Solving the Duplicate Problem

StreetCred - July 6, 2020

We started StreetCred to solve the unsolved challenges we saw in the map data world. Recently we turned our attention to the headache-inducing issue of duplicates, a perennial problem for any user-generated mapping tool. After noticing how easy it was for users to map existing places, we launched a product feature that reduced duplicate submissions by 50%.

To play StreetCred, you walk around your neighborhood, competing to add places to the map and improve the places that are already there. This real-time crowdsourced model helps maintain data freshness and accuracy, but it also makes it easy for players to add the same place multiple times. We saw a lot of players mistakenly adding duplicate places, rather than improving the existing ones.

The game mechanics don’t incentivize this behavior, because players get the same “reward” (the chance to drop a cube) whether they add a place or improve an existing one. That made this appear to be more of a UI issue: players might simply be unaware of what they should be doing.

To address this, we developed a new feature.

Adding a place Entering the location Dupe detected
The new feature detects potential duplicates and prompts the player to improve the place instead.

When you add a place, the app searches the database for places with a similar name within 200 meters of the provided lat/lon. If it finds any, you see a list of options for what the place could actually be. Then you get the chance to select one of these and improve it instead.

After we launched this feature, duplicate submissions fell by 50%. While we still have some work to do on refining the process, this evidence suggests that a well-crafted UI design and game incentives can help us solve some of the map world’s most perplexing challenges. Stay tuned for more updates on our progress.

Outdoor advertisement created by DDB, Denmark for McDonald's.