More Analytics Isn’t Always Better

LogiAnalytics sells a self-service business intelligence solution, and they’d probably like to see Fitbit using it. Their sales pitch comes in the form of a blog post ("Why Fitbit Product Managers should be focused on Analytics") critiquing Fitbit for not giving its users a more full-featured, self-service analytics dashboard as part of the fitness product.

The post starts by lauding Fitbit as an innovator, but quickly pivots to suggest Fitbit hasn’t “kept pace” in the realm of self-service data capabilities. Fitbit’s simple dashboard "may have been acceptable some years ago but not anymore."

The evidence: someone pulled data out of Fitbit and visualized it in a different dashboard tool. Never mind that the someone is a demonstration project by another dashboard tool. The LogiAnalytics author concludes that this is proof that "Fitbit doesn’t meet user needs anymore and provides a workaround for customers to export data to another platform for improved self-service analysis.” The post goes on to paint a doomsday scenario for Fitbit: "Instead of instantiating itself into the daily lives of users, it is separating itself out and losing the user by becoming a device that measures steps (and other things)."

It’s quite a leap based on some questionable assumptions. My guess is that more self-service BI in a consumer device probably isn’t what Fitbit product managers should be concerned with. Here’s an alternative set of lessons for companies looking to integrate analytics into their products: 

  1. The plural of anecdote isn’t data. When you find one example of an analysis approach by your customers, don’t feel compelled to add it to your next product release. A better approach: find out what is the pain that is motivating the customer, then talk to other customers to see whether that pain is common.
  2. Know your user segments. I’d love to hear from Fitbit about how they’ve thought about user segments and meeting analytical needs. Given their success, I expect they’ve done their research. At Juice we recently segmented our Juicebox app users into five analytical user segments: Explorers, Light Explorers, Number Checkers, Table Downloaders, and Freshman.
  3. Apps, not busy dashboards. Fitbit has clearly emphasized their mobile app as the primary mechanism for interacting with fitness data. The app interface shows how they’ve thought about making data useful. The data is delivered in service to different user needs: weight loss, competition with friends, step activity, sleep tracking. I wrote about the post-dashboard world here.
  4. Don’t make consumers feel like they are at work. The idea of baking a self-service BI solution into Fitbit feels like a mismatch in expectations. As a consumer, I want simple, easy, fun, and direct. I wouldn't use those words for any self-service business intelligence solutions, regardless of how important they might be in a particular business context.
  5. Hackers want hacking tools. Fitbit created a well-documented API that gives technically-savvy users the ability to pull their data out, combine it with other data sources, and present it any way they like. That’s the ultimate flexibility — not a dashboard with extra nobs or trend lines.

To be fair, Logi offers some legitimate points:

  • "Through our extensive work in BI and analytics, Logi has found that when users get access to data, over time they begin to want access to more data and want to be able to do more with it."
  • "...for smart companies that seek to keep an ongoing relationship with the user know the philosophy of 'don’t lose the user' is critical to maintaining hardware revenue..."

To give LogiAnalytics some credit, they do point out that once users get access to data, the users will always ask more and better questions. Another important message: analytics on top of a device is a good way to enhance and solidify customer relationships. But choosing the right way to deliver that data is the challenge, one that Fitbit seems to have taken head-on.