I started working at Wongnai. Wongnai is Thailand’s Yelp. We’re a restaurant review platform(we’ve expanded to many other domains). I was assigned to the Growth Team (cross-functional team). The team’s goal is to increase the number of Wongnai App’s weekly active users.
The WAU’s growth has slowed down. To achieve exponential growth we looked at increasing retention. We focused on the weekly active user and weekly retention because typically people dine out once a week.
Setting up the analytics
Kicking off the team we started tracking the app. We were already using Google Analytics but we experience heavy sampling since we tracked our web traffic into the same property. The app traffic is tiny compared to web traffic so the numbers would be sampled and severely unreliable. So we looked at another third-party analytics tool.
- Mixpanel: the pricing is way out of reach for our use case
- Google Firebase: Didn’t support web tracking (Even though our main focus is the app that might (and did) change in the future)
In the end, we chose Facebook Analytics because it was free and has many features and at the time there was no sampling of data.
Churned and Retained Users
Besides the usual event tracking, we mark users into green and red groups. Once a user installs the app that would start counting as week zero if on week one the user comes back to use the app they would fall into the green group. On the contrary, if the user doesn’t come back then they would go into the red group.
Comparing churned and retained
The key here is to find the behavioral difference between the green and the red groups. We look for a feature, content, a button that the green group uses and the red doesn’t. Then we try to improve that feature or get more people to use it.
We listed all the features and look at two metrics frequency of use and volume of use. Even though the green group uses saves 5 times more often than the red group but the volume of use is only 500. It means that it’s far harder to get people to use the save feature.
Factor = ( Total Events By Green Group / Number of Green Group Users ) / ( Total Events By Red Group / Number of Red Group Users )
We compared data in one particular timespan.
| Action / Feature | Total Events By Green Group | Green Group Users | Total Events By Red Group | Red Group Users | Factor |
| Share Restaurant | 100 | 400 | 19 | 600 | 7.89 |
| Bookmark Restaurant | 500 | 400 | 140 | 600 | 5.36 |
| See All Photos | 3000 | 400 | 1200 | 600 | 3.75 |
I’m using mock numbers but these represent the real insight.
Getting the persona down
On the other spectrum, we want qualitative data so we started crafting the persona we randomly interview our users. As a guide, we separated users into 4 groups
- Top Contributor: Elites contributor users who post a lot of reviews, photos to the platform.
- Medium User: Users who contribute from time to time.
- Viewer: Users who come in regularly to read the content and browse restaurants but never post anything.
- Churner: Users who created an account but having use the app for a while.
I interview people on how they pick restaurants. A few examples of the questions.
- How often do you dine out?
- Do you ALWAYS look for review before dining out ?
- Do you extensively read reviews ?
- Who they dine out with?
We ended up with 2 personas. The personas became the basis of many changes to Wongnai web and app.
This was done in early 2017. I’m building this portfolio in mid-2019 this persona definitely can use a revisit.