UX competitive analysis: a practical guide

What UX competitive analysis is

UX competitive analysis is an expert review of how competing products handle the same user jobs as yours. You walk their public flows, from the first headline to the point where signup or payment blocks the way, and you judge them against a consistent set of criteria. The output is a set of annotated screenshots and a short list of design decisions, not a spreadsheet of scores nobody reads.

It answers a question every design team hits: "how do the others do this?". Before redesigning a pricing page, a signup form or a navigation system, it is worth an afternoon to see how five comparable products solved the same problem, what patterns repeat, and where everyone leaves the same gap.

The value is not in copying rivals. It is in knowing which patterns users already understand, and where nobody has done the obvious thing well.

What it is not

UX competitive analysis is often confused with usability testing, and the two make different claims. Usability testing puts real people in front of an interface and watches where they struggle. Competitive analysis is your own expert read of a public interface. It cannot tell you where users actually give up on a rival's site, because you cannot observe their users. It can tell you what the interface asks of a visitor, and where that ask looks heavy.

• Not usability testing. No participants, no sessions, no task completion rates. Your judgement, applied consistently, is the instrument.
• Not analytics. You cannot see a competitor's funnel numbers, and honest analysis never pretends to. It reads intent from what is on the page.
• Not a feature audit. Listing what rivals have is inventory. This method asks how well each flow works for a first-time visitor, which is a different question.

If the wider strategy questions interest you (pricing structure, positioning, proof), the complete guide to competitor teardowns covers the full-site version of this exercise. This guide stays on the UX layer.

Choose three to five comparators

More than five comparators and the analysis takes days and the findings blur. Fewer than three and one odd site skews everything. The useful mix is direct rivals plus one outsider:

• Two or three direct rivals. Products your users actually compare you against. These show the patterns your market already expects.
• One adjacent product. Same audience, different job. It reveals conventions your users bring with them from elsewhere in their day.
• One best-in-class outsider. A product famous for the flow you are studying, even if it is in a different category entirely. Studying a well-known checkout when you are redesigning signup raises the bar beyond what your direct rivals settle for.

Write down why each comparator made the list. When you present the findings, "chosen because their onboarding is the one prospects mention" survives questioning. "They seemed relevant" does not.

What to evaluate

Walk the same path on every site so the comparison holds: landing page, primary navigation, the main product or pricing page, and the start of signup, stopping before any real form submission. On each stop, look at five things.

• The onboarding promise. What does the site claim a new user will get, and how fast? Compare the headline's promise to what the signup flow actually asks for first.
• Navigation. How many top-level items, what they are called, and whether a first-time visitor could predict what sits behind each label.
• Forms. Count the fields on every form in the path. Note what is optional, what is demanded up front, and whether a card is required before value.
• Mobile parity. Load every page at phone width too. Cut sections, shortened headlines and moved calls to action are findings, not accidents.
• Accessibility signals visible from outside. Text contrast, text size, whether the page works at 200 percent zoom, whether images of text replace real text, whether focus states exist when you tab through. You cannot audit a rival's code, but these surface signals are visible to any visitor.

A working heuristics table

Score each comparator against the same short list. A three-point scale (clear, mixed, weak) is enough; finer grades invite arguments about the scale instead of the finding.

Capture the evidence

Every judgement in the table needs a screenshot behind it, because rival sites change without notice and because a claim with a picture gets believed. Doing this by hand means full-page captures of four or five pages per comparator, at desktop and phone width, with cookie banners cropped out. Across five comparators that is forty or more captures before the thinking starts. The manual methods are covered in how to take a full-page screenshot.

Stravue's version of the step: paste each comparator's address and what you're after ("onboarding", "pricing"), pick from the pages it proposes, and it captures them full-page on desktop and phone with cookie banners and chat widgets removed. Pin numbered notes straight on the captures, then export the set as PowerPoint, PDF or a share link. The judging stays yours; only the tab work goes. A worked single-site example shows the output.

Turn findings into design decisions

1. 1

   Sort findings into patterns and gaps

   A pattern is something most comparators do the same way, which users likely expect. A gap is something nobody does well. Patterns argue for convention; gaps argue for opportunity.

2. 2

   Attach each finding to a screenshot

   Every claim in the write-up points at a pinned capture. "Three of five ask for a card before showing the product" lands because the three screenshots sit right there.

3. 3

   Write decisions, not observations

   "Their forms are shorter" is an observation. "Cut signup to three fields and move company size to after first login" is a decision someone can act on this sprint.

4. 4

   Present it as a short deck

   One capture per slide, the pin numbers on the image, one takeaway line underneath. Structure and slide order are covered in how to present competitor research.

An analysis that ends in three decisions changes a product. One that ends in a scored spreadsheet changes a folder.

Frequently asked questions