The Build-Measure-Learn cycle

There are two main ways to develop a product:

  1. The traditional approach where years are spent in the lab developing a product that looks great on paper – often with cool innovative technologies. However, what often gets built this way is a great product that doesn’t actually appeal to customers.

  2. The modern “startup way” where a company puts the product on the market as soon as it has the basic functionalities in place to collect feedback and iterate. Especially with software products and services, product development continues even after the product has been officially released.

This second approach is much more in line with what Eric Ries calls the "build-measure-learn cycle", which replaces the traditional (and long) product development path. This process follows three steps:

  1. Build a Minimum Viable Product (MVP)

  2. Collect feedback on the MVP by exposing it to potential customers

  3. Iterate the product based on the feedback (not only verbal feedback but how they behave – for example, buying your product)

Initially, you develop an early version of the product – an MVP – with just enough features to begin testing its market potential. The product is then given out to customers to test the following: Do people use it? Will they keep using it? Do they spend money on it?

Once the MVP version of the product has been tested, it can be further developed based on the test results and feedback. This process is important in the early stages of a startup but it should continue through the entire lifecycle of the company.


An example of build-measure-learn

The first round of a normal software startup’s build-measure-learn cycle could look something like this:

1. After validating the idea with potential customers, the startup develops a simple MVP using no-code tools or a minimal amount of coding that has a couple of key functionalities valued by the customer (BUILD)

2. They ship the MVP to those potential customers they’ve interviewed and collect feedback about the solution (MEASURE)

3. The feedback from customers reveals that only some of the proposed functionalities provide any value, and that more emphasis should be put on those features (LEARN)

4. Using the feedback, they now proceed to code the core functionalities and develop them further while leaving the useless features out (BUILD)

The Minimum Viable Product

A Minimum Viable Product (MVP) is the very first version of the actual product: it has the basic functionalities and is already able to deliver limited value to the end customer. The MVP should be as simple as possible, and only contain the bare minimum of working features and core functionalities to serve as the first touchpoint for potential customers. An MVP validates whether the product can solve a customer problem and deliver value to the market, often through a pilot, which we will talk more about in Chapter 5.

In one of the most popular startup books of all time, The Lean Startup, a few practical examples of different types of MVPs are proposed:

1) The Concierge MVP

Here the idea is to serve only one customer to demonstrate what your product or service will be like. Everything is done manually, meaning this approach is not scalable, but the goal is to show that the service is useful. In general, something useful to do in the beginning is “to do things that don’t scale”. If the real product is relatively straightforward to create, you can in the beginning test how customers see the value of your product by doing the solution manually.

Example: The CEO of Food on the Table visited one customer in person, talked with her, and then created (and delivered) customized recipes in person. This helped Food on the Table to acquire feedback and learn what worked and what didn’t before creating the actual service.

2) The Wizard of Oz MVP

In this MVP, there’s “someone behind the curtain” making it seem like technology is at work. But in reality, a person (or people) are doing everything manually. This is different from the Concierge MVP as there is a facade of a working technology product.

Example: Zappos.com began with an online storefront – but when a customer bought a pair of shoes, there were no supply chains or warehouses. The CEO went and bought the shoes in a store and mailed them to the customer. Again, though this approach was not scalable, it proved that the idea worked: people would buy shoes online if they were able to try them on and send them back, risk-free.

3) Other Types of MVP

“Other” - a catch-all term for whatever kind of MVP a person can invent. E.g. creating a product that works (also operationally) like the intended complete product but just with the most critical features. There is no right or wrong way to create an MVP and you can be creative with it, just make sure that it delivers value to customers and is done in the simplest possible way.

Measuring the MVP

The second part of the build-measure-learn cycle is to measure the MVP by exposing it to potential customers and collecting feedback to determine how the product should develop.

Who you choose to test your MVP on will affect the usefulness and relevance of their feedback. In addition to the product’s regular users, consider testing extreme users as well – those who are likely to use the product significantly more than regular users, and those who wouldn’t be likely to use it at all. This will help you to get a new perspective on your product, as the heavy users and non-users are likely to be more vocal and opinionated regarding what works and what doesn’t.

To test your MVP, you always need a hypothesis, and you need to provide a framework and gently guide the customer to test it. One way to think about feedback gathering is to use the so-called Kano model, that separates the features and functionalities into three categories:

1. Basic: These are the features that the customer requires and takes for granted. The MVP’s task is to prove their viability.

2. Performance: Additional features and requirements that customers think about and are in the back of their minds when evaluating options. Their value grows linearly – the more you have, the better your product is.

3. Excitement: Unexpected features and functionalities the customer may not think about, but which are the innovations that ultimately separate your product from others.

Another way to guide feedback gathering is to categorize the features into six key attributes known as “FURPSS”:

  1. Functionality

  2. Usability

  3. Reliability

  4. Performance

  5. Supportability

  6. Security

In many business-to-consumer (B2C) products, for example social platforms, mobile applications, or games, a key metric is user retention – in other words, how well does the product hold on to its user base? While testing their product, these companies will often “soft launch” the MVP version by publishing it for a small number of users in markets where the cost of acquiring customers is low. The purpose of a soft launch is to collect data on user behavior that will allow the product to be improved.

This approach is especially popular with so-called “freemium” or free trial products where the startup’s aim is to first get the user to spend as much time as possible on the product (retention) and only then convert them to a paying user (conversion). After various experiments and adjustments to the product’s features and settings, the product is then tested again. If the retention increases, the changes are retained. If not, the changes will be discarded and reverted to the original version. This will help get real-time feedback on the development process.


Retention is calculated using a simple formula: for any given day, retention is based on the percentage of users who have signed up to use the product. So for example, if a hundred people signed up on Monday and there are fifty users on Tuesday, the D1 retention, or first-day user retention, is 50%. On the following Monday, if there are ten remaining users who signed up for the product a week earlier, D7 retention is 10%. And so on.

A common rule of thumb for a hit product is the 40/20/10 rule. In other words, the product should be able to hold 40% of users on the first day, 20% after a week, and 10% after a month. Long-term retention in particular is a very demanding challenge for B2C products.

Iterating the idea

The third and final part of the build-measure-learn cycle is to iterate the product based on the data and feedback that you’ve measured.

According to The Lean Startup, iteration is when you need to turn the data you’ve gathered (the “measure” part) into actual changes to your product. This is when you need to decide if your project should “persevere” (continue to develop along the same lines the product has now) or “pivot” (change some or all aspects of the product strategy).

To do this, you need to analyze the data received and focus on what is relevant in the feedback that will help you to create a product that is actually desired by customers. You also need to decide if the changes needed are sustainable – in other words, can you maintain the level of service or produce the new product.

If the feedback for the product is generally okay, you probably will still change or add something based on feedback to make the solution better, which is normal product development. After this, you continue with the improved product and test it with the same and new potential customers. Sometimes user data shows that it’s not worth developing the product as is, but you should go back to the drawing board and build a completely new product. This is called a pivot, and is a typical transition in the startup world. A pivot means that a startup transfers its resources to a whole new product or to a new customer group. Doing this is not an easy task – but is vital if your startup is to have the best chance to succeed. Remember you should not love your product, love the problem.


Insights from an entrepreneur – Wolt

Elias Aalto is the co-founder of Wolt, a technology company known for its food delivery platform. Today Wolt has over 500 employees and the service is available in approximately 50 cities, mainly in Europe. Elias shares with us how Wolt got started and how they developed the product.

What is the problem Wolt wanted to solve?

“When I was pitched the idea to be part of founding Wolt, I was sitting in a cafe with one of my co-founders, Miki Kuusi. The queue to the cashier was crazy long. We were thinking that why can’t we just order from the table? Why do we have to do it face-to-face with a person behind the counter? So the initial problem we started solving was eating at a restaurant and ordering take-away. It made little sense that you had to queue and wait when you could do it with your phone beforehand.

“Later on, we also expanded to delivering food. Delivery itself was part of our initial plans, but we were hoping that some general last-mile logistics company would crack the problem and we could outsource it to them. But nobody started doing it well so we had to build it ourselves.

“In the food delivery business, our service was in a way already validated. People want to order food. There had been websites for that for many years – and even before that, people ordered by calling. Our idea was to solve the problem better and to do it for mobile phones. I think that for new companies, the best problems to solve are the ones that have already been validated as business cases. Then you simply build the next generation version of the solution. When you are creating a new company, every step you take is like a roll of the dice and you hope that your guess is right – so if you take multiple steps, your chances of getting them all right are quite small. But if someone has already validated some of the steps for you, it makes it a bit easier.”

How did you develop the first product?

“When developing the product, we opened Sketch and started designing the actual user interface directly. We also tried different interactions to understand how it would feel for a user, like expanding menu items, etc. We were mainly testing the user interface on ourselves. That is an advantage of creating consumer products where you are also part of the target audience.

“I think we could skip many validation phases for our product because our problem was quite trivial and the use case also clear. I had been creating mobile apps exclusively for five years, so I knew what mobile apps should feel like – and we all knew how to order food. If you are building something totally new where you have no idea if it will work at all, you have a lot more validation to do.”

How did you prioritize what to include in the first product you launched?

“We didn’t have a backend for investor demos or even the first in-store trials. We didn’t do anything that was not needed: the user could only create an account, add a credit card, choose the restaurant, place an order, etc. We had a search function, but of course the search was done locally on the device because we didn’t have a backend for it. The user couldn’t remove the account either. Also, they couldn’t change their email address, phone number or other info. We just handled everything like this with customer support. Our COO once said that “customer support allows you to launch tech products two years before they’re ready,” and I think this is true. You don’t need to build all the fancy systems in the beginning.

What was your MVP?

“I think many understand the word MVP incorrectly and they use it to justify very mediocre or bad products. You should do less for an MVP, but the things you do should be great. My pet peeve is non-native apps. Nobody cares if your product is available for all devices if the product is bad. You have to create the facade of a great user experience – the user must believe that everything is working smoothly – even if it’s just smoke and mirrors. For example with Wolt in the beginning, our user interface had whole views that were just screenshot pictures from photoshop, like the boxes and lines in the app, and then I just added the text fields on top of them. If the text had laid out on multiple lines the interface would not have adjusted.

“If you go bankrupt, nobody is coming to collect the technical debt from you. Sometimes the only leverage you have is technical debt and it allows you to do things faster in order for you to reach the next milestone like revenue from customers or financing so that you can have more resources and do things more properly later.”

When did you launch?

“We launched in spring 2015 in Helsinki, and had maybe 20 restaurants available on the service. We got some press for the launch as we had a somewhat well-known team, as well as funding. So we got users and we were growing, but were growing too slowly – our daily sales were maybe €500. We thought about many different options: we launched Wolt for use for example in one nightclub in Helsinki and we were also thinking about having the service in sport arenas, etc. But we didn’t focus on any of those and, looking back, I’m quite happy about that. One fast food-chain also wanted to buy a white label solution from us but we declined as it would have been too much consulting work and would have slowed us down.”

When did you realize the need for a pivot?

“Our takeaway business was growing, but it was just growing too slowly. We realized that our product was not an amazing experience for the first-time user. You had to create an account, add a credit card and so on so for the takeaway purchase – there was a lot that was required from the user compared to the value they were getting. If the user was not happy, they probably wouldn’t come back organically and we couldn’t afford to spend money on marketing to acquire them again as our transactions were relatively small like 20 or 30 euros.

“Then Foodora came to Finland and launched its delivery service. There was still no great last-mile logistics firm to whom we could outsource the deliveries. We also understood that with deliveries, the experience is great even the first time. Creating an account and adding a credit card while sitting on your couch at home is not much work compared to having to jump on your bike and getting the food yourself. So we realized that to compete, we would have to start doing deliveries.

“We launched our delivery service in a couple weeks – at first in collaboration with a small company called Feddy. They had a service where you could send a text message to get basically anything delivered. They would search for what you needed, come back with a price and deliver it. So our very first deliveries worked like this: when we got the order, we texted Feddy to handle the pickup and delivery. It was extremely unreliable and inefficient. Eventually, a bunch of couriers from Feddy called us and suggested that they should come and work for us, and they basically built our delivery operations and are still working at Wolt.

“This is again an example that you don’t have to build an automated system in the beginning, you can do it manually to test if it works. As the deliveries started growing rapidly, we suddenly had maybe 200 deliveries a day and we had 10 people just managing them. At that point we knew there was enough demand to build a system to automate the processes.

“We grew very fast – around 25% week over week – and we were able to beat Foodora in our home market, which was essential. We would have had no chance to convince investors in the next funding round that we can enter foreign markets successfully if we can’t even win our home market. I believe that if Foodora had not entered the Finnish market, we wouldn’t have started doing deliveries in time. In that case, we would have either gone bankrupt or simply grown very slowly and nobody would have wanted to invest in us. So I’m actually very thankful for the ‘Pink F’.

“Today our main competitive advantage might be the effectiveness of our delivery operations. We just opened our service in Seinäjoki, which is a small Finnish city. No other food delivery company has been successful in such small cities.”

Setting priorities

When building a startup, it might be difficult to determine what task to work on next and which activities you should focus on. It is fairly easy for founders to lose focus and end up spending time on things that don’t actually help their startup move forward.

An important thing to keep in mind is that all of the time you put into your startup should be spent on activities that provide real value for your (potential) customers. At this point in your startup’s journey, these activities should be closely related to only two things:

  1. Talking to people who use the product or service

  2. Building and iterating your product

Creating habits

Successful products often have the ability to create a new habit in the user. Nir Eyal, famously known for the best-selling book, Hooked, presents a simple circle to explain how this new habit can be formed. The first is the input, for example, a beep from the phone. This will result in an action, such as checking your Facebook feed. The app must be able to reward your action. In the case of Facebook, this means that a red ball appears on the feed – which shows that someone has remembered you.

It’s also essential that there is variability in the rewards so that the user does not get tired of doing it. In social media, this diversity is built into the product because users are constantly producing content. In the gaming industry, one of the most common mechanisms is “loot boxes”, or bundles of rewards that contain a variety of random prizes. Ultimately, the result is a commitment to the product – where users are actively waiting for a new reward. The product is designed so that it rewards a player with activities that matter to them, ultimately forming an emotional bond.

Creating new habits is not only essential in social media services. Creating new habits and educating consumers is vital for new products and other services as well. If you are for example introducing more sustainable fibers for clothes, you need to educate your consumer about the benefits and create a habit for them to always check what fabrics clothes are made from when they enter a store and choosing the more sustainable version. The same applies to new foods – if you’re introducing something like edible insects, you’ve got to create a lot of new habits in order to get consumers to buy your product on a daily basis.


Perhaps in today's world, especially in the software market, instead of an MVP it might be worthwhile to aim to create what’s been called a "Minimum Lovable Product". In extremely competitive fields, the core functionality of the product alone is not enough, what really succeeds are products that are so tangible and valuable that their users feel ultimately attached to them and want to come back.


Insights from an entrepreneur – Rens Original

When discussing startups, you mostly hear talk about software startups, but there are many globally successful hardware startups as well. Building a hardware startup is different. Jesse Tran, the CEO and co-founder of Rens Original – a company creating the world's first sneaker made from coffee and plastic waste – shares with us his story about developing a hardware startup.

What’s your background?

“After doing my bachelor’s degree I worked at an e-commerce startup called DealDash where I learned a lot. After that, I created my own company Factory Finder – it was a platform for helping small and medium-sized fashion brands find quality and responsible manufacturers. Unfortunately, Factory Finder didn't scale really well. It was more like a consulting business and language barriers and cultural differences between Finland and China made it more difficult. So I moved on and founded Rens Original.”

How did you end up founding Rens Original?

“My co-founder at Rens Original, Son Chu, was one of my employees at Factory Finder and he was really good. So I caught up with Sun and told him about this crazy idea. He and I are both sneakerheads, so we wanted to make something cool and sustainable. Most of the sustainable products in the market, not only shoes, look really boring. And we felt that many people feel the same way. That’s why we wanted to make sustainable shoes where we didn’t promote sustainability as the most important point, but the coolness.”

How did you iterate the first idea?

“We started in fall 2017, and back then it was actually called Revolt. With Revolt, we wanted to create plain canvas shoes from organic cotton on which we could print anything with 3D printing. We did some test printing on cotton shoes, but the silhouette of that kind of a shoe didn’t look good and the material is not very innovative or sustainable because organic cotton is still cotton. So that’s why we did more research and ended up finding a lot of cool materials – and eventually we found one made from coffee waste, by which I mean used coffee grounds.

“We thought that this is a great story and the material has great functionality. It’s anti-bacterial and dries fast. We’re originally from Vietnam, and Vietnam is one of the biggest coffee producers in the world and Finns drink the most coffee in the world. So we thought that this is a great story. There was nobody making footwear out of this material. We were the first ones in the world and still are.

“We loved the original idea, but we didn’t want to fall in love with it – and we have continued to move on to better ideas and concepts. Currently, we’re making new products for Rens. We are not a coffee shoe company, we are an innovative sustainable footwear company.”

How did you create the first versions of your hardware products?

“For the first prototype (when it was still called Revolt), we actually cut the upper part of a Converse shoe and cut the sole from an Adidas shoe. Then we glued them together and printed something on the shoe with a 3D printer. We did it ourselves and we had a friend who is a shoemaker helping us. It was really rough since we did everything manually – this was the first prototype. That original concept didn't take off, but we learned from it and got really confident.

“My suggestion is that you learn as much as possible and try to do the first prototype no matter how ugly it is, you need to create something you can touch. After that first one, you will have many more ideas and something to develop.

“In the software world they can do an MVP and go to market. With hardware, if you launch with an MVP you can destroy your reputation and brand. Hardware products often also have more regulations about quality. You don’t need to have a perfect product, but you need to have a workable product. Many hardware startups learn from the software startup community, create a bad product and go to the market and wonder why no one is buying. So you need to start with creating some physical prototype, but you can’t launch with a bad prototype.”

What can you tell us about using partners?

“After creating the first prototype, we worked with factories to improve the shoe a lot. I had several connections through Factory Finder, but we also searched for several new ones. I went to China to meet with them. There are 12 factories involved in making our current shoe and I’ve met all of them.

“It’s been a long process, but now they are partners and really trustworthy. I understand everything the factories talk about and the terms they use. You need to understand the industry. You need to learn it by yourself step by step. Otherwise, if you randomly go to China, everybody is going to trick you and you’ll end up with a bad product and marketing that is not true, and an unethical and irresponsible business.”

Next section
III. Reaching customers