How to Research Your Competition Before Starting a Business

How to Research Your Competition Before Starting a Business

Short answer: audit seven surfaces rather than one — search results, AI assistant answers, marketplaces, review ecosystems, communities, published pricing, and how competitors position themselves.

Then stop asking “who else does this?” and start asking “what do all of them fail at, and can I become the obvious answer for that?” Competitor research that produces a list of names has failed. Competitor research that produces a gap has succeeded.

There’s also a competitor almost nobody writes down: non-consumption. Doing it yourself, doing nothing, or simply living with the problem. In plenty of small markets, that’s the market leader by a wide margin.

Define competition properly first

Four tiers, and a serious analysis includes all of them.

Direct competitors — same solution, same customer. Easy to find, and rarely where the risk actually is.

Indirect competitors — a different solution to the same job. A meal-prep service competes with supermarket ready meals, the local takeaway and a slow cooker.

Substitutes — the customer moves the budget somewhere else entirely. A gym membership competes with a bike, a running app and a holiday.

Non-consumption — the customer does nothing, or does it themselves. Ask any accountant what they lose deals to. It’s a spreadsheet and a Sunday afternoon.

Most founders analyse the first tier obsessively and then lose to the third and fourth. Write down at least two examples of each before you start.

The seven-surface audit

Work through these in order. Each takes somewhere between twenty minutes and two hours.

Surface 1: Search results

Search your core commercial terms — the phrases a buyer would use, not the ones you would.

For each of the top ten results, record who they are (a real competitor, a directory, a marketplace, a publisher, an aggregator), what the page actually offers, and what it’s missing that a buyer would want.

The shape of the results page tells you a great deal on its own:

Surface 2: AI assistant answers

Buyers increasingly ask an assistant before they search, and what comes back is a shortlist of two or three names with reasons attached. If you’re not on it, you’re not in the consideration set at all.

Run your buyer’s real questions through the major assistants: best provider for a particular customer type in a particular place, how to choose within the category, alternatives to a named competitor.

Record four things:

• Who gets named, and whether it matches the search results. Often it doesn’t, and that gap is itself the opportunity. 
• What reasons the answer gives for each recommendation. Those reasons reveal which attributes the model found corroborated across multiple sources, which tells you what evidence exists publicly about each competitor.
• What sources get cited — directories, review platforms, trade press, competitor blogs. Those citation sources are your visibility targets.
• What the answer says about the category, especially the criteria it tells buyers to use. Those criteria are now shaping purchase decisions at scale.

The commercial point: assistants tend to recommend what is describable and corroborated. A competitor with clear service descriptions, consistent details across profiles, visible expertise and third-party mentions gets named. 

A better competitor with a vague website doesn’t. That asymmetry is one of the most exploitable gaps in small markets right now, mostly because so few businesses are working on it deliberately.

Run this quarterly rather than once. The answers move.

Surface 3: Marketplaces and platforms

Where they’re relevant to your category — Etsy, Amazon, Checkatrade, Bark, Fiverr, Upwork, Booking, TripAdvisor, and whatever is sector-specific to you.

Extract price ranges, review counts as a proxy for volume, how listings are titled, what the top sellers include that others don’t, and what buyers complain about in the midrange reviews.

Marketplaces are the fastest way to learn genuine market prices, because those prices are transactional rather than aspirational.

Surface 4: Reviews

This is where your competitors document their own weaknesses, for free, in public.

Pull the last fifty reviews for your three closest competitors and categorise every complaint. 

What you’re hunting for is patterns that repeat across several competitors, because those are category-level failures, and category-level failures are positioning opportunities.

A typical result looks like this:

Read that table carefully. Nobody is failing at the work itself. Everybody is failing at responsiveness, quoting honesty and communication. The winning position isn’t “better quality” — it’s a fixed quote, answered within the hour, with daily updates. That’s buildable, provable and immediately differentiating.

Surface 5: Communities

Forums, professional groups, local Facebook groups, subreddits, trade associations. Search competitor names and category terms inside them.

What you get here and nowhere else is unfiltered recommendation behaviour. When somebody asks “can anyone recommend a…”, who gets named, and what words do people use to justify the recommendation? Those words are your future copy, because they’re the language the market genuinely uses.

Surface 6: Pricing

Build a price map. Where prices aren’t published, get quotes as a customer — legitimately, and without wasting anyone’s serious time.

Record the price, what’s included, what’s excluded, any minimum commitment, and how the quote arrives: instantly, after a site visit, or after three days of chasing. How a quote is delivered is often as differentiating as the number on it.

Then plot the spread. Tight clustering means a commoditised market where you should compete on something other than price. A wide spread means positions are available at both ends, and you should pick one deliberately rather than drifting into the middle.

Surface 7: Positioning and entity clarity

For each main competitor, write one sentence describing what they claim to be. If you can’t, that’s a finding — a competitor with no clear position is beatable by one with a clear position.

Then check consistency: is their business information the same across their website, directory listings, review profiles and social accounts? Do they have named experts, published policies, visible proof? Thin, inconsistent information is a real competitive weakness in a market where both search engines and AI assistants are working to establish who a business is and whether it can be relied on.

The teardown, one page per competitor

For your top three, fill this in and keep it to a single page each.

• Positioning claim: what they say they are
• Actual strength: what customers genuinely praise
• Actual weakness: what customers repeatedly complain about
• Price and structure: headline price, what’s bundled, minimums
• Acquisition: where their customers come from — search, referral, marketplace, paid, physical, partnership
• Visibility footprint: search presence, how often assistants name them, review volume, third-party coverage
• Serving whom: their apparent core segment
• Ignoring whom: who they turn away, price out, or serve badly
• Time to copy: if they copied your intended difference, how long would it take them?

Sit with that last line. If a competitor could neutralise your entire advantage in a fortnight, you don’t have a strategy — you have a feature.

Reading whether the market is overcrowded

Overcrowding is rarely a question of competitor count. It’s a question of whether those competitors are differentiated from each other.

Many competitors all saying the same thing isn’t crowded, it’s undifferentiated — and that’s an opportunity. It’s also the most common state of small local markets.

Many competitors each clearly serving a distinct niche is genuine maturity. You’ll need a niche nobody has taken, or a materially better operating model.

Few competitors, all struggling means the demand or the economics don’t work. Find out which before assuming you’re simply smarter than they are.

Few competitors, all thriving means something is protecting them. Identify the barrier — licensing, relationships, capital, supply access — and work out honestly whether you can clear it.

Four gaps worth attacking

From the audit, look specifically for these four. They’re the ones a new business can realistically win.

The service gap. Everyone fails at the same non-core thing: responsiveness, communication, cleanliness, punctuality, follow-up. Cheap to fix, immediately visible to customers, and surprisingly hard for an established firm with entrenched habits to copy.

The segment gap. A specific customer type is being served generically. Specialising for them beats being generally good at everything.

The visibility gap. Competitors deliver well but describe themselves badly, so their expertise is invisible to search engines, AI assistants and first-time buyers. Being the clearest, best-documented, most consistently described business in a category is a real advantage, and an unusually durable one, because catching up takes competitors months of sustained effort.

The proof gap. Everyone makes claims and nobody evidences them. Published data, real case examples, transparent pricing, guarantees and named expertise convert far better than adjectives.

Three of those four require no capital at all. They require noticing.

What to produce at the end

Not a list of competitors. Three things:

1. A one-page positioning map showing where each competitor sits and where the empty space is.
2. A ranked list of the category-level failures you intend to fix.
3. A single sentence: “Unlike [competitors], we [specific difference], which matters to [segment] because [consequence].”

If you can’t write that sentence yet, don’t launch. Write it first — Why Would Customers Choose You Over Everyone Else? is the guide for exactly that.

Frequently asked questions

How many competitors should I research in detail?

Three in depth, and roughly ten at a glance. Beyond that you get diminishing returns and a growing risk of paralysis. The three in depth should be your most likely head-to-head rivals; the ten at a glance exist to reveal category-level patterns, which is where the actual opportunity usually sits.

What if I can’t find any competitors at all?

Widen the definition before drawing conclusions. Include indirect solutions, substitutes and doing nothing — then ask specifically who tried this and stopped, because that’s the question with the most valuable answer. A market with no competitors is far more often unservable than unserved, and the cost of establishing which is a couple of afternoons.

Is it ethical to get quotes from competitors?

It’s normal commercial practice, within limits. Requesting published information, getting a standard quote or reading public reviews is fair. What isn’t fair is wasting a small business’s time with a lengthy fake enquiry, misrepresenting yourself to extract confidential terms, or taking up a site visit you have no intention of proceeding with. Keep it to what you’d be comfortable having them read about later.

How often should I redo this research?

The full seven-surface audit once a year, and the AI assistant surface quarterly, since those answers shift fastest. Reviews are worth a monthly skim — a sudden pattern of complaints about a competitor is a live opportunity, and a sudden pattern about you is an early warning.

My competitors are much bigger than me. Is that hopeless?

No, and size is often a constraint disguised as an advantage. Larger competitors carry overheads, procurement processes, standardised offerings and slow decision cycles, all of which prevent them from serving smaller or unusual customers well. 

Look for what they can’t profitably do: the small job, the odd requirement, the customer who wants to speak to the owner. Attacking a large competitor at their strongest point is what’s hopeless — not competing with them at all.

Should I tell customers how I’m different from a named competitor?

Compare on dimensions rather than by name, in most cases. Naming rivals gives them free attention, invites arguments about accuracy, and can read as insecure. Saying what you do — fixed quotes, response within an hour, a named person on every job — lets buyers make the comparison themselves, which is more persuasive anyway because they reached it on their own.

What’s the fastest version of this if I only have a day?

Do surfaces one, two and four: search results, AI assistant answers, and the fifty-review complaint audit. Together they take about four hours and give you the results shape, the current recommendation set and the category-level failures. That’s enough to draft a positioning sentence, which is the actual deliverable. The remaining surfaces sharpen it rather than change the answer.

How do I know if a gap is real or just unprofitable?

Check whether anyone previously served it and stopped, and whether the customers in that gap can pay. 

A genuine gap has customers with budget who are currently being served badly. An unprofitable gap has customers who can’t pay enough to cover the cost of serving them — which is precisely why the space is empty. If a gap looks obvious and nobody has taken it, assume the second explanation until you’ve disproved it.