How to solve a product design question?

How to solve a product design question?

Examples of Product Design Questions-

Designing From Scratch

  • Design a refrigerator for kids
  • Design a music streaming service from scratch
  • How would you design an online presence of a brick and mortar coffee shop?

Design features for an existing product

  • Design a recommendation engine for Youtube?
  • You are a PM at Meta. What would you build for WhatsApp going ahead?

Step 1: Persona Hypothesis

Always start with the people who are going to use your products. It’s never a good idea to jump straight to the features or the solution even if you have one in mind.

Start with thinking about what defines the users -

Demographics - Age, Gender, Married/Single, Families/ Bachelors, Language spoken

Occupation Type -Travelling / Non-Travelling, Business / Service

Income Level - Low/High , Fixed / Variable

Geography - Tier 1 city / Tier 2 city, Country, Areas defined on population density, Areas based on internet access

Device - iOS / Android users , Windows / Mac users, App / Web users

Affinity for the product - Strong need (has a make-shift solution in place but will pay for a more elegant solution) / Moderate need (Needs it sometimes, may or may not pay) / No Need (Doesn’t have any requirement for the products)

Think about the question that is put in front of you. If you are designing a refrigerator for kids - you can divide them as 1-4 years old, 4-10, pre-teens and teenagers as it is easy to think of distinct things these age-groups might want. If you are designing a music streaming service - age might not be the best indicator to define distinct needs but can define the engagement levels (young people listening to more music and being more tech savvy). Income level can help you determine payment model later on. For food delivery services - family and bachelors might have very distinct buying behaviours and needs.

You need to decide which parameters for bucketing the user make sense in the context of the question and then choose them. The list above is quite comprehensive but not complete - so try to create new ones if they make sense to you (for example education levels for ed-tech products).

When making a presentation, it is better to create actual personas - i.e. fictional people with relevant characteristics to highlight the classification system that you have in mind and the different needs they might have,

When giving an interview, it is usually better to talk about the characteristics themselves like age, education, affinity for the product, etc. as it might be easier for the interviewer to follow.

Step 2: Problem Hypothesis

Once you know who you are solving the problem for - the problems easily follow from the personas. The ideal way to identify problems is to first list out all the problems a persona might face irrespective of prioritisation or how big or small they seem to you. Even if you have some idea about which problems you want to solve for- listing all of them helps in the step that follows.

Once you have identified problems that each personas have faced - find the intersection. This broadens the scope of the solution. Even if eventually you choose to solve a niche problem for your highest value customer - doing this exercise gives you context becuase it is difficult to simultaneously think of how 10 problems and 4 user groups are mapped to each other.

This is how you would write it down on a piece of paper or a Google Sheet during an interview


Step 3: Solution Hypothesis


Now that you know what problem you are solving, it’s time to suggest the solutions to the problems. Depending on the problems - you might have multiple single features or buckets of features. Like a recommendation engine is a feature, but an app for song discovery is a bucket of multiple features. Either approach is fine. If you are given the challenge to build a standalone product - an app or a device- you will have to list several individual features and then prioritise. If it’s deciding the roadmap of an existing product - you might choose buckets of features - like payment support for WhatsApp or live audio rooms in YouTube. Just use your judgement to see what fits better.


It also helps to bucket the solutions into some sort of a user journey funnel. I like the AIDAOR funnel. It helps me identify what am I solving for- the end goal wrt to the product.

Note: For product design problems most solutions fall in Onboarding and Retention. That’s fine. Mention that you’re focusing on those parts of the funnel. Or use a more concise funnel. This is what works for me.


Once you have the problems mapped to the personas and solutions you have to prioritise how to build the product. The best way to go about it is find the maximum value you can create with the least amount of effort. Product managers usually use the ICE score to determine which feature is best to develop. ICE scoring has 3 parameters-

  • Interest - This is basically expected demand. A good proxy is the share of users who face this problem. Score it out of 10 with 10 being all users needing a solution.
  • Confidence - How confident are you that a solution to this problem that will get adoption? This question can be answered in many ways. For me-too products, you can see if competitors have solved for similar problems and the success they have add. You can use your own subjective opinion and anecdotal evidence in interviews. You can say that you will interview customers and decide as you don’t know enough about the space. Confidence is part data and part intuition, so it’s quite subjective. Score it out of 10 as you see fit with 10 being the most confidence. Avoid the extremes of 0 and 10 ideally.
  • Effort - This is what investment the problem would require in terms of time and money. Of-course the effort you choose is subjective. People usually aren’t grilled on this but have a rationale when assigning scores. Score it out of 10 with 10 being the easiest problem to solve on the list.
  • ICE score out of 1000 = I*C*E
  • image

Step 4: Usability Hypothesis

Now that you have the solution you are building- you have to delve into the details of the solution. This part is very specific to what ever you have proposed so I will not offer you a format but rather some pointers to keep in mind when going about this phase.

What is usability?

No person just lives with their problems as is. They find ways around them. Before Spotify- people bought music on iTunes or downloaded it using BitTorrent. Some used YouTube to listen. Others stuck to CDs, Cassettes and Vinyl for their higher quality. Others listened to the radio. Spotify made the process of listening to music much easier and cheaper. It added music discovery- a previously under-solved for problem. This solution was more “usable” than the previous solution. Talking about what users are currently doing and what additional “value” or “utility” you can create for them showcases maturity. Usability is the difference in between current and proposed solutions.

How do I make sure my solution is usable?

  • Break it down into an MVP that is easy to test out.
  • Look at existing data that might be available internally or externally. This will be a Q&A with the interviewer. You will have to think of proxies that can predict the need for something in the market. Like the high number of cobblers suggest a need for more durable shoes. High download rates for videos in an app suggest the need for low-bandwidth streaming support. Similarly look for data to suggest usability. Note: The difference between this and what we did earlier to calculate the ICE score is just nuance- we will be looking at more specific metrics- specially if the feature is a broader bucket like we discussed earlier.

Is usability just confined to pre-release?

No. Usability includes considering what value your solution will create, releasing an MVP and then measuring its success. So success metrics are a big part of the usability question.

If you are building a new feature for an existing product then go back to the funnel and depending on the part of the funnel you are solving for - choose relevant metrics for that part of the funnel. You also have to identify which metics will be important for the particular product you are working on and where it is on it’s lifecycle.

If you are building a completely new product then focusing on both onboarding and retention is the key. The metrics mostly remain the same.



Once you have the Personas → Problems → Solutions → Usability then you have a proper product idea developed through the modern product management process. Answering any question in this level of detail and structure will usually lead you to a great solution.