Referral Program & Machine Learning Algorithm
I’m really excited to announce that this week I’m co-hosting a dear friend of mine who’s also one of the talented experts in Computer Science, Machine Learning (ML)/Deep Learning (DL) I’ve been working with- Ronen Chen. Together with the #customerconversionchallenge, we decided to add a new edge to the challenge and to combine it with the perspective of ML/DL.
So this week, we’re going to explore Referral Programs as a powerful conversion & growth engine while using ML/AI algorithm to make it a smarter, faster, and efficient process that could do some magic to the company’s growth.
The best-known example of a referral program is the example of Dropbox. Google it, and you will find lots of articles about it. I will give you a short background with the most significant traction regarding it.
Friend referrals are one of the strongest engines for growth and for converting users. Saas, B2B2C and B2C products should consider adding it to their strategy since in some cases, and if you do it right, it’s stronger than any paid marketing tool. Another significant aspect of the referral program is that it’s very simple to measure its success since it’s a direct marketing tool with a designated link to each user that refers it to a friend.
In the Dropbox case, the sender has an incentive to spread the word about Dropbox — getting extra space, and the referee also has an incentive for signing up — more space than if they just signed up through the normal process. A Win-Win situation.
The most important thing to emphasize in Dropbox’s case are the numbers :)
They used the word to mouth and viral strategies-
35% of daily signups from the referral program
20% from shared folders and other viral features
Sustained 15–20%+ month-over-month growth since launch
You can view a great Slideshare of Dropbox’s CEO Drew Houston on their growth strategy.
Now that we understand better the incentive to use a referral program let’s move forward on how to make this program a smarter, faster, and efficient process while using ML/AI algorithms. The simplest example is the registration process. During the registration process, for example, via Facebook/Google, public data can be easily collected using a dedicated Deep Learning Algorithm. It means that we can learn about the users’ behavior, communities, interests, and friends. Then we could use this data for smart retargeting new customers. We could even personalize the content and the offering to the users by understanding more about their behavior. For example, we could offer a personalized template based on the users’ interests to provide to their friends. It’s based on data and not assumptions, which are the most significant difference when using ML/AI algorithms. And the bottom line is that It saves time, money, and a lot of resources.
I know it’s just the tip of the iceberg, and it’s such an exciting topic to discuss. We are creating a webinar on this subject, so if you wish to receive an update, please register here.
Thank you, Ronen, for adding your value to this article and looking forward to more co-hosting articles in the future.
Cheers for now,
Keren
