1. Define the category by substitution
Not by industry code, and not by who sells a similar product. By what the buyer does instead of buying from you.
That definition usually includes something uncomfortable: doing nothing, an internal spreadsheet, a freelancer, a completely different product category that solves the same evening. Leave those out and the map will be tidy and wrong, because the biggest competitor in most categories is inertia and it never appears on a vendor list.
2. Find the set instead of listing it from memory
The names you can recall are the names you already think about. Five public sources give you the rest, and none of them costs anything.
Their own comparison pages. Companies publish exactly who they consider a threat, and they are rarely wrong about it.
Search results for the category term, including the comparison and alternatives queries buyers actually use.
Answer engines. Ask several models who the main options are. The answers are a real signal even when they are wrong, because they are what your buyers are now being told.
Review-site categories and directories, which enumerate the long tail nobody remembers.
Ad libraries, where anyone bidding on your category has to show themselves publicly.
Stop at fifteen names. A map with forty dots is a database.
3. Choose axes you can measure, before you plot anybody
Two by two grids get drawn after the fact, with axes selected because they place the author in the top right. That is not a map, it is a self-portrait, and everybody in the room can tell.
Pick axes that are readable from public data and settle them before you look: entry price against range width, self-serve against sales-assisted, breadth against depth, a specific segment against everyone. If you cannot fill an axis from something public, the axis is a belief. Beliefs belong in the write-up, not on the chart.
This is the same rule as choosing benchmark measures before you know who wins them, and for the same reason: after the fact, any set of numbers can be made to say what was already decided.
4. Fill it from what is public
Prices and the shape of the ladder from their own store or pricing page. Range and depth from their catalogue. Positioning from the homepage headline and the words it has been through. Motion from publishing cadence, live ads and open roles.
Two rules keep it usable. Record the source next to each cell, so a number can be checked six weeks later without redoing the work. And keep measured facts visually distinct from anything modelled: a price is a fact, a traffic estimate is not, and a map that mixes them silently is a map you cannot defend in the meeting where it matters.
5. Date it, then rebuild it
A market map is accurate on the day it is drawn and decays quietly. Tiers appear, positioning drifts, a new entrant shows up in the answer engines before anybody in your company has heard the name.
Put the date on the artefact. Then make it cheap to redo, because the version that gets rebuilt every quarter is worth more than the beautiful one that gets rebuilt never. That is also the difference between this and the report you were considering buying: not the quality of the first draft, but whether there is a second one.
What the map is for
Two things, and neither is decoration. It shows where the empty space is, which is the cheapest ground on the board because nobody is defending it. And it shows who you are actually being compared against, which is often not who you assumed, and that changes what your own pages need to say.
If the map makes you want to move to where the crowd is, read the note on copying first. The interesting part of any map is the part with nothing in it.
Pick one competitor. See what we find.
Your first benchmark is free. One domain, measured against you, no card and no call.