low volume. A 90/month keyword that converts at 10% for a $5,000 service is worth more than a 5,000/month keyword that converts at nothing.
2. Difficulty — Can You Realistically Rank
The score is calculated identically for everyone and knows nothing about your domain, which is why it is so often misread. A KD 25 term is trivial for a DR 70 site and a serious project for a DR 15 site.
What to check: the difficulty score as a first filter, then the actual SERP. Look at the domain ratings of the ranking pages, whether any weak result is holding a position, and whether the top results are genuinely comprehensive or merely well-linked.
Our guide to keyword difficulty explained covers reading the score against your own authority, including a rough ceiling by domain rating.
The three-minute SERP check beats the score every time. If the top ten are all major brands, the score is understating difficulty. If there is a five-year-old un-updated post at position six, it is overstating it.
3. Intent — What the Searcher Wants
The dimension most often skipped, and the one that most reliably wastes effort when ignored.
What to check: search the term and look at what ranks. Blog posts mean informational. Comparisons mean commercial investigation. Product and service pages mean transactional. A map pack means local.
Why it matters: a blog post targeting a transactional query will not rank, because Google is showing service pages for that query. The page is not bad; it is the wrong type. This mismatch is covered with examples in SEO keyword examples.
Also watch for intent that has shifted. Terms change meaning over time, and a SERP that used to show guides may now show tools.
4. Business Value — Would Ranking Actually Help
The dimension tools cannot measure and the one that should carry the most weight.
What to check:
- How close is this searcher to needing what you sell?
- What is the CPC? Advertisers paying $15 a click have established the term produces customers.
- Could you write a natural, non-forced call to action on this page?
- Does this searcher resemble your actual customers?
A useful test: imagine you rank number one tomorrow. What happens? If the honest answer is “some traffic, no enquiries,” the term is a poor priority regardless of volume.
This is where high-CPC vertical terms often beat high-volume generic ones. seo for plumbers at 2,200/month with a $12 CPC is worth far more to an agency than a 20,000/month term with a $0.05 CPC.
Scoring in Practice
A simple approach that works without over-engineering. For each keyword, rate 1–5 on:
- Volume relative to your niche’s norms
- Winnability — difficulty against your domain authority
- Intent match — does the SERP show the page type you can build
- Business value — proximity to revenue
Then weight them. For most businesses, business value and winnability should carry more weight than volume. A reasonable weighting is 30% business value, 30% winnability, 25% intent match, 15% volume.
The specific numbers matter less than the discipline of scoring all four rather than sorting by volume and calling it a plan.
Clustering Before Prioritising
Do this before you score, not after. Keywords that share an intent belong on one page, and scoring them individually inflates their apparent value.
how to do keyword research, how to conduct keyword research, and keyword research steps are one page, not three. Combine their volume and treat the cluster as a single opportunity. The SERP overlap test in how many keywords to target per page is the reliable way to decide what groups.
Skipping this step produces content plans with three articles competing for one query — keyword cannibalization built in from the start.
What Analysis Should Produce
A finished analysis is not a spreadsheet of thousands of terms. It is:
- A prioritised list of clusters, ordered by expected return
- A page type assigned to each — guide, comparison, service page, location page
- An owner URL for each cluster, existing or planned, which is your keyword map
- An explicit not-now list — terms you have deliberately deferred, with the reason
That last item saves more time than it sounds. Without it, the same unwinnable head terms get re-proposed every quarter by someone looking at a volume column.
Common Analysis Failures
Sorting by volume and taking the top twenty. Produces a plan of terms you cannot rank for.
Filtering to KD under 10 and taking everything. Produces a plan of terms nobody valuable searches.
Ignoring the SERP entirely. Produces pages of the wrong type.
Analysing keywords individually rather than as clusters. Produces cannibalisation.
Never revisiting. Difficulty changes as your domain grows. Terms out of reach last year may be viable now, and a stale not-now list is worth re-scoring annually.
Research Finds, Analysis Decides
The framework reduces to a single habit: before committing to any keyword, check all four dimensions, not just the one your tool sorts by.
Volume tells you the ceiling. Difficulty tells you whether you can reach it. Intent tells you what to build. Business value tells you whether reaching it matters. A term that fails any one of those is a term worth deferring, however attractive the others look.
Most wasted content budget comes from skipping this step — publishing against a list that was never evaluated, then wondering why traffic arrived without customers, or why good articles never ranked at all.
If you want a keyword strategy analysed against your domain authority and your actual commercial goals, our SEO services start with exactly that assessment. Reach out to the team at blogthememachine.com, and subscribe to our newsletter below for more practical SEO guides.
