Competitive Analysis

For a long time, competitors entered the room as a single sentence: "I hear they do that too." Plans that took weeks to put together wobbled the moment somebody said it, and nobody in the room could check whether it was true. The material simply wasn't there. Who had looked at whose product, when they looked, what the price was, which kind of customer that company actually sells to — none of it lived anywhere a team could reach.

So decisions followed whatever anyone had heard most recently. A line sales brought back from a customer visit, one slide the founder saw at a conference, an impression an engineer formed trying a rival tool over the weekend. The same company would get described two opposite ways in the same meeting — "they only go after enterprises" and "I've seen them land small teams lately" — and since neither claim had anything behind it, the louder voice won.

Pagedeck lived through this. Sofia came back from a customer and reported that a competitor had quoted half of Pagedeck's price. Within a day there was a proposal to cut prices. When someone finally opened that competitor's actual price sheet, the half price turned out to apply only to contracts of fifty seats or more paid a year in advance. For the ten-person teams that make up most of Pagedeck's customers, it was the more expensive option. Checking took twenty minutes. The missing twenty minutes cost a week.

The information was never hard to get. Pricing pages and feature lists are public, and any competitor with a free plan can be signed up for and clicked through. What was missing was the habit of putting all of it in one place. When each fact lives in one person's head, a thing somebody verified six months ago may as well not exist by the next meeting. The cost of that isn't only the occasional wrong call. Every time the question comes back, either somebody repeats the same twenty minutes of digging, or nobody does and the room argues from memory again.

Competitive analysis is the procedure that turns those impressions into a record. You settle the list of competitors first, decide which items you are going to compare, and write each item down alongside the source you saw it in. Then "I hear they do that too" turns into "that feature is on their top tier only, and it caps saved documents at ten." There is no way to respond to the first sentence. There are several ways to respond to the second.

Its place among product decisions is just as specific. The strategy branch of the product field starts with a long-range goal in one sentence, narrows at product strategy into what the team will and will not do, and narrows again at target market to the customer group worth chasing. The instant you name that group, someone is already sitting in it. Competitive analysis is where you find out who, and when the gap it exposes turns out to match real demand, the result is product-market fit (PMF). Reversing the order breaks the work. Count competitors before the target market is settled and every document tool on earth qualifies, and no decision can be made from a list like that.

The terms filed under competitive analysis split three ways: what you record it on (tools), who you record (subjects), and how you read what you recorded (method).

Tools are the container for the comparison. There is one, the competitive matrix: a table with comparison items down the side and companies across the top. The items are not only features. Entry price, how much the free tier covers, the customer size they mostly sell to, the one sentence they lead with, and anything they shipped in the last six months all belong there, because that combination is what shows where a rival is spending its effort. The table earns its place by exposing blanks rather than by looking tidy. If two of five price cells are empty, the fact that those two companies don't publish pricing is itself a finding. The row for the sentence each company leads with works the same way: when three rivals all lead with speed and none of them mentions the people who have to read the document afterward, that silence is a position nobody is holding.

Fill cells with what you saw, not with what you heard. Price comes from the pricing page, feature scope from screens you signed up for and clicked through, customer size from the case studies they publish themselves. Noting the source next to each cell means you can re-verify it in twenty minutes half a year later, and it means that when someone in a meeting says otherwise you can point at where the number came from. A cell filled in from "someone in sales mentioned it" is unrecoverable within a month — even the person who said it won't remember — and it drags down trust in the whole table.

Subjects is the branch that decides who goes across the top. This is where beginners and experienced teams part ways, because listing only companies in your own category drops everything customers actually use instead of you. Two terms cover that blind spot.

1. Substitute — a different category of thing that solves the same problem for the customer. Pagedeck sells a collaborative document tool, but the way a small team actually keeps meeting notes is a long message in a chat room or one tab of a spreadsheet. Those cost nothing, which is why the strongest rival usually lives in this row. Leave it out and there is no explanation for why free users never move to a paid plan.
2. Indirect competitor — a company selling something else that takes the same customer's budget and the same customer's hours. If a team adopts exactly one new tool this year, a project tracker and a document tool compete for that single slot even though their features never overlap. Asking customers who chose against you what they bought instead fills this row on its own.

Method is the branch that decides what to do with the filled table. Benchmarking belongs here: measuring your own level against a leading example. It is worth noticing that it runs in the opposite direction from competitive analysis. Analysis reads the table looking for what nobody is doing. Benchmarking takes what rivals do well, treats it as a baseline, and measures how far behind you are. If signup takes two steps at a competitor and six at your company, that is not a gap to move into — it's a place you are behind, and measuring that distance is benchmarking's job. Both readings come off the same table, which is why teams that keep one usually run both passes in the same sitting.

The three branches run in sequence. Subjects decides who you look at, tools records what you find, method reads the result. Drop any one and the rest spin in place. Build the table before settling the subjects and you end up with the three companies everyone already knows. Benchmark without a table and you measure against a different company each time, so you never learn whether you improved.

"Let's run the competitive analysis again. That table is from March, so their pricing has probably moved."

The record is old enough that it can no longer support a decision today.

"This isn't a competitive analysis, it's a feature list. I can't see what we're choosing not to do."

The work stopped at copying down rival features and never reached a judgment.

"Add substitutes too. If customers don't use us, they just type it into a chat room."

The real rival sits outside the product category, and it's free.

At Pagedeck, Mia opens this work two weeks before quarterly planning. Sofia collects the deals lost over the last three months along with whatever the customer said they picked instead, and Ravi fills in each company's pricing page and recent release news. Grace doesn't read the whole table. She asks one line: "Which cell in here can only we fill?"

The slowest part of the exercise is never filling cells — it's deciding who belongs in the table at all. Eight candidates come up at first, and cutting the ones whose customer size doesn't overlap with Pagedeck's leaves four. The test for keeping a company is not how well known it is but a single question: have we lost a deal because of them? Before that test existed, the biggest name in the market always took the first column, while the deals Pagedeck actually lost went to two far less famous companies.

When Grace's question gets an answer, that cell goes to the top of next quarter's product roadmap. When it doesn't, the team looks at narrowing who it sells to rather than building more. Last quarter the answer was that every company edits documents about equally well, but nobody connects meeting notes to the tasks that come out of them — and that one empty cell became the first line of the quarterly plan. The table never makes the decision. It puts something decidable in front of people.

The most common mistake is treating the output as a list of features to catch up on. The cells rivals have filled and you haven't are right there in the table, so filling them looks like the obvious next job. But those cells reflect what a rival chose after looking at its own customers, not what your customers asked for. Work through the empty cells in order and two quarters later your product looks nearly identical to theirs, and once products look identical the only thing left for a customer to compare is price. The first thing to look for in the table is not the cell you're missing but the cell nobody has filled. Defending the cells only you have comes out of the same table: if one row is yours alone among five companies, a quarter spent widening that lead usually returns more than a quarter spent closing four gaps.

The second mistake is treating the document as something you build once and keep. It takes days to produce, so there's a pull toward getting years of use out of it, and the file sitting in a folder makes it look current. The contents go stale every time a rival changes something — pricing shifts, the free tier shrinks, the core customer moves upmarket. Feed a six-month-old table into a framework like SWOT analysis and the strategy that comes out attacks a weakness the competitor already fixed. Teams that avoid this re-verify at least the pricing and plan rows each quarter and write the verification date beside the cell. Rebuilding the whole table is unnecessary. Correcting one cell whenever a rival ships a feature or moves a price means that on planning day the current table is already waiting, with no multi-day scramble in front of it.