uxmetrics.com

UX Metrics is an all-in-one UX research and information architecture (IA) testing platform. It bundles card sorts, tree tests, surveys, usability tests, prototype tests, and interviews into a single workflow, and it lets you run studies with your own participants or recruit them through the platform. If you need to validate navigation, labels, or task flows without stitching together several tools, it's built for that. A free tier covers up to 3 participants per study, which is enough to pilot a method before committing further.

What you can do with it

The platform's core claim is that every major research method lives in one place, so you don't move data between tools or rebuild a study when you switch methods.

Because these share one workflow, a common pattern is to run a tree test to find a structural problem, then follow up with a card sort to see how users would reorganize the same content.

How participants work

You have two paths:

• Bring your own participants. Share a link or embed the study, and run it with people you already have access to.
• Recruit through the platform. The site describes recruiting "your way, at any stage," so recruitment is positioned as flexible rather than a fixed panel.

For a first project, bringing your own participants is the lower-friction option: you control who takes part and can start immediately.

Where the AI fits

UX Metrics advertises AI "powered by your full research context." In practice, this means analysis assistance that draws on the study you actually ran rather than generic suggestions. Treat it as a way to speed up synthesis of open-ended responses and patterns — not as a replacement for reading your own data, especially on small samples.

When it's a good fit

Choose UX Metrics if:

• You run more than one research method and want them in one place.
• IA work (card sorts, tree tests) is central to what you do.
• You want to start small — the free tier's 3 participants per study lets you test the workflow before paying.
• You have your own participant pool and mainly need the tooling.

Look elsewhere if you only ever run one method and already have a tool you like, or if you need a large recruited panel with guaranteed demographics — check the pricing page for what recruitment actually includes at each tier.

Getting started

1. Pick one method that matches a real question you have — for example, a tree test on a navigation menu users keep missing.
2. Create the study and add your tasks or content.
3. Run it with up to 3 participants on the free tier to confirm the study works as intended.
4. Review the analysis, then scale participants or add a follow-up method once the design is sound.

The main pitfall at this stage is writing tasks that are too vague to interpret; a small pilot run surfaces that before you spend on a larger sample.

Yes. UX Metrics is built so you can run studies with your own participants rather than being locked into a built-in panel. The site states you can "recruit your way, at any stage," and the platform supports card sorts, tree tests, surveys, usability tests, and more using your own participants. The free tier covers up to 3 participants per study, so you can validate the workflow before committing to a larger sample.

What "use your own participants" means in practice

When a tool says you can bring your own participants, it usually means you handle recruitment and distribution yourself, then send people into the study. With UX Metrics, the relevant signals are:

• Recruit your way — the platform does not force you through a single panel provider.
• At any stage — you can run studies with your own audience whether you are early in discovery or validating a near-final design.
• Multiple methods in one workflow — card sorts, tree tests, surveys, prototypes, interviews, and more, so the same participant group can be used across different study types.

This matters if you already have access to a user base, a customer list, a community, or a pool of testers and want to avoid paying panel fees or waiting on external recruitment.

Participant limits and the free tier

The site's description states the platform is free for up to 3 participants per study. That is the key constraint to plan around:

If your goal is a quick internal pilot or a sanity check with a couple of colleagues, the free tier is enough. If you need statistically meaningful results from a larger group, check the pricing page for the plan that fits your sample size.

How to run a study with your own participants

1. Choose your method. Decide whether you need a card sort, tree test, survey, usability test, prototype test, or interview.
2. Build the study in UX Metrics using the all-in-one workflow.
3. Recruit your participants through your own channels — email, community, customer list, or internal testers.
4. Share the study link or embed with those participants. The site mentions links, embeds, live URLs, and shareable analysis as part of the workflow.
5. Collect and analyze results using the platform's shareable analysis features.

The expected result is that responses from your own participants flow into the same analysis view, so you can review findings without exporting to a separate tool.

Common sticking points

• Free tier cap. Three participants per study is fine for a pilot but not for quantitative confidence. Plan your sample size before you start.
• Recruitment is on you. Because you bring your own participants, you are responsible for sourcing, scheduling, and incentivizing them. The platform does not appear to supply a panel.
• Method fit. Not every method suits every participant pool. A tree test works well with remote, unmoderated participants; interviews may need scheduling.
• Sharing and analysis. If you need to hand results to stakeholders, confirm how the shareable analysis and links work for your plan.

When this setup fits

Use UX Metrics with your own participants if you already have a user group, want to avoid panel costs, and need multiple research methods in one place. If you need a large, pre-recruited sample or advanced statistical testing, verify the paid plan details on the pricing page before committing.

UX Metrics supports card sorts, tree tests, surveys, usability tests, prototypes, and interviews — all inside a single workflow, and all runnable with your own participants. That combination is the main reason to consider it: if you need to test information architecture and then follow up with a survey or a prototype task without moving your participants to a different tool, this platform is built for that. It is free for up to 3 participants per study, so you can validate the workflow before committing.

Methods by research goal

The platform groups its methods around what you are trying to learn, not around a fixed menu order. The table below maps each method to the question it answers.

Card sorts and tree tests for information architecture

These two are the IA-specific pair, and they answer different questions. A card sort is generative: you give participants a set of items and they group and name them, which tells you how users mentally organize your content. A tree test is evaluative: you give participants your existing hierarchy and a task, and you see whether they can find the right destination.

The practical sequence is card sort first to propose a structure, tree test second to check that the structure works. Running both in one platform matters because the labels and categories from the card sort feed directly into the tree you test.

Surveys and usability tests

Surveys let you collect structured responses at a scale a single moderated session cannot reach. Usability tests give you task-level evidence: whether someone completes a task, how long it takes, and where they get stuck.

Use a survey when you need breadth on attitudes or preferences, and a usability test when you need to watch behavior on a specific flow. If you already have a usability finding, a short survey can tell you how widespread the problem is.

Prototypes and interviews

Prototype testing lets you put a design in front of people before development, so you catch structural problems while changes are still cheap. Interviews add the "why" — the reasoning behind a choice that a task completion metric will not explain.

A common pattern: run a prototype test to find where people fail, then interview a few of those participants to understand what they expected instead.

Why "one workflow" is the actual feature

The methods are individually common. What the platform emphasizes is that they share one workflow, one participant pool, and one place for analysis — the site describes links, embeds, live URLs, and shareable analysis as part of the setup. In practice this means:

• Recruit once, reuse participants. You can run a card sort, then invite the same people into a tree test or survey without rebuilding a panel.
• Keep context together. Findings from a tree test sit next to the survey that followed it, which makes the "what happened and why" story easier to assemble.
• Share results as links or embeds. Stakeholders can view analysis without a walkthrough from you.

The site also states that AI features are powered by your full research context, which is only meaningful if your methods actually live in one place — another argument for consolidating rather than stitching tools together.

Choosing between methods

If you are unsure where to start, work backward from the decision you need to make:

• Restructuring navigation or categories → card sort, then tree test.
• Checking whether a specific task is completable → usability test.
• Measuring how common an attitude or preference is → survey.
• Validating a design before build → prototype test.
• Understanding motivation behind observed behavior → interview.

You do not need every method. Pick the one that matches your decision, and add a second only when the first leaves a question it cannot answer.

Getting started and what to check

The platform is free for up to 3 participants per study, which is enough to run a pilot of any method above and confirm the setup works for your team. Before committing to a paid tier, check the pricing page for participant limits and feature differences, since the free allowance is per study rather than unlimited.

Two practical checks worth doing during a pilot:

1. Run one card sort and one tree test on the same content. If the labels transfer cleanly between them, the single-workflow claim holds for your use case.
2. Test the sharing path. Send a shareable analysis link to a colleague and confirm they can read it without your account.

If both pass, the platform fits teams that want IA testing and broader UX research in one place. If you only ever run one method, a single-purpose tool may be simpler — the value here comes from combining methods.

UX Metrics applies AI on top of your own research context rather than as a standalone tool. According to the site, its AI is "powered by your full research context," and it sits inside the same platform that runs card sorts, tree tests, surveys, prototypes, and interviews. In practice, that means AI is positioned as a layer over the data you collect in a study, not a separate product you configure on its own.

What "full research context" means here

The phrase matters because it describes the input the AI works from. UX Metrics keeps every method in one workflow, so a single study can include multiple activities — for example, a card sort followed by a survey, or a tree test alongside a prototype task. Because those activities live in the same platform, the AI can draw on the combined study material instead of only one isolated result set.

The site does not publish a detailed breakdown of exactly which data points feed the AI, so treat "full research context" as a description of scope rather than a technical specification. What is supported by the page evidence is the claim itself: AI powered by your full research context, within an all-in-one workflow.

Where AI fits in the research process

Based on how the platform is described, AI is tied to the analysis and insight side of research rather than to study setup or participant recruitment. The relevant signals:

• Analysis and insight generation — the AI is presented as part of how results are turned into findings, alongside shareable analysis.
• Across methods — it is not limited to one research method; the platform lists card sorts, tree tests, surveys, prototypes, interviews, and more in a single workflow.
• Recruitment stays separate — the site describes recruiting "your way, at any stage," which is a participant-sourcing capability, not an AI function.

So if your goal is faster interpretation of study data, AI is relevant. If your goal is finding participants, that is handled by recruitment features instead.

A concrete example

Suppose you run a tree test to check whether users can find a pricing page, then follow it with a short survey asking how confident they felt. Both activities sit in the same study. When you move to analysis, the AI has the tree test paths and the survey responses together, so it can reason about the study as a whole rather than treating each method as a disconnected export. That is the practical difference the "full research context" framing is pointing at.

What to check before relying on it

• Confirm which methods your study uses. The AI's usefulness depends on having enough collected data in the platform to work from.
• Check the pricing page for plan limits. The site links to a pricing page, but the provided material does not state which AI features are included at which tier.
• Note the free tier boundary. The platform is described as free for up to 3 participants per study, which is enough to test the workflow but not to run a full-scale study.

If you want to see how the AI behaves with your own data, the most direct route is to run a small study within the free participant limit and review the analysis output before committing to a larger one.

UX Metrics is built so that research outputs can leave the researcher's account and reach teammates, stakeholders, or clients. Based on the platform's own description, results are distributed through links, embeds, live URLs, and shareable analysis, and every method — card sorts, tree tests, surveys, prototypes, interviews — runs inside one workflow rather than in separate tools. That matters if your main bottleneck is not collecting data but getting other people to look at it and act on it.

What "shareable analysis" means in practice

The site describes its output as shareable analysis rather than a static report export. The practical implication is that you are distributing a view of the results, not a frozen document. Two consequences follow:

• Updates travel with the link. If you keep collecting responses, anyone holding the link sees the current state rather than the state at export time. This is useful for ongoing studies; it is a risk if you need a fixed snapshot for a formal sign-off.
• The audience needs no account of their own to view what you send — the distribution methods listed are links and embeds, which are view-oriented. Note that the site's free tier is defined by participant count, not by viewer count, so sharing with a large internal audience is not the same constraint as running a large study.

The four distribution routes

The distinction that trips people up most is the third row. Live URL testing means participants interact with an actual address, so anything gated behind authentication, a staging password, or a VPN will not behave the way it does for a normal visitor. Decide early whether you are testing the real experience or a prototype of it, because that choice determines whether your results reflect production conditions.

A concrete workflow

Say you have run a tree test on a redesigned navigation and need sign-off from a product manager and two designers.

1. Confirm the study is complete enough to interpret. Partial results are still viewable through a shared link, so decide whether you want reviewers seeing a half-filled sample.
2. Generate the shareable analysis view for that study.
3. Distribute by the route that matches the audience — a link in the project channel for a quick look, an embed in the design doc if the team will return to it repeatedly.
4. State the conditions alongside the link: sample size, participant source, and whether the study is still open. A shareable view does not carry that context on its own.
5. Collect feedback in one place. Because results are viewable rather than exported, comments tend to land in chat or the doc — pick one and say so, or you will be reconciling three threads.

Where this approach fits and where it does not

Choose the link-and-embed model if your stakeholders are internal, your studies are iterative, and you want people checking results repeatedly without asking you for a fresh export.

Prefer a fixed export if you need an immutable record — a client deliverable, a compliance artifact, or a results appendix attached to a dated decision. A live view that changes after the fact is the wrong object for that purpose.

Watch the participant limit, not the viewer limit. The free tier covers up to 3 participants per study. That is a constraint on how much data you can collect, not on how many people can look at it — so a small pilot study can still be shared widely.

Questions the site does not answer

The available material does not specify link permissions, expiry, password protection, or whether embeds are live or static. If any of those determine whether you can use this with an external client, verify them directly before committing a study to the platform. Likewise, the pricing page is linked but no figures are given here, so treat cost as something to confirm rather than assume.