Running experiments and building prototypes go hand in hand. When doing any kind of validation, you need something to show to people in order to get their feedback. That’s what prototypes are for.
What is a Prototype?
When validating ideas for startups, products and services, usually people talk of building a prototype, mockup, proof of concept, or even an MVP or Beta as if these are interchangeable.
This leads to unclear situations and can mean you lose track of what you’re trying to achieve.
If you’re not careful, the image that the words ‘minimum viable product’ conjure up is that of a shiny, usable, ‘feature complete’ product, ready to be rolled out to actual users. This can put you on a slippery slope: instead of using a lean approach, focusing on validating assumptions, what started as an ‘MVP’ becomes a goal in itself, and feature creep and gold plating set in.
Taking the name ‘minimum viable product’ literally , the main differences between a true prototype and an MVP are that a prototype does 1) not need to be functional, and 2) does not need to be a viable product. Understanding this can make your life a lot easier and speed up validation.
A prototype is nothing but a tool to help you validate a specific assumption.
Validation requires prototypes to be able to create strong and objective signals that tell you if your assumption is validated or not. Asking people about their preferences and interviewing them about what they ‘might do’ when a product or service became available is simply too unreliable. They can’t answer such questions truthfully.
The only way to be at least reasonably sure of what people will actually do is by putting them in the situation where the product or service exists, and observing their behaviour.
The function of a prototype is to provide as cheap and simple a way as possible to create that situation.
Therefore, prototypes usually involves faking (aspects of) the actual product. The point is to present users with a situation where they can experience the product or service as if it already exists, in order to observe their behaviour. It is not the goal to build a ‘light’ version of your final product.
On the cheap
Prototypes come in different levels of complexity and detail, ranging from ‘lo-fi’ prototypes to ‘hi-fi’ ones. Lo-fi prototypes are far cheaper and faster to make, so it is worthwhile to try to validate assumptions with the lowest fidelity prototypes you can get away with: this means you can iterate faster (and learn faster) and you can try more things.
It is not necessarily the smallest product imaginable… it is simply the fastest way to get through the Build-Measure-Learn feedback loop with the minimum amount of effort.
So, rather than looking at the hi-fi side of the range, where MVPs, Betas and technical proof-of-concepts live, it makes sense to have recipes for lo-fi prototypes that you can use to quickly test something.
What prototype should you build?
The answer is: the simplest one you can get away with. Because you’re trying to validate or invalidate an assumption, you don’t always need to prototype your actual solution. If you can come up with a prototype that really helps you to validate an assumption and will give you a clear signal, and that prototype looks nothing like what you finally want to build, that is great.
How to Run Experiments for Problem Solution Fit
Quickly probe your startup’s ‘solution space’ with mini propositions
Try to work from your Riskiest Assumptions and be creative in the way you want to test it, define an experiment, and try to come up with the cheapest, simplest way to get a good signal. Look at the list of 20 prototypes below for inspiration, but keep in mind that simpler and faster is better!
#1 Impostor Judo
The easiest way to prototype is not to prototype at all. If there are similar or competing products or services in the market, even if they miss certain features you assume to be key, simply organize a user test of that product or service. This is the easiest way to find out what users think are key features of the product or service, and it will allow you to zoom in on anything they are not happy with.
Example For a mobility startup, we needed to validate if the location of vehicles was an important factor in deciding to register. We located users of three existing car sharing services and interviewed them.
Start pre-selling the product or service before you have created it to gauge interest. Of course you do need to be careful not to trick buyers into buying something you have no intention of delivering, but if you are upfront about it, this strategy can give a clear idea of the interest for your product or service.
#3 Fake Ad
Create a fake advertisement and place it on a screenshot of a website or photoshop it into a magazine. This is a cheap way to give respondents the idea your product or service is legit when you are interviewing them.
#4 Data sheet
Create a data sheet for your product or service, listing the most important features and characteristics. Show this to potential customers to get a sense of what features they find most important.
Create a brochure (digital or using a cheap print-on-demand service) for your product or service. This requires you to come up with plausible copy and branding, but remember that it is only a prototype and that it definitely does not have to be your final branding or name.
#6 Product Box
Create packaging for your product or service, as if it would go on a shelf in a shop. What do people want to know when they pick up the box? What are the key features and information they look for?
Storyboard how your product or service works, as if you are making a comic of the customer journey. Use this to go over the customer journey with potential customers and test assumptions you have for that journey. If you can draw or have an illustrator at hand, you can even change some of these drawings directly with the potential customer.
Create a fake version of the product, and show it to people to see their reaction. It can be a digital mockup that you experience on a device, or a physical one. For physical mockups, this allows you to see how a potential customer responds to qualities like size, weight, shape, and colour.
Example For the book Design A Better Business we created mockups to see what colour cover would be most visible in bookstores. We used books with similar size and weight and stuck on our mockup covers. We placed these in a bookstore and observed. When someone picked them up, we would ask why they did that and what their expectations were.
#9 Paper Prototype
Instead of using programming and design tools to build a digital prototype, simply use paper, scissors, glue, and markers. It allows everyone on the team to contribute, is very cheap to change, and you can show it to people to get their initial reactions quickly.
#10 Digital Mockup
When wireframes or designs of screens for a digital product or service have been created, convert them into a digital mockup that allows you to observe users interacting with your product. Using e.g. Maze you can also send these prototypes out to people to gather feedback.
#11 Landing Page
Create a landing page for your product or service. Measure interactions on the page, and have users sign up.
#12 Fake Button
Add a new button on an existing site for a new feature, and measure how many people click it. The button doesn’t actually work, but leads you to a ‘coming soon’ message, or perhaps to a questionnaire.
#13 Waiting List
Create a waiting list with email addresses, and give out invites or run a simple campaign. Track how fast you sign up people for the waiting list.
#14 Video Trailer
Create a video trailer of how the product works, and use that to give potential users the idea it is a real product or service.
Example Dropbox created a video demonstrating their product well before it was finished to capture sign-ups. (Watch founder Drew Houston talk about this experiment on youtube)
#15 A/B test
Create multiple versions of your prototype. Each version is different only in the aspect you want to measure. See which of the versions is preferred by potential customers.
#16 Play through
When e.g. organizing a complex event, or designing a complicated interaction or business process, do a play through with your team or with potential users to see if you are on the right track. Use board game pawns and counters to play out all the different steps of the interaction and see what kind of decisions users make and what questions they have.
#17 Wizard of Oz
When you’re developing a product or service that depends on a complex piece of software or automation that doesn’t exist yet, you can rig the prototype so that a human takes the place of that software. The users think they are communicating with a system, but behind the scenes actually a user performs the interactions.
Example For a cybersecurity proposition, one of the ideas we proposed was to build a chatbot. We wanted to quickly understand if users would be open to this form of interaction, without building anything. We simply bought a prepaid sim card, setup WhatsApp, and named the account ‘Kevin’, with a nice robot profile picture. Then we created a script and asked respondents if they would drop Kevin a message. Of course, in reality we were doing the chat ourselves, following the script closely.
#18 Concierge Model
This is similar to the Wizard of Oz prototype. A human manually fulfils a (large) part of the interaction. In a Concierge Model, the potential customer doesn’t need to think they are talking to a computer, however.
Example Zappos founder Tony Hsieh simply bought the shoes he offered online in local retailers. Dutch startup 3D Hubs initially served their customers by finding owners of 3D printers in Amsterdam willing to take paid printing jobs.
Another way to reduce complexity is by foregoing the promise of continuity. It is much easier to run a shop for a week than it is to operate year round. Pop-up shops are a great way to see if there is interest for something before going all in.
#20 Raise the Price
A nifty way to figure out if there is interest and getting price signals at the same time. The first person pays the lowest price, the second a slightly higher price, and so on. You keep raising the price until it starts to become hard to convince people to buy.
Example For Business Model Generation, Alex Osterwalder used this approach to raise awareness and interest people to be a proofreader.
There are undoubtedly many more kinds of prototype out there, and I’d love to hear about them, so please don’t hesitate to tell me about them in the comments!
-- Keep experimenting 🚀