Anyone working as a product analyst or product manager has a crucial role in ensuring product success. A successful product addresses one or many problems, helping customers overcome obstacles or solving a problem they struggle with.
That’s why one of the initial steps in the product life cycle is defining the problem the product will solve. But how do we define a problem properly? — By asking the right questions.
Analyzing without asking the right questions wastes time, although it’s a common occurrence for most product analysts and managers, myself included. In my 3+ years as a Product Analyst at Taboola, I’ve developed a better understanding of asking the right questions and defining a problem statement that hits the mark every time.
In this post, I’ll share seven questions that’ll improve your approach to defining problems, whether you’re a product analyst or manager.
Start by collaborating with the right people
Albert Einstein once said, “If I had an hour to solve a problem, I’d spend 55 minutes thinking about the problem and 5 minutes thinking about solutions.” He understood that defining it is most of the challenge for problem-solving, no matter the domain.
That’s why here at Taboola, product analysts and product managers work closely to define them. The two roles may have different responsibilities and may work in different business units, but they collaborate closely when it comes to problem identification in any project.
To make sure our problem definition phase is 100% finished before moving forward in the project, I created a question toolkit. It uses very specific questions to get to the heart of the obstacle the project aims to solve.
Use these 7 questions to better define problems
Let’s go over the seven questions using a classic pain point we face at Taboola all the time: the sheer volume of data our platform generates. The Taboola platform creates millions of content recommendations daily based on a machine learning algorithm that fits each user’s interests. We gather massive amounts of data per user to train the model to predict a user’s click.
Question 1: What’s the goal of the analysis?
The first pain point of handling all this data is the cost. It can be expensive to store every piece of data, so we need to know if the value of the data is higher than the storage costs. By analyzing user segments, we’ll discover there’s no value in storing more than a certain number of segments per user. If that’s true, how can we discover that threshold number?
In this scenario, the goal of the analysis is to estimate the impact of limiting the number of segments per user on Taboola revenues (which take storage costs into consideration.) The lower the storage costs, the higher the revenues.
Question 2: What decision do we take based on the analysis?
Answering this question helps focus on key insights. It is even more important to understand the potential decision-making because it’s hard to decide with no significant analysis results.
In this scenario, a possible decision is setting a threshold of maximum data Taboola stores per user.
Question 3: What metric can support the chosen decision?
Your decision needs supporting metrics, so you need to pick the right ones that support your chosen goal.
In our scenario, you could conduct A/B tests and compare each group’s percentage revenue per user.
Question 4: What are the optional values and benefits of the analysis?
If your questions have offered measurable values that are clear to you and your product team at this stage, congratulations! That’s pretty rare, in my experience. Over the years, I’ve found it hard to expect every analysis to deliver clear and measurable results that show how to proceed going forward. Sometimes the measured value is only one part of the overall project’s value and may be ignored by product managers or other business leaders.
In our scenario, we might find that an optional value of the analysis is that we can reduce data storage costs by $10,000/month, making it likely that it’ll be included in future project planning. However, if we only found it reduced costs by $3,000/month, that data might be ignored.
Question 5: Who are the stakeholders?
Knowing how the stakeholders are for a problem is essential as it’ll dictate the most relevant metrics and can inform how the project is defined later. For example, customers, business leaders, and particular roles in particular business units may be stakeholders for one problem, but only customers may be relevant for another.
Question 6: How accurate should the results be?
Sometimes, complete accuracy isn’t achievable, so you’ve got to decide how to proceed based on the accuracy you have. Do you need high accuracy, or can you proceed with low accuracy and still get the same results or answers?
In our scenario, you may not need to analyze all the stored segments. Analysis of a portion of it might give you a representative answer, and your conclusions will probably be the same.
Question 7: Is this a one-time analysis, or will you do it regularly?
A Taboola dashboard can be a great way to support decision-making, but you don’t do the same analysis every time, right? There are plenty of one-off decisions and analyses you do, like if you’re validating a new product or ideating a new feature. If you choose to make a dashboard, be sure to include the required time you’ll need for it in your planning.
In our scenario, you decide not to create a dashboard in Phase 1 of the project, but you know you’d like to monitor the savings more regularly later on. So you set up an A/B test in Phase 1 and track the results over time in a dashboard.
Generating actionable insights throughout the product life cycle is key for product analysts and managers. Here at Taboola, product analysts and managers often collaborate early to hone in on the right problems that will have the most impact for stakeholders.
This post outlined the seven questions you can ask to help your product team focus on those valuable problems to create products that lead to more success.