Cybrial
Independent design. Ongoing growth.Manchester · Working everywhere

Guide · Getting found online

How to Do a Competitor Analysis That Changes What You Do

By Measured 4 October 2026

The point is not to find out what your competitors are doing. It is to find the thing none of them is doing.

The short answer

A useful competitor analysis answers six questions and ignores everything else: who actually ranks for the searches that matter (not who you think your rivals are), which queries they cover that you do not, what they charge and whether they say so, what they prove versus what they claim, how fast their site is on a phone, and where their reviews come from. Most of that is free to find. The output should be a short list of gaps you can act on this month — not a scored matrix. If the analysis does not change what you build next, it was an exercise.

Your competitors are probably not who you think

Start here, because almost every analysis starts with the wrong names.

The businesses you compete with for customers are often not the ones you compete with in search. Search a handful of the terms a customer would actually type and write down who is on page one. That list will usually include directories, a national brand, a marketplace and two local firms you have never heard of — and may well not include the rival you have been watching for years.

Both lists matter and they are different jobs. The commercial list tells you about pricing and positioning. The search list tells you who is standing between you and the enquiry. Keep them separate and label them, because conflating them is how businesses end up benchmarking against a competitor who gets no traffic.

Use an incognito window, and remember results are personalised by location. What you see from your own desk is not what a customer twelve miles away sees.

The six things worth measuring

What they rank for that you do not

The most actionable of the six. Take the pages on their site — their sitemap at /sitemap.xml usually lists every one — and read the titles. You are looking for topics, services or locations they have a page for and you do not. A competitor with twelve service pages against your one is not better at marketing; they have simply given twelve searches somewhere to land.

What they charge, and whether they say so

Check every competitor for a published price. In most trades and professional services almost nobody publishes, which makes doing so a genuine differentiator rather than a risk. Where everybody publishes, you need to know where you sit before a customer tells you.

What they prove, against what they claim

Count the specific, checkable facts on a competitor's site: named clients, dated figures, real photographs, numbers with sources. Then count the adjectives. Most sites are almost entirely adjectives, which is the opening — evidence is the one thing a competitor cannot copy from you by Friday.

Speed on a phone

Run their homepage and their main service page through a free page-speed tool, on mobile. If they are slow and you are fast, that is a real advantage you can measure and they probably cannot see.

Reviews: how many, how recent, where

Count on the Google Business Profile, not on their website. Recency matters more than total: forty reviews ending two years ago is weaker than twelve from this quarter. Read the one-star ones — those are a free list of what customers in your market actually complain about.

What they are paying for

If a competitor is advertising on a term, somebody has decided it converts. The ads above the organic results tell you which services are worth money in your market, and the public ad libraries let you see what they are running on social. That is paid-for market research they did and you did not.

Getting actual numbers, cheaply

Most of the above is observation. Two things are worth having real figures for, and both are obtainable.

Search demand and difficulty for the terms in your market. You can approximate this free with Google Keyword Planner inside an Ads account, and get it properly from a paid tool. The figure that matters is not volume alone but volume beside difficulty — a term with 1,000 searches that you cannot win is worth less than one with 150 that you can.

And what advertisers pay per click. This is the most underrated number available to a small business, because it is the market telling you what an enquiry is worth to somebody with a budget. A £60 click means people are confident about the value of that customer. A 40p click means they are not.

Same kind of service, same country, same month. Without the difficulty column, the second looks like the easier market because the number is smaller. It is the hardest term in the research.

If that pattern holds in your market — and in our experience it usually does — then the single most valuable output of a competitor analysis is a list of phrases with real demand that nobody has written a page for.

What the output should look like

Not a matrix. A scored grid comparing nine competitors across twelve attributes is a document that gets admired once and never opened again.

What you want is a page of sentences in this shape: "Nobody in this town has a page about X, and X is searched Y times a month." Each one is a decision you can act on.

Sort them by whether you could do it this month. A gap that needs a rebuild is worth knowing about; a gap that needs an afternoon is worth doing before you finish reading.

Then set a date to repeat it. Twice a year is plenty for most businesses, and it is the comparison over time that tells you something — a competitor who added forty pages since the spring has changed strategy, and that is worth noticing early.

Three ways this goes wrong

Copying. The point is to find the gap, not to match the leader feature for feature. A smaller business that does one thing visibly better beats a worse copy of the market leader, and copying hands them the terms of the comparison.

Benchmarking against the wrong business. If the firm you measure yourself against is venture-funded, or national, or twenty years older, the comparison tells you nothing you can act on. Compare against businesses of roughly your shape, and separately note what the giant is doing that you simply cannot.

Treating it as a one-off. A competitor analysis is a snapshot of a market that moves. The version done once at the start of a redesign and never repeated is the one that produces the "we did that research already" conversation three years later.

A two-hour version

  • Search six terms a customer would type, incognito. Write down who appears.
  • Open each competitor's /sitemap.xml and list the pages you do not have an equivalent of.
  • Note who publishes prices and who does not.
  • Count specific checkable facts versus adjectives on the two biggest competitors.
  • Run two of their pages through a mobile speed test, and your own alongside.
  • Count reviews on each Google Business Profile and check the date of the most recent.
  • Read every one-star review in your market. Write down the three complaints that repeat.
  • Note which terms have ads above them.
  • Write the gaps as sentences. Mark the ones you could do this month.

Item seven is the one people skip and the one that most often changes something. Your competitors have already paid for that feedback; it is sitting in public.

Sources

Everything this page relies on.

  1. The roofing and charity difficulty figures are from our own twenty-sector research, published in full on this site's industries pages, pulled 4 September and 4 October 2026.

  2. Tool behaviour — what Keyword Planner shows without active spend, what the ad libraries expose — changes. Treat the method as durable and check the specific tools against their own current documentation.

Common questions

Questions people actually search for.

Who counts as a competitor?

Two different lists. The businesses you compete with for customers, and the ones that actually appear for the searches a customer makes — which will often include directories, a national brand and local firms you have never heard of. Keep them separate; benchmarking against a rival who gets no traffic is a common waste of an afternoon.

What tools do I need?

Less than you would think. An incognito browser window, each competitor's /sitemap.xml, their Google Business Profile, a free mobile page-speed test and the public ad libraries will get you most of the way. Keyword Planner inside a Google Ads account gives rough demand free; a paid tool gives you demand beside difficulty, which is the pairing that matters.

Why does difficulty matter more than volume?

Because a term you cannot win is worth nothing regardless of its size.

What should the analysis actually produce?

A page of sentences shaped like "nobody here has a page about X, and X is searched Y times a month", sorted by whether you could act this month. Not a scored matrix — those get admired once and never reopened.

How often should I redo it?

Twice a year for most businesses. The value is in the comparison over time: a competitor who has added forty pages since the spring has changed strategy, and you want to notice that early rather than eighteen months later.

What is the single most useful thing to look at?

The one-star reviews of every competitor in your market. It is a free, specific list of what customers locally actually complain about, your competitors paid for it by receiving it, and it is sitting in public.

Who is actually in your way?

Send me your site and the six terms you think matter. I will tell you who really ranks for them, which pages your competitors have that you do not, and which gaps are an afternoon rather than a rebuild.