Like many, I rely on various restaurant review sites before making a decision to try a new eatery. However, every major site has a huge flaw. They aggregate ratings throughout the history of the restaurant.
This would be fine if restaurants were run by robots and immune to changes in staff and quality over time. It also doesn’t take into account that as a culture, our opinions change. Trends die, better options come to town, what was universally great years ago may be average or worse today.
So the place above, seems like a reasonable choice with an 84% approval rating. However, if you dig into reviews, there are many people saying the restaurant has gone down hill over the past year or so.
But you can’t gather that quickly, especially from your phone. So how can we fix this?
With very very basic data visualization. Here’s an example of what’s possible:
This has a lot of implications.
- As consumers, we have solved the “restaurant going downhill” problem I outlined above.
- Anyone can spot someone trying to cheat the system (ie – getting all your friends to post positive reviews of your business at one time)
- Business owners can fairly easily judge the customer feedback on various promotions (the Groupon note pointed out in the screen shot)
- Owners have additional external pressure to turn around a struggling business, and their potential customers can actually see when changes have had a positive impact, rather than letting past mistakes continually haunt.
Taking it a tiny step further – I always get annoyed by the list of restaurant details with simply “yes/no” type items. Seemingly everything outside of the snootiest joints on Yelp say “casual” attire, even if you’re expected to show up with nice pants and a tie. We could use, again, very simple visualizations to help people.
As you can tell from my examples, data visualization is not a strength of mine. But a bigger takeaway is that this can be done very simply and provide a lot of value for the consumer. I’d love to see review sites attempt to go this route.