Add as preferred on Google
How a Repository Health Dashboard Helped Open edX Patch certifi in 48 HoursRead More
Concept Testing in 2027: Everything You Need to Get Started
Key Takeaways
Concept testing replaces guesses about demand with reactions from likely buyers before engineering spend, and a first test can run on free or low-cost tools.
- Test after the idea can be described, before production code.
- Start with one testing goal and decide in advance which result means build, change or drop the idea.
- Recruit buyers with the problem and budget; friends tend to be too kind.
- Surveys measure how many people like an idea, interviews explain why, prototype tests show understanding, and A/B tests compare two versions.
- For a first test on a small budget, we recommend sequential monadic surveys with randomized order, which need a smaller sample.
- InVision has shut down and Adobe XD is in maintenance mode.
- Leading questions, wrong audiences and treating one positive survey as proof skew results, because saying you would buy costs participants nothing.
Use the cheapest method that answers your goal, retest only what you change, and move to a prototype when evidence supports building.
A team can spend six months building an app and only then learn that buyers wanted something else. Concept testing catches that problem early. You show a description or mockup of your product idea to people who match your target customer and measure how they react, all before anyone commits an engineering budget. For founders and product managers planning 2027 launches, a first concept test needs no research agency and can run on free or low-cost tools. The harder parts are knowing when to test, which method suits your goal and which tools still exist, and each gets its own section below, along with a seven-step process and the mistakes that skew results.
What is concept testing and when should you run it?
Concept testing is early-stage research that checks whether a product idea makes sense to likely buyers and solves a problem they care about. Timing matters. It happens after you can describe the idea in a few sentences and before anyone writes production code. Test any earlier and people have too little to react to; test after development starts and most of the budget is already spent, with findings that arrive too late to change much.
What you show people is the concept stimulus. It can be a one-paragraph concept statement, a landing page, a clickable mockup or a short video. Our post on validating a product idea before you build it makes the point that prototyping doesn't need to produce something functional, and gives the example of Dropbox, which validated its idea with a simple explainer video before development began.
Why test a concept before building it?
Testing a concept lowers build risk because it swaps guesses about demand for reactions from people who might pay for the product. The gains show up in four places.
- You learn whether the idea is viable while a change still costs a page edit.
- Real reactions show which features buyers care about and the words they use to describe the problem, which later helps with product copy.
- Mistakes found in a mockup are cheaper to fix than mistakes found in shipped code.
- People who helped shape a product can become its first users, though a test alone won't create loyalty.
How to do concept testing in seven steps
Steps 5 to 7 repeat until the results meet the goal set in step 1. The running example is a founder testing an invoicing app for freelance designers.
Step 1. Write one testing goal
A concept test needs one decision it must inform. "Will freelance designers pay a monthly fee for automatic invoicing?" gives the test a clear pass or fail. A vague goal such as "get feedback on the app" produces opinions nobody can act on. Before the test starts, write down which result would lead you to build the product, which would lead you to change it, and which would tell you to drop the idea and move on.
Step 2. Recruit people who match your buyer
Test with people who have the problem and the authority or budget to solve it. Friends tend to be too kind. A short screener asking about role and current workaround filters out the wrong people; interviews need a handful of participants per customer segment, while surveys need enough responses that one loud opinion cannot swing the result.
Audience fit can change the outcome entirely. Arbisoft saw weak interest in Xiangqi.com, its Chinese chess platform, from Western users, then localized the product for the Chinese market and saw a substantial increase in daily active users. A concept that fails with one group can still succeed with another.
Step 3. Prepare the concept stimulus
Every participant should see the same version of the concept. For the invoicing app, that might be one page covering the problem (late payments and manual invoices), the proposed solution, three features and the price. Keep it short. A ten-screen prototype wears participants out before they get around to judging the idea, and a stimulus without a price tells you nothing about willingness to pay.
Step 4. Pick a method that matches the goal
Each method answers a different question. Surveys measure how many people like an idea, interviews and focus groups explain why, prototype tests show whether people understand how it works, and A/B tests show which of two versions performs better. The invoicing founder wants a price answer from many designers, which points to a survey with a pricing question followed by a few interviews to explain the numbers. Cost and fit for each method come next.
Step 5. Analyze feedback for themes
Group open-ended comments into themes and count how often each one appears. One complaint is an anecdote. When most participants raise the same confusion about how invoices get sent, that is a design problem to fix before building. Weigh behavior above stated interest, because a click on a "Join the waitlist" button says more than a polite comment that the idea sounds useful.
SurveyMonkey and Maze both now offer AI features for analyzing responses and spotting trends. Our view is that these summaries are a useful first pass, and that reading a sample of raw answers yourself still catches sarcasm and the exact words buyers use.
Step 6. Refine the concept
Change the concept based on the strongest themes. That could mean a new design, a different feature set, sharper positioning or a lower price. If the result missed the threshold you set in step 1 by a wide margin, dropping the idea is a valid outcome and often a cheap lesson.
Step 7. Retest the changed concept
Run a smaller round on the changes with fresh participants, since people who saw the first version will compare instead of reacting to it on its own. Stop when the results meet your goal. AI tools have shortened these loops. Arbisoft's product managers now use AI prototyping and codegen agents to create functional prototypes quickly, which narrows the gap between one test round and the next.
Which concept testing method fits your goal and budget?
The cheapest method that answers your step 1 goal is usually the right one. Five common concept testing methods appear in the table, with what each reveals and what it costs in time and money.
| Method | What it tells you | Effort and cost | Best for |
|---|---|---|---|
| Online survey | How many people like the idea, which features rank highest, reactions to price | Low | Measuring appeal across a large audience |
| Interviews or focus groups | Why people react as they do, in their own words | Medium | Early ideas where you still need to learn the problem |
| Prototype test | Whether people understand the concept and can follow its main flow | Medium | Apps and digital services with a clickable mockup |
| A/B test | Which of two versions gets more sign-ups or clicks | Low to medium | Choosing between two messages, designs or price points |
| Digital simulation (AR or VR) | How people react to a physical product in a realistic virtual setting | High | Hardware, gadgets and physical spaces |
Survey-based concept tests come in two common designs. In a monadic test, each respondent sees one concept in isolation, which reduces order bias but needs a separate audience for every concept. In a sequential monadic test, each respondent rates several concepts, which needs a much smaller sample, although answers may be influenced by the concepts shown earlier. For a first test on a small budget, we recommend sequential monadic with the order randomized.
Which concept testing tools are still current?
A first concept test usually needs a survey tool and a prototyping tool, plus a recruiting platform if you cannot reach buyers yourself. Some older recommendations no longer hold. InVision shut down its platform at the end of 2024, with account access ending on 31 December 2024, according to Silicon Republic's January 2024 report. Adobe told The Register in January 2024 that Adobe XD is in maintenance mode and that the company has no plans to invest further in the product. New prototyping work is better started in another tool.
Prototyping has also moved toward AI. Figma announced in July 2025 that Figma Make, a prompt-to-app tool for creating interactive, high-fidelity prototypes, had reached general availability. A founder can now describe the invoicing app in plain language and get a clickable draft to refine, though it still needs a designer's review before participants see it.
| Tool | Type | Role in a concept test | Best for |
|---|---|---|---|
| Google Forms | Survey | Builds forms and surveys to collect reactions | Quick, low-budget surveys |
| SurveyMonkey | Survey | Surveys with AI features for spotting trends | Larger surveys that need analysis help |
| Figma and Figma Make | Prototyping | Clickable mockups; Make turns prompts into interactive prototypes | Digital product concepts |
| Maze | Research platform | Tests ideas, prototypes and messaging, with AI-powered analysis | Unmoderated prototype and concept tests |
| UserTesting | Participant network | Feedback on concepts and prototypes from recruited participants | Reaching a specific audience quickly |
| BetaTesting | Tester community | Targeted testers for field tests, surveys and moderated interviews | In-home and field usage tests |
Pricing and free tiers change often enough that it's worth checking each vendor's current plans before you commit.
Concept testing mistakes that skew results
Five mistakes are worth checking for while you plan the test.
- Testing too early or too late. A vague idea gets vague reactions, and a test run after development starts arrives when most of the budget is spent.
- Testing with the wrong audience. Feedback from people who will never buy the product points the roadmap in the wrong direction.
- Overcomplicating the test. Long surveys and too many concepts tire participants, and tired participants give careless answers.
- Asking leading questions. "How much would this app improve your week?" assumes the app helps. A neutral question such as "How do you handle invoicing today?" lets the participant describe the problem first.
- Treating one positive survey as proof. People are polite in research settings, and saying they would buy something costs them nothing.
Frequently asked questions
What is the difference between concept testing and usability testing?
A concept test asks whether a product is worth building. Usability testing comes later and checks whether people can complete tasks in a working design or product, which needs something far more finished than the description or mockup a concept test shows to likely buyers. A prototype-based concept test can surface some usability problems early, but a full usability test still belongs later in development.
Can you run concept testing without a prototype?
Yes, a one-paragraph concept statement, a landing page with a waitlist button, a storyboard or a short explainer video all work as stimulus. Prototypes help most when the question is whether people understand how the product works.
How many concepts should you test at once?
Our view is that a first concept test should compare two or three versions at most, such as two price points or two positioning statements. More versions need more participants in a monadic design and tire people out in a sequential monadic design. Testing one variable at a time shows which change caused the difference.
Does a positive concept test mean the product will succeed?
No, a positive result only means likely buyers understood the concept and reacted well to it under test conditions. Pricing, competition, execution and timing still decide success. Treat a good result as permission to move to a prototype or MVP, and keep measuring real behavior after launch.
Moving from concept testing to a working prototype
A concept test earns its cost when it changes a decision. Start with one goal and the cheapest method that answers it, and retest only what you change. When the evidence supports building, the next step is a prototype that holds up with real users, and that's the work Arbisoft's UX and product design team does through user research, persona creation, interactive prototypes and usability testing with real users. Talk to the team about turning your tested concept into a prototype your engineers can build from.
...Loading Related Blogs
