How to find your competitors' pricing — legally, from public sources

Pricing research fails in two opposite ways. Either you give up at the first "request a demo" and price by gut, or you get resourceful in ways that create real problems — booking sales calls under a made-up company, signing up as a fake buyer, scraping a portal you agreed not to scrape. The first leaves you blind; the second trades a small research win for a reputational and contractual risk that is never worth it.

There is a wide middle. Prices are sticky things: they get quoted in proposals, printed in tenders, mentioned in reviews, complained about in forums, and summarised by resellers. Most of what you need is already published by someone. The work is knowing where it surfaces and how to make numbers from different sources mean the same thing.

Start with what is simply published

Before the clever sources, exhaust the obvious ones properly, because people skim them and miss half.

• The pricing page, including its edges. Read the footnotes, the annual-versus-monthly toggle, the minimum seat counts, the "starting from," and what each tier excludes. The exclusions tell you where the real revenue is.
• Docs, help centres, and terms. Overage rates, add-on fees, support tiers, and cancellation terms often live in documentation long after they have been scrubbed from marketing.
• App stores and marketplaces. Mobile stores, plugin directories, and cloud marketplaces list in-app and subscription tiers publicly, sometimes at more granularity than the vendor's own site.
• Resellers and partners. Channel partners and integrators publish list prices the vendor keeps vague, because their job is to quote.
• Archived versions of the pricing page. Public web archives show what a price was a year ago. Direction of travel matters as much as the current number: a price that has doubled in two years is telling you about confidence and about who the customer now is.

Then the places prices leak

When there is no published number, look for the situations that force someone else to write it down.

• Public procurement and tenders. Government and public-sector contract awards are frequently published with values and vendors. If your competitor sells to any public body, there may be an exact contract figure in a public register.
• Review sites and their free-text fields. Reviewers routinely name what they pay, especially in the "what do you dislike" section, which is where price complaints live. Treat single figures as anecdotes and clusters as evidence.
• Community threads. Forums, subreddits, and trade groups produce "what are you all paying for X" threads on a loop. These are the most honest pricing data on the internet and the least tidy.
• Case studies, decks, and conference talks. A vendor's own case study sometimes states contract size, and a customer presenting their stack will occasionally put the annual cost on a slide.
• Job listings and RFPs from buyers. A buyer's RFP names budget ranges. A job listing sometimes names the tools and their budget owner.
• Filings, where they exist. Public companies disclose revenue and often average revenue per customer, which bounds what typical deals must be. That is not list price, but it is a hard ceiling on your guesses.

For local and service businesses specifically, the trail is easier: quotes get posted in local community groups, service-marketplace profiles show typical job values, and many trades publish call-out fees and hourly rates outright.

Make the numbers comparable, or they are useless

The most common mistake is not failing to find prices — it is comparing prices that are not the same thing. A number without its unit and its bundle is decoration.

Normalise every figure onto one line before comparing:

• Unit. Per seat, per site, per project, per month, per year, per unit of usage. Convert everything to one period and one unit, and state which.
• Bundle. What is included at that price — onboarding, support, the integration the buyer needs? A cheaper price with mandatory paid setup can be the more expensive option.
• Commitment and floor. Annual prepay, minimum term, minimum seats. A "$20 per seat" product with a ten-seat minimum has a $200 real entry price.
• Discount reality. List price is a ceiling in any market with sales-led deals. If reviews cluster well below list, list is aspirational.
• Confidence and date. Label every number: published, quoted in a review, inferred, or unverified — plus when it was true. A forum post from 2022 is a historical data point, not a current price.

Put those in a table with one row per competitor and one column per attribute, including the source URL for each figure. If a cell has no source, it is an assumption and should be marked as one rather than quietly rounded into fact. That distinction is the difference between a pricing table you can act on and one that just launders your guesses.

What to do with the picture

Competitor prices do not tell you your price. They tell you what the market has trained buyers to expect, and where the empty positions are. Read the spread rather than the average: a market with a cluster at the bottom, a cluster at the top, and nothing in the middle is describing a gap — and possibly describing why the middle does not work.

Also read what the price is attached to. If every competitor charges per seat and one charges per outcome, that is a positioning difference, not a pricing difference, and it is usually more interesting than the number. Where you sit relative to the cluster should be a deliberate statement, and if you land in the middle by accident, you have chosen the one position that says nothing.

The lines not to cross

Two categories of tactic tend to get suggested and are not worth it.

Deception. Booking a sales call as a fake prospect, misrepresenting who you are to get a quote, or creating an account under false details. Beyond the ethics, it burns a relationship in what is usually a small industry, and a screenshotted email lasts forever. The information gained is a single data point you could usually approximate from public sources.

Terms violations. Scraping content behind a login or in breach of a site's terms, or automating against a platform that prohibits it. If a tactic depends on a loophole, it is the wrong tactic — and the resulting data is exactly the kind you cannot cite, publish, or show a client.

One more, and it is a legal matter rather than a taste one: discussing prices directly with competitors, or coordinating on them, is a serious problem in most jurisdictions. Researching published prices is normal competitive behaviour; agreeing on prices with a rival is not. Keep the research one-directional and documentary. If you are ever unsure where a specific tactic sits, that is a question for a lawyer in your jurisdiction, not for a guide.

Where this goes wrong

• Stopping at "contact us." That page is one source of many, and usually the least informative.
• Comparing unlike units. Per seat against per project against per month is not a comparison. Normalise first.
• Treating list price as transaction price. In sales-led markets the real number is lower and lives in reviews and contract registers.
• Letting one anecdote set the benchmark. A single forum figure is a data point. Three consistent ones from independent sources are a finding.
• Using stale numbers. Date every figure. Pricing pages change quietly and often.
• Losing the sources. A number you cannot trace is a number you cannot defend to yourself in three months. Record the URL as you go, not afterwards.
• Pricing purely by comparison. Competitor prices bound the conversation; your costs, your positioning, and what the outcome is worth to the buyer decide the number.

What good looks like

A single table: each real competitor, their entry price and their typical price, both normalised to the same unit and period, what is bundled, the minimum commitment, a source link per figure, a confidence label, and a date. Below it, two sentences on where the market's price clusters sit and which position is empty. That is enough to make a pricing decision, and — because every cell is traceable — enough to revisit it later without starting over.