Optimizing Features For Data Products
Before big data was a known term, Dacheng Zhao has pioneered innovative data products that make an impact at scale.
Google Product Lead on Data Products (Part 3)
One of the toughest challenges for product managers is deciding what to do with specific features. In other words, you’re constantly wrestling with what to keep vs. delete. Most importantly, you need to figure out when it’s best to come out with something new. Data products are no different. Unfortunately, I don’t have an easy way of explaining how to come up with the right answer.
However, it’s useful to compare the “ideal approach” to the “practical approach” for optimizing features in data products. Ultimately, the right answer comes from striving for the “ideal” and then applying the “practical” to build your feature set.
The “Ideal” Approach.
As product managers, you’re probably familiar with defining a “north star” for your product. However, this exercise goes much further than a specific product solution or feature. Instead, you need to figure out the most fundamental problem that you are looking to solve. Generally speaking, this is a problem statement much more than it is a product statement. For example, it can be “we exist to expedite communication” or something that speaks to the soul of your product.
The advantage of making decisions based on your north star is that every feature you pick is ultimately connected to your mission. For example, you’re able to determine if individual features or data products are aligning with your core values. Ultimately, these are the features that you should keep. Conversely, any feature that is underperforming or is not in alignment with your mission is perfect for replacement.
The “Practical” Approach.
As great as it is to go through the “ideal approach,” the real world forces us to operate a bit differently. Simply put, the practical approach takes much longer than the ideal approach indicates. In other words, you’re slowly chipping away at getting your features to align with your core values.
The biggest reason for this is that your initial mission statement will look completely different within six months of starting out. Ultimately, you have to constantly refine and adapt your focus to match the realities of how your data products are being used. For example, we started with nine separate products when I joined Google – and my goal has been to consolidate all of them into one cohesive solution. As you can imagine, going from nine to one doesn’t happen overnight. With this, you need to be prepared to methodically iterate and focus on small wins to stay on target.
In the end, the biggest challenge you’ll face is allocating resources to develop your core product. Said differently, we’re all striving to make our core product satisfy the tenets of our mission. That said, there can be friction about taking resources away from other products to push toward something bigger. Ultimately, this journey takes time and the important thing is to constantly keep the mission in mind based on your north star. However, just remember that there’s never a bad time to pivot!