Ratings and Feedback Loops

Once upon a time, you encountered something new like a restaurant, a product, or a service and if you liked it, you told people about it. If you didn’t it didn’t spread. That all changed with the internet.

The internet is filled with two things: data and algorithms. It isn’t much more than that at the core. The data is text, video, sound and that covers things like ratings. The algorithms choose up what to serve us. Yet, in the quest for finding what is best, we turned to ratings, commonly done out of five stars. We can compute average ratings with a direct number, comparable against other products. Compared to that previous era where you heard about something through a friend, you can now search all types of that same product and draw a numeric comparison. This seems like a great thing…

… on the surface.

As a single product, or at least a narrow subset, becomes the dominant winners a feedback loop is created, and the other competition soon dies away. That doesn’t seem particularly bad either, after all, if the best products are winning, things are getting better, right?

Take Netflix though, their algorithm needs you to rate content so that it can serve similar content to other people that they know they will like it. The problem is as those top winners emerge, and everyone starts to be shown the same content, it seems like there is nothing new on the platform at all. That leads to cancellations. So does someone seeing ratings that are too low and choosing not to watch even after finding something new. They may have enjoyed the movie they skipped just fine, but since others didn’t, and rated as such, they lost the opportunity. Netflix may lose them as a customer as a result.

As algorithms take over, we are growing in feedback loops. Winners win because they are winners. Losers lose because they are losers. Sometimes those make sense, sometimes they don’t at all. Finding a way to break these loops is going to be a key to being able to make something happen for yourself in the future.

Imagine you’re a job seeker and data is available about you prior to hire, your sales numbers, your promotions, your credentials and more. Starting with a great resume, you get a great opportunity, and that snowballs to the next one and so on. If you didn’t start that way, you will perhaps snowball the other way, get a lesser opportunity which doesn’t bare fruit, making the next opportunity worse and so on. It happens today, but with AI starting to do recruiting, it will get exacerbated. I’m not telling you this to make you despair, but to face the reality that the world is changed because of the availability of data.

Differentiation and creating new categories to be judged by are going to be the only tool you have to break into areas where there is an established brand and winners. I don’t have a foolproof plan for all scenarios that I can share yet, but I have thought quite a bit about feedback loops and reading them may give you some ideas. You can read about them on the following posts:

  1. An interesting thing about feedback
  2. Binging with Babish and Feedback loops
  3. Is it worth doing if it doesn’t get likes?
  4. Tomorrowland
  5. Everywhere I look I see fear
  6. Standup Comedy
  7. Iterative Design
  8. Data creating leaders that follow

Hopefully, there is perspective found in these. The following conclusions can at least be drawn:

  1. Have some convictions at a philosophical level about the work you do. Don’t be a leader that follows.
  2. Continue to hone your craft and iterate the work you do.
  3. Ask the audience you serve how you can make it better.

If you do these things, you can create a product or service that fits an audience, becomes the top rated item and becomes a winner that wins, even in a space where there was a previous winner.