Prompts to Map a Competitive Landscape

Evidence about a competitor arrives with things attached to it. A price has a currency. A quote has a page and a date. A judgement has a confidence. A stated target market is often a range rather than a value.

A grid can express one thing. Where the dot sits. Everything else is dropped at the moment of placing, and the drop is silent, because a dot with no evidence behind it looks exactly like a dot with a quoted page behind it.

This is not a complaint about charts. It is the reason the working document and the picture have to be two different objects. Keep the evidence, the dates and the confidence in a table beside the map, and treat the map as the summary rather than as the analysis. A landscape template is where that table lives, one row per player, with a source column that a grid has nowhere to put.

Ask for a landscape without naming the axes and you will get axes. Ours came back as ease of use against depth of functionality, with a suggestion of small business against enterprise as an alternative.

Both are the category’s own marketing language. Every CRM on the market claims to be easy to use; it is close to the most common sentence in the category. An axis built on a claim that every company makes cannot separate any two of them using anything they publish, so the placements have to come from somewhere else. What they come from is how each company is generally regarded.

The two axes also decide what kind of object you have made. Price against capability gives a product comparison; buyer against buyer gives market segmentation with dots on it. Both are called a perceptual map, and only one of them will answer the question you started with.

An axis is a claim, not a formatting choice

Naming an axis asserts that this variable explains something about this market. That is a real claim and it can be wrong. It is also the only part of the exercise that cannot be delegated, because it depends on the decision you are trying to take.

Which gives a short test, and it is worth running before the drawing rather than after. Take each company and ask what you would do differently if it sat one quadrant over. If the answer is nothing, that axis is a label rather than a variable. Axes drawn from what each company publishes about itself are the cheapest to defend, and competitor positioning is where those sentences live.

Two inputs, and the second one is usually treated as a style choice. It is the analysis. The first input is where most of the work sits, and the work is capture rather than thinking.

Capture numbers exactly as the page printed them, currency symbol included. That instruction sounds fussy until a set spans several countries, and then it is the difference between a price axis and a fiction. More on why below, because we hit it on the first try.

Three prompts, and the middle one runs first. Pulling the segment statements out before choosing an axis is what tells you whether the axis you had in mind exists in anybody’s published words.

The third prompt has to be a separate message on a finished map. Appended to the request that draws the map, an instruction to challenge the axes produces a paragraph of hedging under a confident grid. Sent on its own, against a map that already exists, it produces a list of quadrant moves that change nothing, which is the finding.

We asked for a CRM landscape with no evidence at all. Then we ran it again with real material: six vendors, and the entry paid price exactly as each vendor’s own page served it on 24 September 2026. The same material carried every statement those pages make about who the product is for.

Two of six prices were never published in the axis currency

Six pricing pages, one connection, one afternoon. Three currencies came back. Attio quoted euros, four vendors quoted US dollars, and Zoho quoted Indian rupees on a page whose text was in French. Salesforce was the only one of the six offering a currency selector.

To put six vendors on one price axis, somebody has to convert. The converted figure then sits on the chart as the price, at a rate with no date and no source, and it appears on none of the vendors’ pages. A reader who checks Zoho’s page for the dollar figure on your map will not find it.

The country a page was served to is the field nobody records. Capturing it matters for any comparison, and it matters more here, because a competitor messaging comparison at least shows a reader the language it was reading. A dot does not.

This was the result we did not expect, and it is the strongest argument on the page. The second axis of almost every landscape is who the product is for. We pulled every segment statement the six vendors publish. Not one of them states a single segment.

So the axis does not exist in anybody’s published words. To draw the map, each range has to be collapsed into a point, six times, by whoever is holding the pen. Every collapse is a judgement. On the grid every one is a dot.

The information was in the two that broke the pattern

Four of the six landed in one corner, which is the usual outcome and is structural rather than bad luck. Entry price and target segment are not independent variables. A product priced at nine dollars a seat is priced that way because of who it is for, so plotting both plots one thing twice, and the result is a diagonal.

The two that broke the diagonal were the two worth talking about. Salesforce publishes a twenty-five dollar entry price and declines the segment question, which means a low entry price is being used as a door rather than as a statement about who it serves. Attio publishes the highest entry price in the set and describes itself as being for every stage of growth. Both are visible in one sentence of prose. On the map they were dots in a corner with four others.

Scale is the axis people reach for next, and it is the one with a trap in it. Relative market share divides your revenue by the largest competitor’s, so the series breaks the moment a different company becomes the largest. A market share analysis keeps the working, which a dot cannot.

A landscape ages one dot at a time, and it never announces it. A price moves, a vendor adds a segment page, a company rewrites the line describing who it is for. The map on the slide does not change, because a map has no mechanism for changing.

That is why the evidence table beside it matters more than the picture. A placement whose evidence carries a date can be re-checked; a placement that only exists as a dot cannot be questioned, only redrawn from memory. Redrawing a map quarterly from the same year-old pages is a refresh that updates nothing.

Some of this is not a tooling problem. Choosing two axes that separate a set, and knowing which quadrant boundary would change a decision, is judgement, and nobody should want it automated. What can be kept current is the evidence under each dot: the prices, the segment statements, the positioning lines, and the dates each one was last true.

Flares follows competitor pricing, packaging and positioning and dates every change it finds. A landscape then gets redrawn because a placement moved, rather than because a quarter ended, and the evidence table underneath it stays something a sceptical reader can check line by line. Where the same set has to be reported on rather than drawn, a competitive intelligence report carries the dates a grid cannot.