competitive analysis frameworkcompetitor analysis frameworkporters five forcesperceptual mapframeworks

Competitive analysis frameworks: which one answers your question

26 September 2026 · 9 min read · by the Competite team

A competitive analysis framework is a set of boxes, and the boxes decide what you notice. That is the whole reason picking one matters: SWOT will never tell you whether a market is worth entering, and Porter’s Five Forces will never tell you what to charge on Tuesday. This compares the five you will actually meet by the question each answers, says when each is the wrong tool, and ends with the one most early-stage founders should use instead of any of them.

What is a competitive analysis framework?

It is a fixed set of questions you ask about every competitor, so that two of them can be compared and nothing is forgotten. That is all it is, and the fixedness is both the point and the trap.

A framework makes you notice what its boxes ask about, and quietly stops you noticing everything else. Choose one built for market structure and you will produce a market-structure answer even when your real question was a price. So the useful decision is not which framework is best; it is which question you are actually trying to answer this week.

Which is why the same five companies can produce two contradictory documents in one week. Run Five Forces and you get a page about how hard this market is to enter. Run a pricing matrix and you get a number you can put on a page by Friday. Neither is wrong, and only one of them was the question you had.

How do the five frameworks differ?

By the question each is built to answer, and by very little else. All five arrange the same public facts about the same companies. What separates them is which facts they send you out to find, and which they quietly let you leave out.

When is Porter’s Five Forces worth the afternoon?

When you are deciding whether to be in a market at all. Porter’s own framing asks how much power buyers have, how much suppliers have, how easily a new company can arrive, what could replace the category outright, and how hard the existing firms fight.

Those are structural questions with slow answers, which is exactly why it is misapplied. A founder already six months into building runs Five Forces, produces a page about barriers to entry, and has learned nothing they can act on before Friday. The framework was not wrong; the question was already settled.

Use it once, at the start, or when something structural moves: a platform changes its terms, a supplier consolidates, an incumbent starts giving your category away free.

Why does SWOT produce so little?

Because it asks four open questions with no test for what belongs in an answer. Anything fits, and the thing that fits most easily is an adjective. "Strong brand" goes in a box and nothing follows from it.

Compare that with a row in a pricing table. "Entry plan caps at 8 images, footnote on their pricing page, read 14 September" can be checked, dated and argued with. "Strong brand" can only be agreed with, which is why nobody ever disagrees with a SWOT and nothing ever comes of one.

SWOT is a summary instrument, not a discovery one. It is genuinely good at compressing an analysis for somebody who was not there, which is what a board paper needs. It is poor at producing findings, because none of its four questions tells you where to look. Two rules fix most of it, and they are in the competitor SWOT analysis guide.

What makes a perceptual map useful or useless?

The two axes, and nothing else. A map with the right axes shows you an empty quadrant worth occupying. A map with flattering axes shows you alone in a corner because you picked the two dimensions you happen to win.

The tell is where you end up. If the map puts you alone in the corner you were hoping for, you chose the axes after deciding the answer. A map that is doing its job usually puts you somewhere slightly uncomfortable, next to at least one company you would rather not be next to.

The honest way to pick axes is to take them from sales conversations rather than from a whiteboard: the two things buyers actually weighed out loud when they chose. If nobody has ever mentioned a dimension unprompted, it is not an axis. The longer version, including how to read an empty quadrant three ways, is in competitive landscape analysis.

When should you use a feature and pricing matrix?

Whenever the decision is about a real buyer choosing between you and somebody specific. It is the least fashionable framework on this page and the one that changes the most decisions.

It is also the only one here a buyer could read over your shoulder without wincing, because every cell is either their words or a competitor’s. That is a fair test for any competitive document: if showing it to a customer would embarrass you, it was written to reassure the team rather than to decide anything.

Its rules are simple and almost nobody follows them: rows are questions a buyer actually asked, every cell carries the quote and the date it was read, and at least one row says you are behind. A matrix where you win every row means you chose the wrong company or the wrong rows. There is a table to copy in the competitor analysis template and a filled example in this worked analysis.

Where does jobs to be done fit?

It reframes who the competition is. The question is not which companies sell what you sell, but what people are hiring to get the job done, which is usually a spreadsheet, a freelancer, or carrying on as they are.

That reframing is worth more early than any grid, because at an early stage the status quo wins more deals than every named company put together. What it cannot do is produce a number. It tells you who you are really up against; it does not tell you what to charge. Pair it with a pricing read rather than using it alone, and see direct and indirect competitors for how the kinds differ.

Which framework should an early-stage founder use?

Usually none of them, in the form they are taught. The decision in front of a founder is almost always what to charge or what to say, and the fastest route to both is a short matrix on three to five real competitors with a quote in every cell.

Note the order that implies. SWOT last, not first. It is a way of writing up an analysis, and using it as a way of doing one is the single most common mistake in this category.

How often should a competitive analysis be redone?

The structural parts about once a year. Anything with a number in it about once a quarter, because that is how fast a pricing page moves.

This is the part no framework covers, and it is the part that decides whether the work was worth doing. A framework produces a document with a date on it; four months later the rows about who a company sells to are still right and the rows with prices are quietly wrong, and nobody announced it. Reading the same pages on a schedule is what keeps it true, which is what Competite is for: it watches the pages the numbers came from and tells you the day one moves.

A framework decides what you notice. Pick it from the decision you owe, not from the one you have heard of.