How Data helps you master the Berlin Restaurant jungle.

12:30 pm in Berlin Mitte. You just walked out of your trendy start-up office to grab something for lunch. You went to Dolores just yesterday, you already had Pizza and Döner this week, and now you’re faced with the daily dilemma of what to eat. Luckily today it’ll be just you and your smartphone for lunch. As we all know, agreeing on a lunch spot can be tedious once you exceed the optimal headcount of 1. But you like your colleagues of course. They are great.

Even by yourself it can be hard to find a decent spot. Berlin’s blessing and curse is a sheer endless variety of culinary delights, sprinkled on top of a significant amount of tourist traps.

Thank [insert deity or alike here], you know your way around Google Maps Explore. With a couple of filters you can get a list of nearby places with a certain minimum rating, that you haven’t visited and that are open now (ok, they’d be nuts if they weren’t right now). You may be amazed but in Berlin there are always places that fit these criteria. In most of the cases, you’ll be satisfied with what you get. The density of ratings in a city as Berlin creates enough statistical confidence that a 4-star place will not disappoint you. In some cases you may even find a new favourite. One to add to the list of “worth a weekly visit”. And obviously there is a certain chance that the restaurant messes up. Wrong dish, long wait, hair in the soup, it happens.

In all three of these cases, you decide to put the few minutes walk back to the office to good use and share your experience:

In the first case (all fine), you leave a rating confirming the current standing. Why? Because it is statistically more accurate to reflect everyone’s experience, not only those extremely satisfied or dissatisfied. Also, you may leave feedback on what the restaurant could do to be even better. The businesses’ perception of and response to ratings is increasing steadily. You may also ponder on the eternal conundrum: do 3 stars already constitute a “good experience”?

In the second case, leave the according rating of „awesome“, and store the restaurant on one of your lists for future reference. Maybe even add a note.

In the third case go ahead and rate the place “badly”. Important here is to explain why you did so, firstly for the restaurant to reflect and improve based on your feedback, secondly to have other customers understand what they’d potentially be getting into (Too much mayonnaise? Well, maybe someone else loves too much mayonnaise…). Also, don’t destroy the place and assume the worst. Mistakes happen. But your rating is necessary to reflect the restaurant’s “error rate”.

Back at the office you’re still weighing the pros and cons of a “balanced rating approach” (defining a 3-star-rating as “met expectations”) and come to the conclusion that customers would benefit greatly from the resulting additional differentiation, but that you’ll deem it unlikely that you get the mass of reviewers to ever behave that way. Then you briefly feel bad for leaving a 3-star-rating despite being satisfied with the restaurant’s performance. Luckily, you’re distracted by a bunch of other interesting questions: How does Google Maps account for trends? Restaurant owners change, their staff changes, the menu changes. Wouldn’t trends be extremely important? And what about personal taste? Sure, there is the new matching-rating, but does it account for the preferred quantity of mayonnaise? And couldn’t Google Maps solve these tedious joint lunch spot alignment issues? That’d be the Holy Grail!

You calm down by telling yourself that all of this can be solved with enough data points. They managed to solve similar problems in the past and you trust that they will keep doing that in the future. After all, how else will you navigate the Berlin restaurant jungle without getting lost?

This half ironic, half serious local story was written as part of my Connect 2019 application. It is supposed to give credit to features that I love about Google Maps, while also raising a couple of interesting questions about the service and how it could further develop. I hope you enjoyed the read.