Ok, first off that title…if you omit the cognitive bit and it is the business strategy of every new product I have worked on. I have been on existing product teams where the beauty of this strategy equates to extreme Agile.
I have never seen it happen with initial product release of a cognitive product in an existing business. Let’s take the title apart and see why.
“Cheap, Fast and Sexy!” When leadership says that this is their approach to building new products, I worry. I have never seen the business staff a product team to met their deadline. Releases are pushed back or scope is reduced to get something into the market on deadline.
I have been on a product team that was forced to release an incomplete “me too” product over a year late. The product owners thought they knew exactly what the market needed and loved their “new” idea and decided exactly how many developers and designs it would take to build by their deadline. They repeatedly ignored buyer and user feedback. I remember standing in a conference room full of buyers and hearing “We don’t want your fucking data, we want insights [you can help us get] from our own data” yet our scope and direction did not change.
MVP is a slippery concept, it is warped from Minimally Viable Product to Minimally Valuable Product by the pressure of possible sales. This creates friction in the product team and results in churn and delays.
I have been on teams that just couldn’t get out of their own way to release an MVP. The product owner had turned over the keys to the castle, the product team was responsible for defining MVP and prioritizing the work. The team itself churned and argued about what was MVP. Design leadership directed us to push for building a huge vision for the technical product that anyone could use without help.
Now let’s add Cognitive technologies onto the complexity of building an initial MVP. The cognitive MVP’s does not exist! None of the cognitive products I have worked on released a product that would be considered minimal. Rightly so, while this set of machine learning and natural language technologies are incredibly powerful at finding and generating insights from unstructured language they require more time and work then MVP products.
But development leadership has decided that this heavy lift is done only by the data scientist and machine learning engineers in their silo. As a result, business leadership thinks that the product team should be staffed and build an MVP that meets their deadline. Designing products that use cognitive technologies often have to be designed to build trust, transparency and to help user correct their epistemological mismatch. That again is more work than an MVP!
In my experience, I have never been on or seen a cognitive product team release an MVP that met a business deadline.
So, how do we fix all of these compounding problems? Glad you asked! In a future post we will identify methods and team behaviors that can minimize these problems!