A competitor teardown you could hand to an investor
Most competitor analysis is a grid of adjectives — strong brand, good UX, aggressive pricing — that nobody reads twice and nothing follows from. A useful teardown is six facts you can check, each with a link and a date beside it. Here's where each of those six actually shows up, a filled-in example for the business we use across this site, and the blank template it follows.
A competitor teardown is worth doing when a decision is waiting on it — what to charge, which customer to go after, what to say on the homepage, whether to bother entering at all. It's worth almost nothing as a document produced because a template said to produce one. The difference shows up in the last line of every row: if you can't write "so we will…", you've written a description rather than a finding.
Two rules make the rest of this page work. Three to five competitors, not twelve. Twelve shallow profiles beat nothing and lose to three deep ones, and you'll never update twelve. Pick the two doing the most of what you want to do, the one doing it cheapest, and the substitute — doing it themselves, doing nothing, or the tool they already pay for — because the substitute usually has the largest market share and never appears on anyone's list. And every cell gets a source and a date. Prices change, review counts move, the careers page empties. A teardown without dates rots quietly and you keep making decisions from it.
The six things worth knowing
Everything else is decoration. These six, for three to five competitors, is a complete piece of work — and it's half the answer to "should I start this". The other half, sizing the market itself, is on its own page.
- 1 · What they charge
- The number, what's included at that number, and what it costs to get to the next tier. Plus whether they publish it at all — a competitor who quotes by phone has a price that moves, and that's a fact about them worth knowing.
- 2 · What their customers complain about
- Not the star rating. The specific thing that repeats in the three-star reviews. Three competitors sharing one complaint is a description of the gap you're walking into, written by the people who'll pay you.
- 3 · Whether they're growing
- Hiring, opening, filing, posting — or quietly not. A careers page with four roles on it and a company registration in good standing says something a revenue estimate never will.
- 4 · How they get customers
- Ads, search, referral, footfall, a partnership, or one salesperson. This is the one that changes your plan most, because it tells you which channel is already crowded and what it costs to be there.
- 5 · Who they're for — and visibly aren't
- Read their homepage for who it excludes. The customer they turn away, price out or serve badly is the one you can have, and it is usually the smaller, the more urgent or the more awkward one.
- 6 · What they'd do if you took a customer
- Can they cut price and survive? A shop with a lease and four staff can't match a one-person van on price for long, and knows it. This is the difference between a competitor who'll ignore you for two years and one who'll undercut you next month.
What to leave out: revenue estimates from traffic tools (see the note below the table), anything described as "brand strength" or "innovation", a feature matrix with thirty rows and ticks in twenty-eight of them, and a market-share pie chart for a market nobody measures at your scale. If a line can't be checked by someone else in five minutes, it doesn't go in.
Where each one actually shows up
Everything in this table is free and public. The only things you'll need an account for are LinkedIn and a throwaway email address for their mailing list. Work down it with a browser open; a competitor takes about forty minutes the first time and fifteen on each refresh after.
| What you want | Where it is | What to do |
|---|---|---|
| Their list price | Their own site | Try /pricing, /services, /rates and /book — then walk their booking or quote flow as a customer to the last screen before payment. That screen is where the real price and the surcharges live. |
| The price when they don't publish one | Google Business Profile · Yelp · Thumbtack, Angi, Houzz | The Services tab on a Google Business Profile and the marketplace listings often carry a price the website doesn't, and are frequently older — which shows you the direction of travel. Failing that, request a quote as a normal customer, from a normal address. |
| How their prices have moved | Wayback Machine | Put web.archive.org/*/theirsite.com/pricing in the address bar — an asterisk in place of the date lists every snapshot ever taken. Two years of price history in about five minutes, and it works on their homepage, team page and product pages too. |
| Their discounts and current offer | Their email list · ad creative | Join the list with a spare address. Whatever the ad or the welcome email offers is the real price, not the one on the page. |
| What customers actually complain about | Google Maps reviews | Sort by Lowest rating, then read the three-star reviews before the one-stars. One-stars are often a single bad day; three-stars are written by people who were basically satisfied and still had a specific complaint. Tally the reasons in a spreadsheet. |
| The reviews that were filtered out | Yelp | Scroll to the bottom of a Yelp page and open the "not currently recommended" reviews. Often a noticeably different picture from the ones above it. |
| Formal complaints and unfiltered opinion | BBB · Reddit · local Facebook groups | Search their name on the BBB, then run a Google search for site:reddit.com plus their name. In regulated trades, add the state licensing board's complaint search. |
| Whether they're hiring, and for what | Their /careers page · Indeed · LinkedIn Jobs | A job post is the most candid document a company publishes. It names the tools they use, the shift pattern, the size of the team, and — in the growing number of states with pay-transparency rules — the pay band. Four open roles means growth; a page that's been empty for a year means something else. |
| Whether the headcount is going up | LinkedIn company page | The employee count on their company page. It's approximate, and the useful part is the change — screenshot it once a month and you have a growth curve nobody sells. |
| How busy they actually are | Google Maps | Popular Times on their Maps listing, and the rate at which new reviews arrive — count reviews in the last 90 days against the 90 before. For a physical business these are the best free demand proxies that exist, and both are real behaviour rather than an estimate. |
| When they started and whether they're current | Your state's business registry | Search their name for the formation date, status, registered agent and officers. A registration that lapsed into delinquent status, or an entity two years younger than the brand suggests, is a real finding. |
| Every ad they are running right now | Meta Ad Library · Google Ads Transparency Center | Free, no login, and the most underused pair of tools in this table. Search their page or domain: you get every live Facebook and Instagram ad with the date it started, and every live Google search, display and YouTube ad. An ad that's been running for four months is an ad that works — that's their best angle, tested with their money. |
| Whether anyone finds them in search | Google, incognito · their sitemap | Search your five most valuable phrases in an incognito window and write down where they appear and whether they paid to be there. Then open theirsite.com/sitemap.xml and count the pages — a site with 200 pages is running a content strategy; one with 6 is not. |
| Roughly how much traffic they get | Similarweb · Semrush | Both give a free estimate for any domain. Treat it as a magnitude, not a measurement — below roughly 5,000 visits a month the estimates are close to guesses, which for most local competitors means don't use them at all. Never derive a revenue figure from one. |
| Their stack, and whether they can take a card | View source · Wappalyzer or BuiltWith | Right-click, View Source, and search for stripe, shopify, squarespace, calendly. A competitor whose booking page emails a form rather than charging a card is a competitor losing bookings at 9pm on a Sunday. |
| Real financials, when they exist | SEC EDGAR full-text search · ProPublica Nonprofit Explorer | Public company: EDGAR has the filings and full-text search will find your competitor named inside someone else's. Nonprofit: the Form 990 on Nonprofit Explorer gives revenue, expenses and the top salaries. Private and small: there are no real financials, and anyone selling you some is selling a model. |
A filled teardown
This is one page for one competitor, in the format the template below sets out. Northline Bike Repair is the business we use across this site — one founder, a van, mobile bike repair in Denver. The competitor is invented so that nobody real gets torn down on a marketing page; every source named in the right-hand column is real, and each one is where you would genuinely have found that line.
| Line | What we found | Where |
|---|---|---|
| Who they are, and why they're on the list | A twelve-year-old bricks-and-mortar shop three miles from the centre of our service area, six staff, sells bikes and services them. On the list because they own the phrase "bike repair Denver" in search and because half our target customers have already been there once. | Their site, Colorado Secretary of State registry (entity formed 2014, status good standing) |
| Price | $85 basic tune-up, $185 full overhaul, both listed on their booking page. Parts extra and not itemised until the quote. No mobile service, no call-out. Their booking page took a card, which the two other mobile operators don't. | Their /service page, walked to the last screen before payment |
| Price direction | The same tune-up was $65 in 2022 and $75 in 2024. Up roughly 30% in four years — faster than costs, which means they've had pricing power and probably still do. | Wayback Machine snapshots of their service page |
| The complaint that repeats | Turnaround. Nine of the fourteen three-star reviews in the last two years say a version of "had my bike for two weeks". Nobody complains about the work itself or about price. That is a queue, not a quality problem — and a queue is the single most attackable thing on this page. | Google Maps reviews, sorted lowest, three-stars read first |
| Direction | Two open roles on Indeed (mechanic, weekend counter), pay band listed at $22–$26/hr. Review count up from about 6 a month to about 9 over the past year. Growing, and hiring to clear the backlog rather than to open anything new. | Indeed, their careers page, review counts by month from Google Maps |
| How they get customers | Search and footfall, not paid. Nothing in the Meta Ad Library, nothing in the Google Ads Transparency Center. They rank first organically for four of our five phrases, with 60-odd pages on their sitemap and a blog updated monthly. Paid search around them is therefore cheap and uncontested; organic is a two-year fight. | Meta Ad Library, Google Ads Transparency Center, incognito search, their sitemap.xml |
| Who they're for — and aren't | For the enthusiast with a $3,000 bike who will wait two weeks for it to be done properly. Not for the commuter with a flat and a Tuesday, and not for anyone who can't get the bike into a car. Their homepage sells expertise; it never once mentions speed or convenience. | Their homepage and about page, read for what they exclude |
| What they'd do if we took a customer | Very little, at first. A lease and six staff means their price floor is high and a mobile van isn't their competition — until we take enough of the simple, profitable, quick jobs that their mix worsens. Then they'd probably add a collection service rather than cut price, because cutting price breaks their model. | Inference from the cost base above — labelled as inference, not fact |
| So we will… | Sell the two things they structurally can't: same-week and at-your-door. Price the basic tune-up at $60 plus a $25 call-out, which lands under their $85 while earning more per hour because there's no shopfront. Bid on their brand-adjacent search terms while paid search here is empty. Do not compete on overhauls. | — |
| What we couldn't find out | Revenue, margin, and whether the two-week wait is capacity or deliberate. Traffic estimate tools disagreed by 4× on their domain, so we ignored them. Written down rather than guessed. | — |
Three lines in that table did all the work. The repeating complaint gave Northline its positioning, and it came from reading fourteen reviews rather than from a strategy session. The empty Ad Library entry said the paid channel is uncontested, which is worth more than any traffic estimate. And the cost-base line explains why the strategy is safe for about eighteen months — the thing Northline is selling is the thing a leased shop with six staff cannot copy without becoming a different business. Everything else on the page is context.
The template, blank
One page per competitor. Ten lines, in this order, each with the source and the date it was checked. Three to five pages is a complete competitor analysis and takes an afternoon.
- Who they are, and why they're on the list. One sentence each. If you can't say why they're on the list, take them off it.
- Price. The number, what's included, what the next tier costs, and whether they publish it.
- Price direction. What the same thing cost two years ago —
web.archive.org/*/theirsite.com/pricing. - The complaint that repeats. The thing three or more reviews say, in the reviewers' words. Not the star rating.
- Direction. Hiring or not, review velocity up or down, registration current or lapsed — with the evidence.
- How they get customers. The channel, and how you know: ads found or not found, search positions, footfall, partnerships.
- Who they're for — and visibly aren't. Read their homepage for who it turns away.
- What they'd do if you took a customer. Their cost base, and whether they can cut price and survive. Label it inference, because it is.
- So we will… One sentence. What you do differently because of the nine lines above. If it's blank, the page was a waste of an hour.
- What we couldn't find out. Write the gaps down instead of leaving them blank, so the next person doesn't assume you checked.
Four rules that keep it honest
Date every cell, because a price from March is a different fact in September. Write "couldn't find" rather than an estimate wearing a confident font. Don't score competitors out of ten — a composite score hides the one dimension that matters behind nine that don't; instead rank them on the single thing you actually compete on, and say what it is. And refresh on a cadence rather than never: prices and ads every quarter, reviews monthly if you're using them as a demand signal, everything else when something visibly changes.
The version you hand to an investor
Not ten pages. One table: competitors down the side, and across the top only the dimensions a customer actually chooses on — for Northline that's price, wait time and whether you have to travel, and nothing else. Then one line underneath, which is the line the whole document exists for: why they can't easily copy the thing you're better at. If the honest answer is "they could, in a month", an investor will find that out in the meeting, and it's better to have written it down yourself.
Common questions
How many competitors should I analyse?
Three to five. The two doing the most of what you want to do, the one doing it cheapest, and the substitute — customers doing it themselves, doing nothing, or using a tool they already pay for. The substitute usually has the biggest share of the market and almost never appears on anyone's list. Twelve shallow profiles beat nothing and lose badly to three deep ones, and nobody has ever updated twelve.
How do I find a competitor's prices if they don't publish them?
Four places before you resort to asking. The Services tab on their Google Business Profile, their listing on a marketplace like Thumbtack, Angi or Yelp, old versions of their pricing page in the Wayback Machine, and the last screen of their own booking or quote flow. If all four are empty, request a quote as an ordinary customer from an ordinary address — and note that a competitor who won't publish a price has a price that moves, which is itself worth writing down.
Can I see what ads a competitor is running?
Yes, for free and without an account. The Meta Ad Library shows every Facebook and Instagram ad any advertiser is currently running, with the date each one started. The Google Ads Transparency Center shows their live search, display and YouTube ads. An ad that has been running for months is their best-performing angle, tested with their money — and an empty result tells you the paid channel around them is uncontested.
How accurate are website traffic estimates for competitors?
For large sites, roughly right in magnitude. For a local business with a few thousand visits a month, they can be out by a factor of five, and two tools will routinely disagree with each other by 4×. Use them for order of magnitude at most and never derive a revenue figure from one. Counting how many new Google reviews a competitor gets each month is cruder, free, and closer to the truth.
What should I do with a competitor's bad reviews?
Read the three-star ones first, not the one-stars. A one-star is often a single bad day — a missed appointment, a rude phone call — while a three-star is written by someone who was basically satisfied and still had a specific complaint, which means it's about the product rather than the mood. Tally the reasons across all your competitors. A complaint that repeats at three different companies is a structural gap and it's the best positioning you'll find anywhere.
Is a SWOT analysis worth doing?
Only if every cell contains a checkable fact with a source and a date. "Strength: established brand" is not a finding, it's a mood. "$185 for the full service, listed on their booking page, checked 4 September" is. Most SWOT grids fail because adjectives are easier to write than facts, and nothing can follow from an adjective.
How often should I redo a competitor analysis?
Prices and live ads quarterly — both change quietly and both change your pricing. Reviews monthly if you're using review velocity as a demand signal. Hiring, filings and positioning only when something visibly changes. The point of dating every cell is that a refresh becomes fifteen minutes of updating rather than an afternoon of starting again.
Hand the reading to someone else.
Ask Velofound to research a competitor and the teardown lands in your inbox with the sources under every line — beside a website that's already live and taking payment. Free to start.
Start free →