Aptify's CEO Blog

A recent discussion within Aptify’s User Community on trial memberships sparked my interest in exploring this topic with a broader audience and thinking about how trials might tie into previous topics on this blog including Predictive Analysis and Member Engagement Scoring models.

Trial offers are fairly common in the membership world. But just how good are trial memberships in terms of bringing on new members? The simplest answer to this question might be sought by looking at a simple initial conversion ratio. While this number may tell the early part of the story, it is very important to evaluate the trajectory of that member over a longer period of time. As an example, how active is this new member in other areas of the association – a measure of their overall engagement. Furthermore, how likely is it that the member will renew in Year 2, 3 and beyond?

Are there patterns that can be used to predict what might make one individual more likely to convert past a trial, get active, and then becoming actively engaged? The technology is out there to help with both the predictive analytics question as well as creating a method of “scoring” engagement levels. The question of when to offer trial memberships is a great example that should leverage both of these concepts.

Consider it this way – if you can find a few defining characteristics that help assess if a prospect will go from trial to member and beyond there would be many applications. It would be great to build those characteristics into your core operating processes. Imagine a scenario where your AMS (for staff and on the web too) automatically determined the probability of “success” based on several factors and then selectively (and automatically) offered trial memberships to individuals or groups that fit the model.

As I mentioned in earlier posts, none of these modeling techniques are perfect, but they often do reveal interesting trends. At a minimum, they are worth a hard look. If one or more model is good at predictive work for your organization, find ways to use it frequently and in nontrivial ways.

I previously wrote about some of the theory behind using predictive analysis in a member-based organization. This post is intended to take the next step, looking at some specific variables that might be relevant in a membership renewal business case. Here are some quick examples that may prove interesting in a predictive model. The value of these data points may be greater in some organizations than others and in some cases these attributes may have no bearing at all on renewal probability. These are just some common data points that we've seen in some way correlate to renewal probability.

# of Years as a Member

Is there a correlation between the # of years someone is a member and their likelihood to renew. For example, is an individual/organization that has been a member for one year, less likely to renew than a 10 year member? Depending on the organization and the nature of the membership structure, there could be different answers. For your group, do you know the answer?

# of Non-Dues Related Transactions

Members that are active in purchasing books, certifications, or other non-dues items from an organization may be getting more value for their membership, and therefore more likely to renew than others. Is this true for your group?

# of Years of Uninterrupted Attendance at Annual Event

Attendance at Annual Meetings, as well as other events may be a factor in someone’s willingness to renew membership. A committed member that attends annual gatherings, year after year, is more likely to renew, or are they?

# of Committee Assignments

The involvement level of an individual or organization in volunteer activities, such as committee positions, may have a bearing in their overall perception of value from membership. Is this the case for your constituents?

These factors are simply a handful of dozens or hundreds that can be evaluated quite easily with modern predictive analytics tools. Additionally, a key goal of predictive analytics is to find factors in the data that you didn't necessarily think would impact your intended business goal.

Once you have started the process of running these types of analytical tools, you must have an action plan, otherwise there is little value in the analytics. For example, if a predictive model suggests that your most at-risk members are those that have been with the organization between 3 and 7 years and have not attended at least one event in the least 2 years, what do you do? The action element is just as important as the analysis. Remember, no predictive model is perfect, but they get better as you use them more and more. Don't be afraid to act on good, yet imperfect ideas that come from these tools. Make sure you foster an inquisitive culture organizationally and constantly go back to tune the modeling approach.