Fantasy Football and Product
What fantasy football has taught me about data and how to build products.
I began this football season starting a new role with a new company and time to do my research on fantasy football was null. So I did what every PM would (or should) do in their day to day work life is experiment. The experiment that I’m running has the hypothesis that “If I let the projections tell me who to play and who to sit without any other influences I will win”
Team on Autopilot
Yahoo is where my league is hosted which utilizes AccuScore a data analysis product/algorithm which does predictive analysis of players to give the user their predictions on scoring for a game and season. They also offer the ‘Auto-Draft’ feature which takes in these predictions and drafts your team for you. So I put my team on Auto-Pilot and let Yahoo take the wheel. I awoke the next morning to a draft score of A- and a prediction that I would smoke the competition.
Fast forward to present day, I’m 0-6 by purely letting the numbers choose who to play for me. No matter what I think of the player and the team they are playing, if their prediction is lower than another. They ride the pine pony and sit the week out of the lineup. My team has been in the bottom half of points scored and no injuries to any players!
Now, how does this get to building products and using data? Great question. We have all been told just look at the data and it will guide you. But how? Just looking at the conversion number of a particular user flow such as points projected for a particular player doesn’t paint the full picture.
Painting the full picture with bob
We have all been a victim of being siloed into thinking just because the conversion percentage has increased we have won. I too have sung the praises of an experiment that has drastically increased conversion and hitting the ‘give this too all users’ button then later finding out I’ve pissed a good deal of users off or introduced confusion.
Just like in fantasy football there is more than just the final conversion % or projection number. You have your leading indicators such as weather, is it home-away, do they have any injuries, is the team they are playing hot etc. Your conversion flows have these same pieces at the top of the funnel all along with the qualitative data on how your users.
Keep these leading indicators in mind when looking at your final outcomes as focusing just on that final number might be like picking sports ball athletes based on some algorithm. We have to be completely informed when making a decision to push to prod with enough certainty that we aren’t providing our users a terrible experience.
What am I going to do next? I’m riding this rollercoaster to the end since I’ve placed myself dead last without a chance of returning…
Andrew Warner is our first guest writer here at Dignified Product. His background includes research in the Product of the Human Body, Genes, and Human Health. In his time, he’s accomplished a lot, but the research he’s putting out about the projection and statistics of this Fantasy Football product was so compelling that we begged him to let us publish it. You can find him babbling about his Win/Loss ratio along High Street on the weekends.