Nutrafol

Nutrafol is the industry leading hair growth supplement brand, using innovation to design a new approach to hair growth. With 100,000s of customers actively engaging with Nutrafol’s clinically effective and patented products, curating the best customer experience is a top priority for the Nutrafol Team. With this in mind, Nutrafol partnered with Vidora to use machine learning in order to better understand customer behaviors and tailor the best experiences for each individual customer.

Challenges

Nutrafol would like to better customize the experience and offer a better personalized service to each customer using 1st party data. Goals include:

1. An improved onboarding experience to match the best products and services with the right customer
2. Proactive support of all customers by anticipating their future needs before the customers have to ask

In order to accomplish the goals above, Nutrafol assessed the ability of Cortex to do the following:

• Leverage all 1st party data to create customer-centric machine learning decisioning
• Provide accurate and usable machine learning pipelines and models 
• Seamless integration of Cortex machine learning decisioning into the Nutrafol business

Implementation

Nutrafol, an existing Twilio Segment partner, was able to leverage the Vidora Twilio Segment integration to both send data from Twilio Segment to Cortex, but also have Cortex machine learning pipelines send back to Twilio Segment. The process was easy, requiring only a quick UI configuration within both Twilio Segment and Cortex. The entire process of integrating 2 years of historical data, in addition to implementing an ongoing data feed into Cortex took Nutrafol only a few hours to set up.

Setting up Vidora as a destination in Twilio Segment is possible with just a few clicks. Once a source is connected to Vidora, all events and traits will be forwarded to Vidora.

After easily setting up Vidora as a destination, the Nutrafol team was able to quickly send data from Twilio Segment to Cortex but also have Cortex machine learning pipelines send back to Twilio Segment.

Setting up Twilio Segment as a destination in Cortex is simple, allowing data to be sent between the two platforms.

The integration with Twilio Segment included data from multiple sources including: customer analytics, subscription information, and help center tickets. In addition, Nutrafol leveraged Cortex’s ability to match IDs across different ID spaces to merge the various data sources into a cohesive user profile with which to build machine learning decisions.

Using Cortex

Working with Cortex, Nutrafol built a variety of customer-centric models which helped them be proactive around customer engagement and ensure each customer had the best experience. Some of the initial pipelines created included –

• Likelihood to Subscribe
• Likelihood to Cancel a Subscription
• Back-dated machine learning models to test the accuracy of the pipelines during previous time-periods

Building machine learning decisioning in Cortex is easy due to the variety of features and technology Cortex has specifically for working with large-scale consumer data problems. 

Cortex features which Nutrafol leveraged included –

• Consumer data ingestion frameworks for events, user metadata, and conversions
• Automated consumer data feature engineering techniques which generated hundreds of unique features for each pipeline
• Model and feature analysis visualizations to assess the performance of pipelines on both historical and current data

Business Value

Given the rapid growth of Nutrafol, it was important to not only have accurate machine learning technology but also to quickly build and iterate on machine learning pipelines. Cortex was able to meet both the Nutrafol requirements.

• The first production pipelines required less than 4 hours of data integration work by leveraging both the existing Cortex integration with Twilio Segment in addition to the various consumer automation features within Cortex. Data integration included 2 years of historical data and an ongoing feed of customer data across multiple sources.
• Accurate machine learning pipelines. As an example, the pipelines that predicted the top 10% most likely users to subscribe accounted for 77% of the actual subscriptions that occurred.

Conclusion

As a company leading an innovative approach to hair growth, Nutrafol knows that a personalized customer experience is best to help their customers achieve their goals.  Partnering with Vidora allowed Nutrafol to use machine learning to anticipate customer needs in order to tailor the experience specifically to each customer. And this was all accomplished with an easy setup and integration with Twilio Segment, which allowed Nutrafol to transform raw data to machine learning decisions in a matter of days.