What is hypothesis testing?

How can a hypothesis be validated, what is a lean startup methodology and why "Jobs to be done" is the best-known approach to create a perfect PMF.
In this article, we would love to share with you our vision of what the product hypothesis is. We will also look into the fundamentals of the User Story and Jobs to be Done lean startup methods and shed light on what Product-Market Fit is.
These days being fast and effective is paramount. Lots of aspiring entrepreneurs are building startups in order to create useful and profitable products. That’s why nowadays it is vital to validate the idea of your product as soon as possible. You must ensure that the idea behind your product will help it grow into a successful project. Numerous startups fail at the start. To survive in this reality you should launch a product fast, analyze how it works, quality check it with users, and decide whether to bury it or continue developing your product in the way you are positioning it in the market.

You can start with one idea and end up with the need of a pivot. What is a Pivot? A pivot represents a change in a startup's concept — its business model, its product, its field of activity, or targeting. It means you are flexible and able to adjust your project and be ready to give up on your initial ideas if you realize that change will lead to growth and your product will be in demand.
Lean Startup

According to Eric Ries (the author of the Lean Startup methodology), there are two main reasons why startups fail:

- Preoccupation with traditional business calculations, plans, strategy building, and comprehensive market research. But the problem is that in the complete uncertainty, in which startups operate, these classic management techniques do not work.

- The second reason may seem to be the exact opposite of the first one — seeing that traditional management approaches do not work, entrepreneurs abandon all of the management tools. They let things run their course and are guided by the "just do it" principle.

And here comes the Lean Startup methodology, which combines everything that is good in traditional management and the new approaches.
startup robot
A startup operates in an environment of extreme uncertainty, and this needs to be taken into account when launching it. At the initial stage of development, a startup needs to remain flexible in order to learn from mistakes and test the founders' hypotheses as quickly as possible, which means avoiding great expenses and costs. This is the approach at the heart of the Lean method, which aims to help entrepreneurs improve a startup's chances of success.

What does a lean startup consist of?

There are three foundations:
● Business model identification — you have to create a business model here. How it is going to be functional, how you are going to monetize it, how it is going to grow. Calculate unit economics.
● Customer development — it is crucial to start it as soon as possible. You need to discuss with your potential users what you are planning to develop for their service, how it is going to impact their experience. You need to end up with checked, educated guesses from your potential audience.
● Agile development — as soon as you are through with the two points above, you should start developing your product using a flexible methodology. It is important to start building something that will NOT take ages and will not necessarily be perfect, but something with real-life application, something you will be able to start selling to your users.

You should not treat your product as something you can sell forever once you have developed it. Product improvement is an endless process. As soon as you release the first version of your product you will receive tons of new inputs from users. As practice shows, you and your team will not be able to digest those inputs. An ideal option is to start with a smaller audience and grow your users 10 percent a week.

The “build-measure-learn” cycle is the primary principle. The Lean Startup method initially aims at producing a minimum workable product (MVP) and continually soliciting feedback from customers in order to understand their real needs, rather than creating a perfect product. Unless the initial assumptions are borne out in practice, they are wrong and a pivot must be made — a radical change in direction. This approach rules out the risk of a startup creating a product that no one wants.
Before devising a plan for the development of an MVP, first, you have to calculate customer acquisition costs, i.e. how much you will have to pay to attract users to your product. You can have these numbers by benchmarking against competitors. Then you need to deal with conversion assumption, which is a very risky metric. You cannot rely on competitors because then there will be a lot of dependencies. You can miss your unit economics calculation here. So what can be done at this stage to check your anticipated customer conversion? One of the possible ways is to use a “Fake Door”. A simple example: you can create a landing page of a product you are going to sell and buy some targeted ads using Facebook or Google. Your landing page should describe your product in the best way possible, explain application benefits to your potential customers, and have a “Purchase” or “Subscribe” button. The next stage should contain a “thank you” page to appreciate users for paying attention and promise to provide early access as soon as you launch your MVP.

Using those real data you can measure your possible conversion and even receive emails of your potential users to check their motivation and expectations about the product. It is a well-known broadly used practice that professional product managers currently employ.

As a result of this check, you will have numbers. They can either be promising, and your unit economics is going to work, or the results are going to be disappointing, and even at this stage with no MVP ready it is recommended that you critically rethink your idea and maybe even make a pivot. This is the ideal scenario of the Lean Startup practice, but the reality is usually different. Anyway, it is good to remember the roots and check ideas while consuming the least possible resources.
startup robot
Above is the diagram showing that explosive growth is observed shortly after you launch a product — when it is first marketed. Later, as you constantly improve your product that has been launched and is working, your ROI (Return on Investment) decreases. You enjoy the most significant effect at the very beginning, and this effect wanes the more you work on building up the value you are offering to your customers.

That is why it is so important to constantly monitor the market and launch the features/products that your users are starving for. But first, you need to formulate the hypothesis of what your customers are expecting and be able to spread it among your teams and people responsible for your product growth.
Let’s review contemporary approaches to the formulation of hypotheses:
User story — the kernel of this approach to the creation of a hypothesis is the focus on actions that users are going to undertake. According to the User Story technique, we always imagine a user and describe them from their point of view. You need to describe a user with social demographic characteristics. For example, a 25-year-old woman wants to buy a pair of shoes quickly, using her smartphone, during a coffee break. What do we know about her? Official parlance won’t catch her attention; let’s try everyday spoken language style which fits her youthful nature. We can also add the location, where she lives, as well as use research to calculate the number of potential users in the area and try to attract them all. When we apply the user stories approach to hypothesis formulation it is good to use them as jobs for the development team and guide everyone based on what your product is going to do, but we are missing the main goal in this approach. We don’t pay attention to the reason why this young woman wants to wear those shoes — to a party to capture attention, for example.
startup robot
So since the User Story approach is missing important goals, it is better to move a bit further and use the Jobs to be Done technique.

The most famous Jobs to be Done use case is the “Milkshake” case. (It is very popular on the net so you can google it easily). Here is a short story about it. The main goal was to increase sales of milkshakes. Having researched buyers’ feedback, manufacturers met all of their buyers’ expectations (more sugar, bigger cups etc), but this didn’t lead to any increase in sales. The main goal was to find out why users intended to buy a milkshake rather than how they appreciated its taste. The shift in the focus from the person to the reason and obstacles encountered whenever and for whichever reason a user needs something is the key difference. Users don’t buy a product, they buy a difference made to their lives or a better version of themselves. The milkshake secret turned out to be that it wasn't about buyers' preferences in taste or cups. Milkshakes helped them not to get hungry in traffic on the way to work. Its direct competitors in this "job" were not other chains’ shakes, but bananas, coffee and snack bars. Then cocktails were made thicker (to make them more nourishing) and the option of being served through a window right in the street, without getting out of the car, was added. Sales increased.

A user “hires” your product to address their challenges and meet their needs. It is with the ultimate result in mind that you must sell your product to users. The Jobs to be Done methodology implies that you can create a hypothesis based on a user’s situation and their current objective. Whenever a user is in search of a solution to a challenge, they wish to do something to achieve their goal. However, the use of your product, whether an item or a service, will end as soon as its value to your customer ends, even if your product has been effectively developed and remains operational.
The Jobs to be Done technique also envisages certain job limitations. Jobs should be discontinued unless the next step adds value to the product. Services that accompany a user everywhere are particularly popular now. The Chinese market (WeChat) is a perfect example. Developers would love their customers to start using their apps at birth, live with them and eventually die while using them. So when is it the right time for you to stop wishing to add another new feature to your product?

There are the following indicators for this:

  • There are other, better solutions with the functionality that you have in mind, and you won’t want to compete with them.
  • The next step you wish to take can be implemented in many alternative ways. (For example, you want to send some sort of a notice to a user at certain hours and think about adding an alarm to your app for that. You shouldn't do that unless you want to lose the user).
  • Your next step will bring about a dramatic change in your target audience.

While using this technique it is very important to continuously work on your Product Market Fit and improve it. Your PMF is temporary, as the user only hires your product for a limited period of time while they have a need for it. It is great if you deliver what the market wants, and your product sells. But you cannot relax — you need to keep looking for something new. Cell phones are one of the most obvious examples. There used to be lots of successful players in the button phone market. They were continuously improving their products, researching the market, and releasing new and better models. But then smartphones came along, disregarding the fact that Product Market Fit had been achieved. You should always look at what users want and never cease customer development, never stop checking what your product can become tomorrow, rather than which of its current features need to be improved.
Here are the important criteria to determine PMF:
  • the product meets market needs,
  • the product is in demand,
  • 40% of customers say they would be upset if they were unable to use your product,
  • NPS score is over 80% (whether users are willing to recommend your product),
  • ROI from marketing efforts is positive (whether you receive profit/revenue from a user once you've invested in attracting them).
It is critical that you track your Product-Market Fit based on the above indicators in order for you not to miss the fine line between the need to further invest in your current product, resources and the need to make changes, all the way up to a pivot. An integral part of your monitoring is unit economics. If your project does not work the way you expect, perhaps it will be good for addressing your customers’ alternative, unobvious issues.

The most outstanding Pivots:

YouTube — it started as a dating service.
Slack — it used to be an internal video game chat service that failed.
Instagram — it was originally called Burbn and offered too many features (check-ins, scheduling, and making pictures from meetings with friends). It turned out that users did not care much about the check-ins component, but they were glad to use the fast picture-sharing functionality. Its creators benefited from this input to launch a startup that Facebook bought for $1 billion in 2012.
startup robot
So how do you know whether the development path you chose should be altered, or everything is right and you should continue building on the current product concept? Perhaps “it’s just not good enough” and “let's keep on doing that”. Should we work with users, teach them and explain that it is the current version that meets their needs?

There is no obvious solution. You have to work with users, analyze feedback, and handle customer development. You can communicate with users and get new insights. User contacts that you have collected while testing Fake Door can help you with this.
Keep on training. Using the Jobs to be Done technique, think of situations in which some of your users’ need is effectively addressed by your product.

Criteria of a good Jobs to be Done hypothesis:
  • Describes a user’s motivation,
  • Does not mention the solution,
  • Does not describe the action,
  • Does not describe the challenge,
  • It is easy for you to explain it.

Below are examples of good and poor Jobs to be Done hypotheses.
This approach will empower you to develop additional functionality, understand the situation, in which to offer it, and how to present it so that customers pay for it. This will help benefit from important user information and focus on creating value for various members of your team. For example, you are working on a project for someone who wishes to lose weight.
startup robot
Jobs to be Done — What problem, when, and how should we address to create value for our user? Our client wants to not only lose weight, but also to appeal to the opposite sex, to be liked and loved.

User story — What does a user of a certain gender, age, in a certain location do in this situation in order to deal with their problem? A 25–35 year old man in a big city with an above-average income buys a gym membership and starts working out with a trainer.


By combining these two approaches to creating hypotheses we get insight for an app, which you can use to plan your diet, get workout plans from a trainer, track progress, but also chat with fellows who share the same interests, get approval and support, read tips on how to choose new trendy clothes and train communication skills to talk to the opposite sex.

It is also very important that you not lag behind the market with your product, and the development team is key to this. How fast can you show your project to the market before you are beaten by your competitors, how stable and pleasant to use will your product be? We can help you with the Free project vision for your startup, so you get your optimal user journey, software architecture recommendations, and a structured step-by-step plan with a time frame for the realization.
We've also prepared a series of useful materials for you: