How to Use Keyword Research Tools Properly
This is not a guide to which one to buy. It is a guide to what to distrust in all of them, because every number in every keyword tool is an estimate the tool produced. None of it comes from Google.
What They Actually Do
You type a phrase in. The tool returns a list of related searches with numbers beside each one. That is the whole product. Understanding it in those terms immediately changes how much weight you give to what comes out.
The genuinely valuable half. The suggestions.
Discovering phrasings you would not have thought of is real value. People describe problems in ways a business never would. A tool surfaces those.
The half to be careful with. The numbers.
Volume, difficulty, competition and everything else in the columns. All calculated by that tool, using its own method, which block two covers.
What determines your results. The phrase you typed.
A tool cannot suggest what you did not point it at. Give it a trade term and it returns trade terms, which is how businesses end up researching their own vocabulary.
What it cannot do. Judge.
Nothing in the output knows which searches lead to work you want, at margins you accept, in places you serve. That judgement stays with you.
What this page is not. Anti-tool.
Tools are useful and block five sets out where. The problem is treating the output as fact rather than as a prompt.
Every Number In Them Is An Estimate
Google does not publish how many people search for things. So no tool is reading a real total. Each one produces its own figure using its own method, then presents that calculation as though it were a measurement. That single fact explains almost every confusing thing about these products.
Where the figures come from. Modelling.
Sampled behaviour, patterns across similar searches, data bought from elsewhere and assumptions filling the gaps. Reasonable work, though not a count.
The clearest proof. They disagree.
Run the same search through two tools and the numbers differ. If either were reading a real total they would match, so their disagreement is the demonstration.
Why they look authoritative. Precision implies certainty.
A specific number in a clean column reads as fact. Presented as a range with a note about method, nobody would treat it the same way.
What they are still good for. Relative scale.
Whether one search is much larger than another is usually right, even when both figures are wrong. That comparison survives the estimation.
What they are not good for. Forecasting.
Multiplying a volume figure by an expected share to predict visitors produces a number with no basis. It appears in proposals constantly.
Where each figure gets examined. Its own guide.
Our metrics guide works through volume, difficulty and the rest one at a time.
Why We Name None Of Them
You will not find a single keyword tool named anywhere on this site, paid or free. That is a decision rather than an oversight. A guide about tools that names no tools ought to explain itself rather than leave you wondering.
The first reason. Recommendations date.
Products change hands, pricing changes, features are removed and quality moves. A recommendation written today is a liability within a couple of years and nobody goes back to correct it.
The second reason. It reads as endorsement.
Naming one implies we stand behind it. We are not in a position to vouch for somebody else's product on an ongoing basis. Pretending otherwise would be improper.
The third reason. The money.
A great deal of tool recommendation online is paid for through referral arrangements. Since we take none, the simplest way to stay clear of that entire question is to name nothing.
The fourth reason. It distracts.
Which tool you use matters far less than whether you understand what its numbers are. A business with a clear head and free sources beats one with a subscription and misplaced confidence.
What we will tell you. What to look for.
Any tool that shows its suggestions clearly and does not present its estimates as certainties is adequate. That is the whole specification.
What this means practically. Ask us directly.
If you want a view on a specific product, that is a conversation rather than a published recommendation with our name on it forever.
The Free Sources That Beat Them
For a small business covering a handful of subjects, the sources that cost nothing will get you further than a subscription. They are also more direct, because several of them are statements of what people actually type rather than estimates of it.
The best one. Your own phone.
The questions customers ask every week, written down in their words. Demonstrated demand, in the exact phrasing, at no cost.
The search box itself. Direct evidence.
Suggestions appearing as you type are drawn from what people actually search. That is not modelled and it is not an estimate.
The foot of the results page. Related searches.
What else people looked for around the same subject. Free, immediate and frequently the source of the best long tail material.
Your own reporting. The best of all, eventually.
Once pages have been live a while, the searches your site actually appears for beat any external estimate. Our guide covers using it for research.
Where the full list lives. Its own page.
Our free tools guide covers each source, what it does well and its specific limit.
What none of them give you. Volume figures.
Which matters less than people expect, since block two established what those figures actually are.
Where Tools Genuinely Help
There are situations where a paid tool is straightforwardly worth the money. Recognising them stops this page reading as a blanket dismissal. It also lets you judge whether your own situation is one of them.
Scale. The main one.
Researching hundreds of subjects across a large site by hand is impractical. That is what these products were built for and they do it well.
Unfamiliar territory. A new sector.
Working in a trade you know nothing about, the suggestions genuinely teach you the vocabulary faster than anything else available.
Competitor discovery. Genuinely useful.
Seeing which subjects other sites appear for surfaces gaps you would not have found. Our guide on opportunities they miss covers that.
Speed. Hours saved.
Gathering phrasings by hand takes an afternoon. A tool does it in minutes, which matters when the work is somebody's job rather than an evening task.
What they never supply. The commercial judgement.
Which of these subjects your business should own remains a decision made by people who know the business, whatever the software cost.
The fair summary. Time, not insight.
A tool buys speed. It does not buy better decisions. A business expecting the second will be disappointed by the subscription.
Where They Mislead
These are the ways tool output leads businesses wrong. They are consistent enough across products to be worth listing as a set.
Rounded bands. Concealing smallness.
Figures grouped into ranges hide how tiny or uncertain a number really is. Two searches shown identically can be very different in reality.
Difficulty scores about nobody. The commonest.
A difficulty figure is calculated without knowing anything about your site. The same search is easy for one business and impossible for another. One number cannot express both.
Everything looks like demand. The filtering problem.
Nothing in an export distinguishes your customer from a student or a competitor. Our method guide covers how much of a raw list that removes.
The seed trap. You get back what you gave.
Start from your trade term and the tool confirms your vocabulary rather than correcting it. Start from a customer phrasing and the results change completely.
Volume beats value in the sorting. Structural.
The interface invites you to sort by the largest number, which is almost always the wrong ordering for a small business.
Zero does not mean nobody. Especially locally.
A search reported as having no volume can still bring work every week, which is block seven.
The Local Problem
Keyword tools are least reliable exactly where local businesses need them most. Small volumes are the hardest thing to estimate, so the searches that actually produce work for a local firm are the ones the numbers describe worst.
Why small numbers break estimation. Nothing to model from.
Estimating a large national search from sampled behaviour is achievable. Estimating a specific service in one town has almost nothing to work from.
What that produces. Searches shown as empty.
A phrase reported as having no meaningful volume, which a local business knows perfectly well brings enquiries, because they answer them.
The decision that follows. A page not built.
Somebody looks at the figure, concludes there is no demand and does not write the page. That is the concrete cost of trusting the number.
Why the low figure can still be worth it. Concentration.
A small number of searches from people in your area needing your service converts at a rate no broad term approaches. Few visitors, most of them buyers.
What to trust instead. Your own knowledge.
If customers ask for it, the demand exists whatever the column says. Our local research guide covers working without reliable figures.
The rule we apply. Never delete on volume alone.
A search with a low figure and obvious commercial logic stays on the list. The figure is the least reliable thing about it.
Do You Need To Pay
For most of the businesses reading this, no. A subscription is worth it when the work is continuous and large. For a local firm covering a handful of subjects it buys speed you do not need and confidence you should not have.
When paying makes sense. Continuous work.
Somebody researching most weeks across many subjects. Or an agency working across several clients. Then the time saved genuinely exceeds the cost.
When it does not. A one off exercise.
Working out what to write about once and then building the pages. The research part is over in an afternoon and the subscription runs for a year.
What we would suggest instead. A trial.
Do the exercise with free sources first. If you hit a genuine wall, you will know exactly what you need a tool to do rather than buying one hopefully.
What to check before subscribing. The output format.
Whether you can get the suggestions out in a form you can group and annotate. That matters more than the size of the database.
What not to buy it for. Reporting.
A tool that produces impressive documents you never act on is an expense rather than an investment. The value is in the suggestions.
Where the rest sits. The hub.
All twenty-two guides are on the keyword research guide.
None of those
numbers came
from Google.
Google does not publish how many people search for things, so no tool is reading a real total. Each one calculates its own figure using its own method and presents it in a clean column where precision reads as certainty. Run the same search through two of them and they disagree, which is the proof.
How we treat tool output:
A search showing no volume can still bring enquiries every week. Local demand is the hardest thing of all to estimate.
Every guide.
One subject.
What the words mean, what the numbers in a keyword tool actually are, how to find the searches worth targeting, how to turn a list into a publishing plan and what to do when a page you relied on slips.
