Startup Lessons: Event-driven Customer Acquisition

We all know how important it is to track your site’s page views. But page-view information by itself is as useful as having a web site that tells people what you do without providing the means for them to take immediate action.

Enter event-driven analytics. Event-driven analytics is a means of analysing user behaviour through events in real-time. This serves as an immediate feedback loop that you can harness to align your service or product to customer needs and ultimately increase the conversion rate. In this post, I will share a story that highlights the importance of good event-driven analytics.

Analysing and Drawing the Hypothesis

At SweepSouth, we recently had an interesting case where the event based analytics made the key difference that ultimately lead to a subscription-based conversion (yay, another happy customer!). So here is how it all happened.

We keep a close eye on our analytics dashboard for all the real-time user events taking place. On this particular early Tuesday morning, everything looked pretty regular until we noticed a user session that seemed to have gone back and forth from one event to another no less than 6 times during the booking process. This is typical behaviour of a user that has somehow gotten stuck (and in this case, is very adamant to complete the process). Looking at the sequence of events that took place, one could draw a hypothesis with high confidence.

In our booking process, the customer inputs the email address if they aren’t already signed in. The system recognises the email address, then it brings up the “sign-in” dialog without interrupting the booking process. The two specific events that were triggered here were the “Existing User, Sign-in Required” and the “Contact Details Submitted for Validation” events (pretty self-explanatory).

Following the “Existing User, Sign-in Required” event, we could clearly see that our customer (let’s call her Jane Doe) had a problem signing in. At this point we had a light-bulb moment.  “Jane couldn’t remember the sign-in credentials that SweepSouth automatically sent during her previous “onetime” booking”. Easy, we thought, we’ll just pop Jane an email with the “reset password” link to our site. The problem was that she had just left our site with no signs of returning. Really crappy feeling for a startup! 🙁

Weeding out the Assumptions

“No need to give up”, we thought. Jane had already typed in her email so we had a way to drop her a quick note in the hopes of getting her back to the booking process. We spent the next 10 minutes analysing the exact steps she took. Through the analytics, we had come up with a strong hypothesis that Jane wasn’t only stuck on signing-in, but that she was also confused by why we kept blanking out the email address field every time she canceled the sign-in process. This had been obvious to us while building the system, but it turns out that it was less obvious to our customer.

We spent the next 15 minutes coding the changes to enhance our booking process user experience, based on the hypothesis we formed, and we pushed the changes to production. Finally, we popped Jane a nice email with a “thank you for using us in the past” note and mentioned that we had managed to iron out the things she might have had difficulties with a few moments ago.

It was a moment of victory! Jane in fact came back and this time seamlessly completed the process (Yay, happy moment!). We managed to fix the problem that our customer had (and many others could potentially have), and we managed to make a conversion in the process! Even better, it wasn’t a “onetime” booking but was instead a reoccurring/ongoing booking.


Our conversion rates are well above the mean at this moment, and this has a lot to do with good event-driven analytics and the ability to respond really quickly from both the technical and the customer engagement perspective. The key takeaway here is that

If you are not analysing real-time customer behaviour on your site, then understanding your online customers is as good as making not-so-educated guesses.


Social Share :