The BBC article “How we fell out of love with voice assistants” by Katherine Latham is no surprise to me. It reminds me of numerous conversations during my Cortana and Alexa days about the importance of user value and trust.

We tracked and tried to optimize many metrics. When we started paying more attention and tried to optimize business-related metrics, experiences started to suffer. When the business cares more about traffic/number of interactions, teams create chatty experiences to improve those metrics, which doesn’t necessarily mean added value to the user.

The article also highlights the need for trust… privacy and reliability. If the promise for natural experience is broken or the data is mishandled, the user will lose trust. Lost trust translates to less usage.

Finally… I love that the article highlights “time saving” as a key decision for adoption/usage. If the result of an assistive experience is to save users time, then that experience will become a habit. When the experience gets in the way of accomplishing tasks, the device will get disconnected. I have always been in favor of utility-oriented positive metrics when tracking the Key Performance Indicators (KPIs) of a natural experience. Metrics such as “time saved”, “favorability towards the experience”, “usefulness to the user”, “number of tasks successfully accomplished”, etc. Of course, we still need the friction-related metrics such as “time to accomplish a task”, “repeated attempts”, “number of corrections” or synthetic/model-driven metrics that some systems use.