Machine learning turns membership data into actionable insights
Society of Actuaries Case Study
SCROLL
OPPORUNITY
A well-established professional association had a membership database containing a wealth of information on their 33,000+ members. The association wanted to use this data to help understand why some members were more engaged than others. The end goal was to boost attendance at the association’s events. In addition to providing value to members, these events are a key way in which the association sustains itself financially. Event promotion is critical to the association’s marketing efforts. However, with such a large database, the association didn’t know where to start when it came to turning all of that data into marketing insights.
SOLUTION

Machine Learning

Our Analytics + Decision Science team recognized right away that this was the perfect type of problem to solve with machine learning
Broadly speaking, machine learning is the use of statistical algorithms to find patterns within large amounts of data. Machine learning has been the key to recent breakthroughs in artificial intelligence, and there are also lots of common applications that we encounter in our everyday lives, like when Netflix recommends something to watch or Google autocompletes a search query.
In order to solve our client’s problem, we used machine learning to analyze their membership database and organize the members into different groups, or clusters, based on a variety of attributes, including professional experience, membership tenure, past engagement, and others. After the machine learning algorithm broke the database into different clusters, our analysts drilled down on the results. Using tools like principal component analysis, we could see what attributes the algorithm weighted most heavily when dividing up the clusters.
The next step was to work with our Digital Marketing + Strategy team to identify which clusters had the most potential for growing engagement and ultimately boosting attendance at the association’s events. In the end, we found five clusters that had clear potential based on their attributes. Then, we took those five clusters and created five distinct member personas, just like the customer personas we’re familiar with as marketers. Using these personas, our client can create personalized marketing campaigns with specific content tailored to each persona.

On top of this, we even worked on a proof-of-concept machine learning model that would identify which individual members would be most likely to sign up for a new event. This tool could help our client avoid fatiguing members who are not likely to attend an event, as well as determine which members to potentially target using more costly marketing like direct mail.
RESULTS

Our client was thrilled!

5
We delivered 5 actionable member personas built from actual data
5
Based on these personas, we helped our client develop 5 workflows for highly-targeted email campaigns
20
Each workflow consisted of 20 unique paths, targeting each persona with personalized content to guide them toward registering for an event

With machine learning, we were able to parse and analyze a massive amount of data in ways that would have taken a human analyst months or even years to do

This new structure is allowing our client to leverage their large database in a strategic, measurable way with a focus on engagement rather than volume of messages sent — a huge improvement of user experience for the organization’s membership

If you have a big, difficult dataset that you know holds potential but you’re not sure how to get started working with it, give us a call.

We’d love to help your team leverage machine learning to tap into the potential of your data.

VIEW ALL

Case Studies

NEXT CASE STUDY

Data-Driven UX + Site Architecture Boosts Website Performance

Ready to start a project?
let's work together