Hallway Tests — Quick Feedback on Your Prototype

What is a “hallway test”

A hallway usability test or a corridor test — is a quick UX-testing of a developing product/service.

• You can test interfaces, functionality, features, landing pages, CJM storyboards, screen flows, articles, emails, and advertisements.
• It takes 5–10 minutes.
• It’s called a hallway test because you only need to step into the hallway of your office and ask 5–10 people, who are not involved in the product’s creation, to participate.
• The criterion for sufficient testing is when the information starts repeating.
• It can be conducted iteratively in batches of 3–5 tests with subsequent improvements.
• It can be done physically nearby or online.

Advantages of the Method

• Fast — 5–10 minutes per test, 3–4 tests in a series.
• Simple — does not require complex coordination.
• Cheap — does not require contractors or funds.

When to Conduct a Hallway Test

Suppose you have a sketch, draft, mockup, prototype, screenflow, MVP, or another preliminary version of the product. Usually, you might assume that the user:

• reads the texts as intended,
• understands the statements,
• pays attention to the highlights,
• uses the controls,
• moves through the scenario path,
• and performs the desired actions.

Spoiler: Actually, not at all!

Most likely, you are mistaken. These are just hypotheses, and you need to check how it happens in reality. You will be surprised ;)

Why Conduct Hallway Tests

User perception and usage

Testing is needed to first find out how users perceive and use what you are going to create or have already created. The options are:

• Exactly as you expected — magic! Moreover, if this is the first version, then it’s suspicious.
• Not quite as intended —try to identify as many rough edges, obstacles, bottlenecks as you can, then fix them and conduct another series of tests.
• Not at all as intended — it’s great that you found this out early, instead of wasting time and money on developing it. Create another version and test it.

Save time and money on gross mistakes

Secondly, to identify unsuccessful options as early as possible and not waste time and money on their creation.

Not A/B tests

Hallway tests are NOT conducted to choose one of the options — A/B tests exist for this, but their implementation is more complex than it might seem, especially in terms of isolating the experiment and interpreting statistically significant results.

Who can be a Respondent for Testing

1. Team members are usually not suitable because they are biased.
2. Sometimes neighboring team members are enough, but they might also be aware of your product, and their opinions could be even harmful. Moreover, after 10–20 dialogues, new people will be needed.
3. Fresh new users who have never seen your product are often necessary.
4. Sometimes people from a specific industry and with certain experience are needed, such as procurement specialists, supplier selectors, or trainers.
5. Conversely, sometimes people not from the industry are needed.

Look for respondents as you would for problem interviews.

How to Conduct a Hallway Test

1. Prepare the object of study

This can be an application interface, specific functionality, a feature, a landing page, a storyboard or screenflow, an article, an advertisement, or other material reflecting possible user behavior scenarios like CJM.

2. Open all links in advance,

so you don’t have to search and wait for loading during the test.

3. Create and tell a 1-minute legend

to set the context for the user:

• Who they are,
• What their task is,
• How they are trying to solve it, what actions they have already taken,
• How they got to the test object.

Tell the legend to the respondent, for example:

You are a training manager at a large bank. You need to choose an external contractor to conduct a training on Agile & Scrum for 10 product teams. You entered the query in the search and go to the first 5 links. A landing page opens.

4. Explain the Testing Approach

Below is an example, but explain it in your own words:

Voice all your thoughts and actions out loud.

Do not try hard to figure things out. Turn off your ingenuity. If it’s unclear, it’s the product’s fault. So feel free to comment on everything that is unclear. Imagine you came home tired, you are seeing this in the evening and don’t want to think, but more want to grumble and criticize.

I understand that “this is your personal perception”. And that’s just great! It’s your perception that interests me.

If your gaze wanders inconsistently — that’s normal. Just comment aloud what you looked at, what you noticed, and how you understood it.

5. Constantly Ask 3 Questions

1. What do you see? They may tell you: “I entered, I see the headline, subheadline…” To understand what catches their attention and what remains unnoticed.
2. What do you understand? They may reply: “What does this mean? I didn’t understand anything.” How the user perceived the information and what remained unclear is important to you.
3. What do you want to do/what are you doing? They may say, for example: “I scroll further. So, this I skip — not interesting. Aha, I read this… Hmm, this is interesting. What’s next? I click here… Oh, this is not a link, but just underlined — unexpected.”

Ask these same 3 questions on each new block/screen, listen, and take notes. If the user is silent, continue to ask these questions so they verbalize all their thoughts and actions.

6. Record the respondent’s actions and emotions

It’s important not only what they do but also what they do NOT do, although you expect and want that behavior.

Record the respondent’s emotions and where/when they arise. If one block causes excitement, it’s more likely the person will scroll further and explore the next ones. If the user starts to get confused, feels discomfort, increasingly doesn’t understand what, where, and why — they are likely to leave the page soon.

7. Conduct the test with 3–4 respondents

8. Analyze the results

Repeat step 1 — make changes, prepare a new version

Repeat steps 2–8

Conduct another round of tests.

Additional Recommendations

Mouse in Your Hands or the User’s?

• Mouse in your hands, you scroll, click buttons — the preferred option. Pros: the user can’t quickly and silently scroll down, and you get to know their thoughts in slow motion. In this option, don’t ask the user “What to do? Where to click?”, just wait for commands.
• Mouse with the user: you see how they move the mouse, where they try to click, what they highlight. Viewing speed is close to real. BUT! In this option, the respondent speaks less about where they look, what they notice, and what they do — you lose valuable information.

The corridor UX test mimics the usual situation

— the user one-on-one with the service, no one around, no help, no hints, no explanations. The only difference is that they clearly voice their thoughts and click buttons through you.

• Do Not Explain to the User what and how they need to do or why this element is in this place.
• Do Not Answer Questions.
• Do Not Prompt. User stuck and doesn’t know what to do? Great — this is valuable information for you! Ask: “What do you see? What do you think? What will you do?”

Ignore expert judgments

Respondents may share their opinion on color, placement, controls, convenient options, examples of similar services. You are NOT interested in this now. Don’t argue, don’t convince — don’t waste time and energy on this — you don’t need it. Politely listen and then bring the user back to the same 3 questions. You are interested in only the respondent’s behavior and reactions, not their (non)expert opinion.

What NOT to Do

• Do not give control (mouse).
• Do not answer questions.
• Do not pay attention to judgments and opinions.
• Do not explain details.
• Do not prompt.
• Do not say anything besides the 3 questions.
• Do not react with facial expressions.
• Do not persuade.

Limitations of the Hallway usability tests

• You will NOT confirm that your text, path, scenario, or design is good.
• But you will be able to find its errors, which are usually quite a few, especially in the first versions.
• Do users see what we want them to see? Do they pay attention to the highlights we need?
• What do they NOT pay attention to, although we would like them to?
• How do they perceive and understand what they paid attention to? As we intended? Or somehow differently?
• Is it clear enough, interesting, convincing?
• What prevents them from going further? What stops them? What is missing?

Important, Less important, Harmful

Important

• Where they didn’t click
• What they didn’t see
• Reactions and emotions.

Less Important

• Where they clicked
• What they saw.

Harmful

• Judgments, opinions, recommendations.

Acknowledgements

Many thanks to the Product Heroes and Rick.ai teams for spreading knowledge and practices on product development.