The MVP Experiment Canvas
A practical framework that helps you design and launch Minimum Viable Products to validate your business ideas effectively.
(THIS IS A DRAFT, PLEASE LET ME KNOW YOUR THOUGHTS/EDITS BY COMMENTING, THANKS!)
Over the past few years I’ve sat down with dozens of (aspiring) entrepreneurs to talk about their ideas, products and strategies. There have been several themes that they struggle with, but there’s been one question that keeps coming back:
I’ve made this “thing”, how do I get people to use it?
This question is not about acquisition or ‘growth hacking’.
- Not knowing for who you have build your product (who’s problem are you solving?)
- Where to find these people
- How to talk to them and show that your product solves their problem.
Going from a business idea that’s in your head to something that you can show, and deliver to your target customers, is the biggest challenge in the early stages of an idea. Why? Your time is limited (unless you have infinite money and motivation) and you have to figure out if your idea is interesting enough to pursue (with more of your time, money and efforts).
Whether you can code or not, most ideas are about a solution (not about a problem), so most people start on the path of building their idea of that solution, in whatever shape or form.
What they - in my humble opinion - are forgetting, is that their solution can still change a dozen times. You can change any element of the “thing” you (want to) create. Because in the beginning it’s all about delivering the value you think your target group is looking for (and discover if you’re doing it right). Even if you’ve acquired 100+ people you can still change anything you like!
What should be focussed on in these early stages are the fundamental questions around an idea:
- Who is experiencing the problem I want to solve?
- Is it a big enough problem to solve?
- If it is, what’s the value proposition I should pitch to them?
- Where can I find and reach these people to validate my solution?
While searching for the right answers to the questions above, don’t worry too much about your “solution” at that point (you can still change it a dozen times!).
I truly believe that if you really deliver value and solve someone’s problem, they don’t care how you do it.
To take the business idea that’s in your head and turn it into something you can investigate and validate, the “Minimum Viable Product” has been the approach to do so ever since Eric Ries popularised it with his book “The Lean Startup”.
So, what is a Minimum Viable Product (MVP) ?
There have been many discussions ever since the term MVP got more popular. To start I want to share what I believe it is not:
An MVP is not a first release of a rudimentary product. As in: “my final product has 10 features but I’m launching my MVP with 3 core features so people can start using it.”
This technical approach will get you a crappy product that will not help you to validate your idea in the right way (with the right speed and direction). It will probably get you a product that is not minimum, but something that is too big, well-described in this blogpost.
I follow the original definition of Eric Ries:
“A Minimum Viable Product is that version of a new product which allows a team to collect the maximum amount of validated learning about customers with the least effort.”
and the refined version of Ash Maurya:
“A Minimum Viable Product is the smallest thing you can build that delivers customer value (and as a bonus captures some of that value back).”
But, I have some notes to add to these definitions:
- “that version of a new product” = not necessarily a technical product, it could be anything. A talk at a conference to validate a thought concept, a real life workshop to validate an online course, a landing page to validate a SaaS tool, a video to validate a file syncing service, etc.
- The “product” is a temporary combination of: Value Proposition + (Part of your) Target Audience + Distribution channel to reach that audience + a way to establish a Relationship with that audience.
- “Minimum” can be interpreted in multiple ways. It should be enough to test your assumption, but small enough to be disposable (it’s simply a tool to learn and not the V1 of your end product). How minimum should your MVP be, according to Eric Ries? “Probably much more than you think.”
- “Viable” means “capable of working successfully” → Whatever you create as an MVP should work successfully to help you deliver the value to your customer and with the validation of your idea’s riskiest assumptions.
- “customer value” can also be “perceived customer value”, depending on what you are investigating with your MVP. For example: if you want to confirm if you’re talking to the right audience, about solving the right problem, you might do this with a lead generation landing page where you communicate the perceived value for these people: your solution to their problem.
- “least effort” → this is different for anyone and depending on what you want to validate and how. My rule of thumb with the canvas that I’m introducing is that you can have your first MVP experiment up and running within two weeks.
- “Capturing value back” is anything of value that helps you in the validation of your idea. Money is an example, but feedback and the foundation of an ongoing relationship with your chosen target customer is also value.
To me, a Minimum Viable Product is a structured experiment (code or no code) that helps you validate the most riskiest assumption(s) of your idea. Basically anything that helps you to de-risk it’s uncertainty so you can decide if the idea is something you should pursue. It’s disposable and can be repeated in different setups until you reach an acceptable level of (un-)certainty.
As you may know most startups (42%) die because there is no market. So in the early stages of an idea, I believe an MVP is mostly used to investigate the Desirability of an idea: do people want this? Viability (can it work successfully?) and Feasibility (can it be built?) are only relevant challenges when you are more confident that your solution is something that people want.
Using the MVP Experiment Canvas to design a Minimum Viable Product for your business idea
Inspired by this amazing post by Tristan Kromer and Ash Maurya’s original “Experiment Report”, I’ve created an easy to use and free tool that helps you streamline the process of designing an MVP experiment.
I’ve taken Tristan’s idea of using 4 elements of Alex Osterwalder’s Business Model Canvas (the easiest way to map your idea and its business model in one overview, which I recommend doing before you use the MVP Experiment Canvas), this is the foundation of your MVP experiment:
1. Your Customer Segment 👤
Your “big idea” delivers value to a certain target customer group. For your MVP experiment it’s important to look for specific characteristics that can help you to define different customer segments, from that bigger group, that can either pay, receive or decide on your value proposition.
Who’s problem are you solving? / Who are you providing value for? To describe your target customer, check out the Value Proposition Canvas.
- What are their needs / what do they want to achieve?
- What pains are preventing them from fulfilling this need?
- Which gains would make it easier for them to achieve what they want?
Next, you decide which segment (from the bigger group) will experience the most value from your MVP Experiment. But also, can you learn from them? Pick the most viable Customer Segment to work with and move on to step 2.
2. Value Proposition 🎁
What are you offering this Customer Segment? Remember, this is a temporary version because we’re running an MVP Experiment for this specific segment.
Describe your offering with one of the following formats:
My company, (insert of company),
is developing (a defined offering)
to help (a defined audience)
(solve a problem)
with (secret sauce/differentiation).
For (your target customer)
who (statement of need or opportunity)
our (product/service name)
is (product category)
that (statement of benefit).
3. Channel(s) 🚚
How are you going to reach your Customer Segment and “deliver” them your Value Proposition?
The channel that works best depends on the segment you chose. Running an MVP Experiment will also teach you how accessible the segment is through this channel.
The most common channels for an MVP Experiment are:
- Your own Network (direct or referral)
- Online Forums & Social media (search to find discussions on any given topic)
- Meeting IRL (in stores, conferences, door-to-door etc.)
- Cold emailing
- Aggregator websites or blogs
- Business directories and job boards and freelancer sites
For more inspiration, check Gabriel Weinberg’s overview of possible traction channels. Remember that any channel is disposable and that you pick one that does not cost too much time or money to get you results.
4. Customer Engagement ❤️
How are you going to engage with the people who use your MVP? Once you have delivered your value proposition it’s time to learn from their experience, gather feedback and establish a relationship to help you in the validation of your idea.
Because an MVP Experiment is small and contained, there are only a few ways to do this in a qualitative way:
- Follow-up meeting, phone or face-to-face (interview or open conversation)
- Follow-up email with a Call-to-Action (send reply, fill out survey etc.)
If your MVP involves a digital product you could also use a tool like Google Analytics to measure basic conversions and click events, or record user sessions with a tool like Hotjar to get more insights into a specific set of people you’d like to engage with, and the questions you want to ask them based on their actions.
5. Riskiest Assumption(s) 💣
What you are setting out to test (and prove) with your MVP?
Which assumptions if invalidated, will kill your idea? To minimize risk and costs, you have to find out if your Riskiest Assumptions are true, as soon as possible
Assumptions can revolve around 3 subjects:
The Problem (Assumption about your market or customers):
“For the problem to be important enough / solved, it is necessary that ….”
The Solution (Assumption about the value of your proposition):
“For the solution to solve the problem / succeed it is necessary that ….”
The Implementation (Assumption about technical aspects of your solution)
“For the implementation to work / succeed it is necessary that ….”
For most startups and new products though, the riskiest assumption is the existence of a market. That’s also why almost all MVP’s revolve around: “is this something people want?”
For your MVP Experiment, you can use my Riskiest Assumptions Ranking Tool (.xls file) to easily write down and rank your assumptions to determine which one you need to investigate with your MVP Experiment.
6. Experiment Format ⚗️
Now that you know what you want test with your MVP, it’s time to decide what type of experiment you are going to run. As mentioned, most MVP’s are about the Desirability of an idea (do people want this? do they care?).
Here are some examples of what you can do:
- Landing page for leads → “We’re launching soon, sign up to get notified”
- “Fake door test” → Signing up shows a “Woops! We’re not ready yet” page (great example by Joel Gascoigne of Buffer)
- Pre-order landing page → validate demand + willingness to pay (you can deliver the actual product or refund everyone later)
- “Wizard of Oz” → Simulate the automated delivery of your value proposition, but you actually do everything by hand without your customer knowing it. For example, Andrew Mason started Groupon with a simple WordPress site where he posted daily deals for a local pizza shop. He manually created PDF coupons as he received orders and emailed them via his personal email.
- “High bar smoke test” → Test customer’s willingness to pay without letting them pay directly (For example: they know about your pricing but have to fill out a long intake form to sign up).
- “Broken promise smoke test” → You tell X people about your sign up page to test if your value proposition is interesting enough for your customers to tell more people (leading to more than X signups).
7. Experiment Scenario / Workflow 🔄
How will your experiment work? What are the steps your customer and you have to take? Describe them from beginning to end.
For example: in 1999, when Zappos (bought for $1.2B by Amazon in 2009) was just an idea of Nick Swinmurn, the most riskiest assumption about selling shoes online… was that people were actually willing to buy shoes online (as apposed to a physical store where they could try them on).
- Swinmurn took photos of shoes in a local shoe store and put them online.
- When a pair sold, he bought the shoes and shipped them to the customer himself.
- He gathered feedback from customers and evidence around the business model.
This was obviously not sustainable for the long run, but after completing enough orders, Swinmurn lowered the amount of uncertainty around his idea of starting an online shoe store and decided to pursue it.
Another example: the workflow of Uber’s MVP.
8. Metrics 📈
Depending on the type of experiment you choose to run, you also have to decide what metrics you will measure.
Metrics should be:
- Specific → a conversion is what exactly?
- Understandable → everyone should be able to understand them
- Comparative → a ratio or a rate that you can compare over time
- Actionable → Vanity vs. Real metrics
They could be:
- # of interviews
When running your experiment you can measure more than one metric. This is important as measuring more than one can help you find surprising evidence about your riskiest assumption(s).
9. Success Criteria 🏆
You can measure more than one metric but you can only select one metric that defines your experiment’s success. How will you qualify and/or quantify a successful outcome?
“true MVP success is in (a) uncovering long-term product potential and (b) being able to develop a roadmap to unlock that potential.” — Yuval Ariav
- What is your criteria to validate your assumption(s)?
- What would mean successful validation based on that?
- Focus on a metric that has a big impact on your assumption(s)
For example, a conversion success metric: “I will run the experiment with # amount of customers and expect signal X with # of customers.”
Now it is time to launch your MVP! You can run your experiment for as long as you want but I would recommend to run it for 2 weeks, so that you can setup, build, run and analyze one full experiment within one month.
10. Results 🏁
When your experiment has ended it is time to gather its qualitative and/or quantitative results.
- How many customers did you test with?
- Describe what happened?
- What data did you collect? (Refer back to your metrics)
- What user feedback did you receive?
11. Learnings & Insights 🔑
With the results gathered it’s time to analyze them:
- What are your key learnings? (what are the insights you got?)
- What are your key surprises? (did anything happen you did not expect?)
- Did you get enough results? Can you work with them?
What do the results tell you about your Riskiest Assumption(s)?
- Was what happened also what you expected?
- If not, what happened and what does that tell you about your Riskiest Assumption(s)?
- Did you discover new things you need to investigate or test?
- Did you Validate or Invalidate your assumption(s)?
- Or do you need more data to draw a conclusion?
12. Next steps 🚦
What do the results tell you to do next? Are you going to Pivot, Pursue, or Stop?
This is where you change directions and design new experiments to test your assumptions for a new customer segment, problem or value proposition.
(If you pivot you can use the “Validation Canvas” to keep track of your progress.)
Were you able to lower the uncertainty around the Desirability of your Value Proposition? If so, what Assumption(s) should you test next?
If you can obviously conclude there is no long-term product potential (without exploring a pivot), write down your personal learnings and move on to the next idea.
— — — —
That’s it! If you download the canvas you can print it out, put it on a wall to work with, or work with the digital version of the canvas (in PNG, PDF, Google Sheets, Excel or a Numbers file) to set up your experiment.
If you want to learn faster, you could run multiple experiments at a time to test multiple riskiest assumptions with multiple MVP‘s.
- A canvas is just a tool to structure information, and this format is what I came up with based on my experiences and knowledge.
- Go out and explore, research different tools (code or No-Code) and build your own personal “stack” of formats, canvasses and apps that help you take those first steps.
- After defining, building and running a few actual experiments you will get a better feeling at what you’re doing and a better understanding of the approach that suits you.
- I suggest you just start as quickly as possible, it does not matter if your first few customer interviews suck or that you can’t really figure out all the assumptions that you have