# Prototypes Decide when to test a prototype and prepare it for a remote study. To test an interface, you will need material: a prototype, static screens, or images. In product teams, prototypes are usually provided by designers, but sometimes researchers create them themselves if they have the experience and the right tools. To quickly prepare materials and adapt them for testing, a technical brief will be useful. It is usually created by the researcher, who already has an approximate research design in mind. The technical brief should think through and describe: - The required screens and the connections between them. If the materials are static, which exact screens or images are needed? If we show a prototype, from which screen will the flow start and where will it end? Which transitions are needed from one step to another, will they be linear, or will we allow users to go back? - Functionality. Do we need an interactive prototype or a static screen? Which elements need to be active? Which actions does the respondent need to perform? - Target action. What main action are we checking, for example placing an order? Which actions will be considered indicators that the task was completed correctly? - Element states. Many interfaces have elements that change status, such as color, enabled state, or activity. To help respondents correctly understand what is happening on screens during testing, you need to think through the state of key elements, for example toggles on or off, buttons active or inactive. - Data on screens. If you are testing an interface with personal data, products, or other objects, think through what they will display so they are suitable for different respondents while creating a relatively relevant situation for each of them, for example account balance, dish options, locations. If you already have a prototype or know that the test material will be a prototype, check its readiness using our checklist. And if you still need to decide what to show respondents and at which stage of the project to start testing, let’s figure it out together. [Prototype preparation checklist](https://wynde.evgeniichekov.com/en/guide/prototipy-kakie-byvayut-i-kak-podgotovit-k-testirovaniyu#chek-list-kak-podgotovit-prototip-i-chto-sdelat-posle-testa) ## When and what to test To get feedback on an interface concept, you can start researching it at the earliest stage of idea formation. Moreover, the earlier you start, the more confidently you can move forward in designing the user experience. Testing material can include: - Static screens - Interactive prototypes, from low to high fidelity - Test environments and beta versions of the product - A live product version, production Another nonstandard way to get user data for interface design is co-creation, where the client creates or edits a layout together with you in real time according to their vision. Such a session can be organized offline, by drawing on paper, or in online tools. This format helps understand the user’s mental model and generate new ideas. Different levels of prototype and material detail correspond to different stages of concept development. The scale on which the desired level of detail grows can be illustrated like this: ![Test material can become more detailed as a product moves from concept to release.](https://wynde.evgeniichekov.com/guide/illustrations/c8e3b849-83fa-497c-83e7-579b1dae3666.png) This means that at the idea stage, you can test very rough prototypes or static screens that show the concept to understand whether you are moving in the right direction. However, the closer you are to releasing new functionality, the more important it is to be confident that everything works as expected throughout the whole flow and that corner cases are considered. Example: If you are creating a new app for a music streaming service, one example of material for an early test could be a simple static home screen. With it, you can check whether a potential user sees all the important blocks they care about. After that, you can design the full interface with more confidence. But before moving to development, another testing stage with a detailed interactive prototype will be useful to make sure the user flow works without issues. Pathway lets you use static images, detailed prototypes, a link to an already working website, and even record a respondent’s actions in the browser as test material. This makes our service an ideal tool for testing at every stage of your project development. ![An interactive prototype is open on a mobile screen from a Figma link.](https://wynde.evgeniichekov.com/guide/illustrations/bd513ccf-ee7a-492e-a913-a30020ab4dc1.png) ## Specifics of prototypes for unmoderated testing The unmoderated testing format determines how prototypes should be prepared. Here is what this means: - Start point and flow length. In moderated testing, we can afford to start far away, for example directly from the product home page, even if we are testing functionality hidden in the profile, so we can also check how intuitive the path to the target action is for the respondent. In unmoderated testing, however, the participant usually spends less time with the prototype, and unnecessary flow length increases the risk that they never reach the target area. If it is not critical for the research design, you can start 1–2 screens closer to the target area, for example directly in the profile. - Number of degrees of freedom. In moderated testing, the researcher can gently guide the respondent if they go the wrong way and bring them back to the target flow. When a person works with the prototype alone, you need to minimize the chance of getting lost or stuck: think through clear transitions without dead ends and limit prototype interactivity to only the necessary actions. - Stability level. For unmoderated testing, the prototype needs to be carefully checked for possible errors. Without a moderator, the user will not be able to report that the prototype broke, and you will not be able to fix it quickly. If you want to compare several prototype versions, for example to test flow completion with two visual-solution variants, you must strictly ensure that only the control variable changes between versions, for example different profile icons. Otherwise, the results cannot be considered relevant because it will be impossible to filter out the influence of other changes. This rule applies to both moderated and unmoderated testing. ## Checklist: how to prepare a prototype and what to do after the test Before the study - [ ] Create a prototype for all tested devices: iOS, Android if there are meaningful differences, mobile and desktop web, and software-based applications. - [ ] Check that the prototype matches the respondent’s device. - [ ] Move the prototype into a separate project, not just a new page. This reduces the risk of opening the wrong layout. - [ ] Provide link access without requiring app installation. Make sure the link is available to everyone. - [ ] Make names, prices, locations, payment methods, and other information inside the prototype realistic. Milk cannot cost 1,000 rubles, and “John Doe” will not make the test feel realistic. - [ ] Remove extra screens or branches that are not related to the test. If some were created for demo purposes, delete them. - [ ] Make sure all tasks and screens form a logical, consistent story and match the research objectives. - [ ] Eliminate dead ends. The user should always be able to continue the scenario when moving from one task to another, even if they chose the wrong path. - [ ] Make all fields clickable and empty, without autofill. Pre-filled fields confuse users and make them think something was added automatically. If there are many fields to fill in and you need to test whether they are clear, Figma may not be suitable. In that case, consider other prototyping tools, such as Axure. - [ ] Hide hotspots, the clickable-area highlights. Remove them before creating the link. If you cannot disable active zones in Figma, ask the respondent to do so before the test starts and provide clear instructions. You can also edit the link by adding “&hotspot-hints=0” to the end or changing the value from “1” to “0.” - [ ] Check that the URL does not reveal the purpose of the test or the company name. For example, “onboarding_check” is bad; “problem17” is good. If brand anonymity matters, remove any mention of your company from the link. - [ ] Make sure the URL does not reveal what you are testing. For example, “onboarding_check” is bad; “problem17” is good. If you need brand anonymity, the link should not contain a reference to your company. - [ ] Split heavy prototypes into parts and number them. Services like Zoom often struggle with the load. - [ ] For unmoderated testing, split the prototype into several parts, each dedicated to a specific task. The researcher can help you define what those tasks should be. - [ ] Add the ability to go back at any stage. This lets users repeat their steps if they get confused, and lets you or the researcher ask additional questions. - [ ] If you compare several screens, the differences should be only in the tested element. - [ ] If needed, insert intermediate screens for context, for example “two months later.” Additional checklist for international products - [ ] The interface should match the country where the test is being conducted. Pay special attention to the displayed banks, names, and country. - [ ] All prototype text should be in the appropriate language, depending on the test language. After the study is complete - [ ] Adjust access settings or delete the prototype. If you delete it, make sure the prototype was in a separate project and was moved back to the main one after completion.