BOTs 7 Day Retention & DAUs

7 day retention is a tricky thing. It is slightly scary, too. A bunch of users enter your product and only a fraction are retained. We have all seen these reports – 80%, 90% drop. It’s enough to keep you up at night. Unfortunately, this is the nature of our world when one wants to increase DAUs.

Let’s skip the discussion regarding product market fit. We assume product market fit has been established, but what about your product is making end user stick around?

Do you know? Does anyone know? Are you sure? Is the company rallied around this effort? It’s different for each company. Facebook isn’t shy about what works for them. The focus should be on core behaviors. For Facebook it required end users to add 7 friends in 10 day. Slack discussed 2,000 messages between a given team of 50. This was a predictor for 93% retention for Slack.

For our group, we had a similar story. We had a few predictors of retention, and one very simplistic behavior was “messages sent”. The perfect opportunity arose because we just launched our BOT. The BOT engaged with end user the moment they arrived inside the product.

Armed with this information, it was easy to form a clear hypothesis. We launched an a/b test that focused on increasing “messages sent”.

A few weeks worth of data, and we found by tweaking the BOT experience just a little – We increase “messages sent” by 67%. Not bad.

Did we instantly solve world hunger overnight? No. Did we take 1 giant step in the right direction? Absolutely. Is there more work to be done? Definitely.

When you are optimizing for DAUs, it’s this constant process of refinement:

What is the data telling us?

What experiments can we run to optimize for those behaviors?


If you are asking what behavior analytics products support such insights, you want to explore Amplitude. Integrate and then click on compass. If you’ve never heard of them, take a look here:

Look out for a post later. I will discuss Amplitude in more detail.

Brion Hickey