Prompts for Competitive SWOT Analysis

Strengths and weaknesses are statements about a company: what it does well, where it is thin, what it charges, what it has shipped. A competitor’s own pages are fair evidence for all of that, and a careful reader can check every entry.

Opportunities and threats are statements about a market acting on that company. Regulation, a shift in buying behaviour, a new entrant, a category consolidating. Nothing on a vendor’s website is evidence for any of it, because a vendor does not publish the forces acting on them.

The format is a promise of completeness

Every published SWOT example has four full boxes with three or four bullets each. That is what the format looks like, so that is what gets produced. The pressure comes from the shape, not from any particular tool. A person filling in a SWOT template by hand reaches for the same filler in the same box for the same reason. The top half is fed by the same evidence a positioning teardown and a feature comparison already produce, which is why those two are worth running first.

The detail that matters is the second input, and specifically that it sits in a separate folder. Keep market evidence out of the first pass and you can see how much of the grid your competitor evidence actually supports. Where that second folder is thin, a competitive market read is the job that fills it, and it answers to different sources.

The first builds it. The second is worth running on SWOTs you already have, including ones written by people, and it usually cuts more than you expect. The third handles the bottom half separately, which is the only honest way to produce it. The fourth turns four lists into a decision.

The instruction that changes the output most is the permission to leave a quadrant thin. It sounds like a small caveat. It is the difference between a document that reports what you know and one that reports what the shape demanded.

We asked for a SWOT of a real product with nothing supplied, and then asked for the same thing with one saved page and permission to leave boxes empty.

Read the threats in the first run again, and put any other vendor in that category into them. Every bullet still reads correctly. Put your own company in and they still read correctly. That is the whole problem in one move. Those four lines are not findings about a competitor. They are things that are true of software categories, and they exist because two boxes were still empty.

Strengths survive, threats do not

Worth saying plainly, because the interesting result is rarely the one you set out to find. Two of the weaknesses the model produced from memory could be checked against the live pricing page, and they survived the check. Recall about a widely documented product is decent. The failure is not that the model does not know things. It is that the format asked for eight bullets, and only four of them could have had a source. That is a property of the grid rather than of the model, which is why SWOT analysis is taught as a prompt for gathering rather than as a deliverable.

The instinct when a box comes back thin is to go and find something for it. That instinct produces most of the filler in most SWOT documents, and it destroys the most useful output the exercise has.

An empty threats quadrant does not mean there are no threats. It means you hold no evidence about them. That is a different statement, and a far more useful one, because it names the gathering you have not done. A competitive landscape map cannot record that gap at all, which is worth knowing before you reach for one: a grid holds an empty cell and a scatter plot has no way to. Fill the same box in from general knowledge and it now claims you have thought about the outside world. Nobody downstream will ask again.

Two sourced entries beat eight plausible ones

A grid with two strengths, two weaknesses and two empty boxes looks unfinished. It is also an honest account of what a day of reading established. And it survives the one thing that kills competitive documents: somebody senior asking where a particular line came from. The version with eight bullets does not, and it usually fails in public.

The other habit worth keeping is the closing list from the fourth prompt: the entries that are true, properly sourced, and change nothing. Naming them stops the grid being read as a to-do list. It is also a short list of exactly the things not worth a meeting. The weaknesses half has one source a competitor does not control: competitor reviews, where the complaints are dated and written by people with no reason to be kind.

The half of the grid with the best evidence is also the half that expires fastest, and it expires in one direction. Weaknesses get fixed. A competitor reads their own reviews, sees the complaint you built a differentiator on, and ships something about it.

Nothing in a SWOT document marks that it has happened. A weakness noted in March sits in the same box in October, looking exactly as it did. The first sign it closed is usually a rep being corrected by a buyer who has just read the release notes. That is the most expensive way to find out.

The bottom half of the grid stays manual, and it should. Deciding which market forces act on a competitor differently from everyone else is analysis rather than observation. That is why the swap test exists, and why no feed will ever run it for you.

Flares watches product changes, pricing and positioning across the companies you track, and dates each one. A weakness that stops being true then reaches you as a dated change, rather than as a correction in a live call. What a fixed weakness means for your position is still your judgement.

Kept that way, a SWOT stops being the document somebody rewrites from scratch every year. It becomes one where two cells changed and the rest still stands. That is the only version anybody keeps opening. What leaves the grid goes into a competitor battlecard, and only the sourced cells should survive the trip. The same is true of the wider competitive analysis it usually sits inside.