We read a lot these days about Artificial Intelligence, or AI for short. It’s a buzzword for sure, and there’s also a lot of noise in the press about how AI will be the downfall of civilization, leading us to some dystopia where ethical decisions will be decided by computers. Well, maybe in the future, but we aren’t there yet. There are plenty of more simple, practical uses being developed today.
At Blinkfire Analytics, we have always used AI and it’s sub-categories Machine Learning and Deep Learning (which NVIDIA explains well) as part of our computer vision based algorithms, but it’s also becoming useful to solve real problems our brand customers have in deciding which sponsorships generate the most ROI, and from our rights holder side which categories of content generate the most value and engagement from fans.
For instance, is it better to sponsor a dasher board in a hockey rink, or sponsor a jersey? Better to do a digital sponsorship of the final score or be on the tail fin of a Formula 1 car? In the past, this was often done by hand, taking a group of posts, calculating some values using an excel spreadsheet, and then extrapolating values. But this is a great problem to solve for AI. The key is having an accurate training set from which to train your machine learning algorithms. One thing we have learned is that it is very difficult to take someone else’s training data (or “off the shelf” training data) and apply it to another domain.
We recently announced that we indexed our 100 millionth post. Since we’ve been indexing content and sponsorship data for a few years now, there was no better place to go but our own data set. It’s also something no one else has, so it helps us be the most accurate source for this type of data.
For instance, content featuring players or execs “Holding a jersey” generates engagement averages different from “Happy Birthday” posts (in any language!). And using Blinkfire Analytics AI driven technology can help analysts that work at our customers generate these reports for their content and sponsorship teams automatically.
These are some rudimentary examples that are just the tip of the iceberg. The future is available today. Stay tuned for more.