Our client, a century-old member-funded outdoor recreation club, was experiencing a multi-year decline in membership. They were anxious to identify the cause of this decline, which was impacting their bottom line. In 2011, they undertook a survey of current, past, and non-members to better understand what compels and prevents people from joining or renewing their memberships.

The analytics team at Catalysis applied a variety of data mining analyses to the survey data to find the keys to retaining existing members and getting prospective members to join. This case study specifically explores the process and tools used to identify causes of attrition, reasons for event attendees not becoming members, and, ultimately, ways to increase membership.

THE ORGANIZATION

To support its mission of helping people explore, conserve, learn about, and enjoy the outdoors, our client offers a wide variety of classes, activities, programs, and services to the public. While most of their offerings are available to non-members, there is often a price advantage for members. In addition, members have access to member-only events, discounted rates for some travel and lodging, lift tickets, and books, as well as access to other outdoor organizations.

Through its classes, the organization offers instruction in a wide variety of outdoor activities from climbing to sea kayaking to photography and folk dancing. They also offer organized activities like bicycle or climbing trips and social events. Programs include youth and family programs, volunteer opportunities, and conservation programs, among others. Services include such things as use of the organization’s library and bookstore, lectures, publications, and lodges.

PRELIMINARY ANALYSIS

The team began their analyses by looking for inconsistencies in the data, as well as normalizing it to enhance the ability to data mine. At this point, the team created a definition of “engaged” (members who joined or renewed) and “not engaged” (those who do neither). Because the data suggested there was little difference between the joiners and the renewers, both groups had the same motivations for membership.

Next, a preliminary analysis was undertaken to determine if data mining would find any correlates to engagement. Finding a clear correlation between membership and paid class attendance (everyone who had attended one or more classes had joined or renewed) the team decided that there was enough information in the data to continue the analysis.

Using a tool called MARSplines, the team performed the next layer of analyses to identify, at a high level, the attributes beyond class attendance that would predict engagement and therefore membership or renewal. In particular, the analysis identified the following three primary predictors and their relative importance according to the MARSplines results:

To validate the results of the MARSplines analysis, the team used respondent familiarity with services, and the number of activities and programs, to predict whether they ultimately joined or renewed. The model proved to be 77% accurate which is a strong indicator of the validity of the model.
Next, the team dug deeper into each of the predictors to look for trends which could indicate areas for improvement.

PREDICTOR #1: ACTIVITIES

The analysis of the activities also began with the use of MARSplines. However, the resulting data mining exercise did not reveal any pattern, suggesting to the team that perhaps there were different kinds of people or segments with different engagement motivations, and that looking at everyone at once was hiding the motivators to membership. So the team drilled down deeper to see if they could identify groups of people based on the types of activities in which they participated.

Using a Kohonen Map Cluster Analysis, the team identified three groups of people among the activities: those who primarily walk (blue), those who do a combination of pedestrian sports like day hikes, scrambles, snowshoeing, and walking (green), and those who are very active in a wide variety of outdoor activities (red).

The three groups had quite different levels of membership engagement. In the graph below, ‘0’ indicates not joining or renewing and ‘1’ indicates those who do join or renew. The pedestrian sport greens are centered over 0 or not join/renew, while the walking blues and the “everything” reds are centered over 1 or join/renew.

PREDICTOR #2: PROGRAMS

After identifying the three activity groups, the team analyzed the impact of participation in a type of program on engagement. Once again, MARSplines was used to determine the relative importance of the types of programs.

Again, the higher the score the more important the attribute is to predicting the outcome which, in this case, is likelihood to join or renew. The results show that participation in outdoor education programs is three times more important to joining/renewing than are the family programs. In addition, the youth programs and lodges have no impact at all.

Based on this analysis, Catalysis recommended the client increase awareness and use of outdoor education, conservation, and local branch programs in order to increase engagement.

PREDICTOR #3: SERVICES

Finally, the team looked at the correlation of the services used with the likelihood to join/renew. Using MARSplines, the team observed the following:

In this case, the importance numbers highlight the need for an emphasis on the website and an appreciation of the articles provided, followed by the bookstore and using the website to search for activities. Further, there is an opportunity to better focus the services: eight of the services had 0 importance vis-à-vis renewal/membership, which means there may be an opportunity to streamline operations and save money by decreasing spend on those services.

THE RECOMMENDATIONS

Based on the analysis, we recommended that our client take the following steps:

– Given the direct correlation between participation in classes and membership, the organization should focus its efforts on getting new members and potential members involved in classes as soon as possible.
– The organization should also:

• Obtain a deeper understanding of the lack of “green” engagement – those members who participate in a range of pedestrian activities.
• Focus on promoting the outdoor education, conservation, and local branch programs.
• Revisit the services offered to ensure that they are focusing on the value provided to members.
• Identify any barriers to use of those services that are under-utilized.

CONCLUSION

The number of activities, programs, and services offered by this organization and the varying impacts that they have on the engagement of their members and prospective members lends itself to the use of sophisticated data analysis tools such as MARSplines and the Kohonen Map Cluster Analyses. Other uses for these types of tools include:

– Developing a deeper understanding of what drives customer behaviors and loyalty; an area that any business from banks to casinos to groceries should consider.

– Identifying what is important to customers; what they focus on for advertising, what drives purchase or membership, or even which attributes of products to emphasize in order to increase sales.

ABOUT CATALYSIS

For nearly 20 years, Catalysis has specialized in the digital integration of award-winning marketing campaigns that drive connected, measurable results. Our clients include Microsoft, Moss Adams, Banner Mattress, Thunder Valley Casino, BabyLegs, and WineBid.

For more information, contact info@catalysis.com or visit our website at www.catalysis.com.

The information contained in this publication is general and is for informational purposes only. Catalysis makes no warranties, express or implied, in this material.