Product Phases and Feedback Loops

Launching a successful product is hard. Lots of different variables have to all come together — the right problem to solve for the user, the right market timing, the right execution. A lot of it has to do with luck, but there are things you can do to get luck on your side and find the right data and feedback loops to help you make more informed decisions.

There are lots of tools, techniques and methods to gather data and to get feedback from your users, your team, and your leadership. What you do and when depends on the questions you’re trying to answer. The questions you should ask depend on where you are in the product cycle.

To know what questions to ask, it helps me to think about the product cycle in phases — Initiate, Build, Launch.

Initiate

In this phase you’re trying to understand if your company or your investors should invest money and resources into solving a particular problem right now.

The feedback loops here are from market analysis, exploratory research and your leadership or investors.

Here are some questions to focus on:

• What is the unique opportunity and how does it fit into the overall vision?
• What is the problem you want to solve and why?
• How does it benefit your users or prospective users?
• Why is it important for the product now?
• What are a few different ways you could solve the problem?
• What is your vision of success and how will you measure it?
• How will you explore your ideas fast and cheap?

Why it works

1. It helps you focus on the user and the problem you’re solving, instead of being inwardly focused.
2. It forces you to have goals and multiple ways of thinking how to get there, instead of just one direction.
3. It encourages fast and cheap validation, rather than long expensive bets.

Build

At this point in the cycle you’re trying to design and build the most useful and usable product for your users. You will do a lot of user research, design, prototyping and coding. The research in this phase should be fast and mostly qualitative.

Qualitative research won’t answer if your product will have product market fit or if a feature will be successful over time, but you can’t do quantitative data driven research yet because you don’t have an MVP. Qualitative research will clue you in to what might be a good idea, what might work, what probably won’t work and if anything you’re designing could be designed better.

In this phase most of the feedback loops should happen with your users and within your core team.

Here are some methods you can use in this phase, and the questions you should be asking:

• Exploratory Research
  What are the user behaviors and what needs are unmet?

Talk to your users, listen for hidden problems, observe workarounds and pain points.

• Ideate
  What are all the different ways you might solve the problem?

Explore a breadth of ideas, think outside the norms, push your imagination.

• Storyboard
  Can users imagine themselves in the context of your story?

Create a series of simple visual storyboards of your most promising ideas and walk your users through them.

• Design
  How can you solve your users’ needs in the simplest and most intuitive way?

Put pen to paper, open up Sketch and create countless iterations.

• Prototype
  What are the most essential flows to focus on for the MVP?

Start working through multiple use-cases and understanding the scope and complexity of the user experience.

• User Test
  Can users quickly and easily accomplish given tasks and do they perceive these tasks to be valuable?

Run users through the most relevant use-cases and observe what happens.

• Build
  What’s the fastest way to build an MVP?

Start building your product as soon as possible without too much throw-away code – as soon as Product somewhat solidifies a general direction.

• Test over time
  Does the product seem to add continuous value over time?

Use an almost finished MVP to do a diary study with your users.

Why it works

1. You’re seeing the actual behaviors of your users and building empathy as you’re building your product.
2. You’re exploring solution spaces by tinkering, building, failing, learning and iterating.
3. Each level of fidelity and each interaction with your users provides invaluable feedback towards your final output.

Launch

You’re still not quite ready to launch, but now you have a working MVP. This phase should actually be called Validate & Launch, but just Launch sounds cleaner.

The beauty of having an MVP is that you can now start experimenting at a statistically significant scale and use data to validate your investment. Before fully launching your product or feature, you need to understand if it adds enough value to your users over time and if it contributes to your overall KPIs.

The tried and true method is to launch a small, but statistically significant experiment to a small segment of your users. If you are a startup and don’t have any users yet, you can do a soft launch.

Depending on the initial results, at this point you might go back a few steps and keep iterating if the results aren’t promising, or you can refine and optimize if the results signal success.

Once you have the data and your assessment of it, it’s a great time to review your product and the results with the leadership and cross-functional teams. If you’re going to launch what you’ve built, there are usually lots of moving parts (especially in bigger organizations) and it’s important to clearly communicate between teams and gain alignment.

In this phase the feedback loops are from quantitative data and from your leadership and cross-functional teams.

Methods and questions to focus on in this phase are:

Statistically significant experiment with a control

• Is the product or feature meeting user needs and adding value over time?
• Is the product or feature likely to move the KPIs in the right direction?
• If it’s a feature, is it breaking anything else (and is that okay)?

A/B Tests

• Which variant of the feature is the most optimized to improve KPIs?

Why it works

1. You’re using real data with real users to validate results.
2. You have an MVP and you can use real data to learn and iterate or refine and optimize.
3. Cross-functional alignment and great communication help orchestrate a successful launch.

The product cycle is not linear, it’s iterative. It means that depending on what you are learning you’ll jump back and forth between phases and methods. What I think does remain fairly constant are the types of things you’re trying to learn in each of the phases, and therefore the feedback loops in each respective phase.