Agility in Product Management: Why 100,000 Interviews? 🎯

Many believe that we need vast amounts of data to create a product, but this isn’t always the case. Forget the 100,000 interviews and that 95% statistical confidence. A few well-targeted conversations can be more than enough.

In practice, resources are limited. Do you really need a mountain of data for every decision? A more agile approach says no. 🏃‍♂️ For instance, in an initial discovery process, 5–10 interviews can reveal patterns and trends that provide a deep understanding of user needs and problems. To give another example, in user testing stages, it’s observed that with 5 users you can detect between 75% and 85% of interface problems (likely the most pressing ones).

As always, it’s about maintaining an agile focus. Instead of pouring resources into exhaustive statistical research all at once, you can invest them in generating a greater number of iterations/experiments and thus refine your product. To me, this is a far more pragmatic way to work.

From my perspective, this is even more applicable in the case of B2B models. Often, we have limited access to prospects and it’s challenging to gather a large sample of candidates. That’s why the focus should be on executing a high-quality discovery process. Let’s not try to reach all those potential clients, but instead extract a lot of valuable information from the touchpoints with them (being careful not to become a feature builder).

Companies at the forefront of innovation have it clear: no more paralysis by analysis — action, learning, and adjustments as new insights are obtained. This approach allows for quick pivoting and finding that perfect fit between product and market. 🔄

The next time you’re thinking about how to collect data for your product, consider whether you really need a large volume of information, or if you can take a more efficient shortcut. Agility in product development is a necessity to stay relevant and competitive in today’s market.

Have you tried quick experimentation techniques in your projects? Or do you believe it’s important to conduct thorough research according to statistical purity?

Share your experiences! 👇