Summary of File Analysis

Membership File Analysis of Selected Markets

David J. LeRoy

Craig Reed

Sue Greer

TRAC Media Services

January 2003

File Analysis Highlights

• Lapsing is the “normal” state for most people on any station’s file. That is, most people on a station’s file are lapsed. Across a 10-year period, using a mega file with some 1.5 million names from eight different stations, our analysis found 29% of the names were current members, 23% had lapsed for two years, 26% were lapsed for 3-5 years and the remaining 22% were lapsed for more than six years.

• Among the approximately half million current members, 23% were new members, 62% had renewed in the last cycle and 15% had just rejoined after lapsing.

• Even many lifers (members who have given 8 of the 10 years covered in the study) have at one point or another lapsed and rejoined. Only 32% of the lifers never lapsed.

• At any given point in time, the lapsed file contains people who will rejoin in the future. If activated, these future rejoins would increase membership by about 30%.

• The problem for station management is identifying which lapsed members have truly lapsed and which ones are only hibernating, waiting for an opportunity to rejoin.

• People the least likely to lapse are those that use a variety of methods to renew their membership (mail, pledge, telemarketing etc.).

• The average duration between donations for people in this sample is 18 months.

• Because the mail renewal model is the norm for most stations, the average renewal gift is in the $40 dollar range. These are also the long-term members who are dropping off the file.

• The file is shrinking because (A) new members are not transitioning into lifers at about the fourth to fifth year; (B) long term renewers are lapsing off the file (many of who are mail renewers). We don’t know why these two groups are lapsing, but that is the purpose of the next phase of this project – the member surveys

• Among demographics variables, income continues to be a poor predictor of membership, while having a four-year college degree is the best predicator but it is far from perfect. (See geo-demographic section below.)

• Looking at geo-demographic variables, one third of the file comes from well-educated neighborhoods, which only comprise 19% of the population. The remainder of the file comes from “mainstream” middle class groups that comprise a large part of the population as well.

• The distribution of members across geo-demographic clusters has shifted very little over the years.

Lapsed Member File Analysis

This portion of the project looks at the membership files for the eight stations in the study. Suffice to say that there were 1.5 million records of both current and lapsed members. Here we look at the combined files. The closing section of this report will also discuss the use of geo-demographic overlays.

Is the File Shrinking?

That is the research question we attempted to answer in this file analysis. And the answer is yes the file is shrinking and it has two causes: (1) as the file ages, the newer members of the station(s) are not maturing into multi-year members but lapsing off the file; (2) in addition, multi year members are lapsing at record rates which in turns also shrinks the file. Moving away, death, vagaries such as changing ones name, and of course, choosing not to renew for various reasons, can cause lapsing. Currently we cannot ascertain what proportion of lapsers should be associated with which cause. These answers will be addressed in the field surveys.

File Structure and Study Design

To gauge the extent of lapsing behavior in the public television system, TRAC examined eight station’s transactional records over a period of 10 years, beginning in 1992 and continuing through 2001. The nine stations were:

• GPTV (Georgia Public Television)
• KRMA (Rocky Mountain PBS)
• KUAT in Tucson
• KVIE in Sacramento
• WETA in Washington, DC
• WHYY in Philadelphia
• WPBT in Miami
• WPT (Wisconsin Public Television)

We tracked year-by-year giving behavior for 1.5 million members at these stations. For each membership gift, we noted:

• Months since previous membership gift
• Amount of gift
• Method of giving

From these patterns, we were able to identify:

• Members who had never lapsed (i.e., who gave consecutive gifts within 12 months of the previous gift)
• Members who had lapsed and rejoined (gave membership gifts after an interval that exceeded 12 months)
• Members who had lapsed and never returned. These we partitioned as Lapsed < 2 years (12 to 24 months), Lapsed 3-5 years (36 to 60 months), & Lapsed 6 years or more.

For a complete set of crosstabulations of the file analysis data, see Appendix A.

Lapsing Defined

For this report, we defined lapsing as any calendar year without a gift. Stations have adopted larger windows (e.g., 15 or 17 months), but yearly giving is the goal and consequently this stricter standard is used in this report.

Renewal rates—the percentage of last year’s members who renew each year—have fallen steadily since 1995, from a high of 61 percent to 56 percent in 2001. During this time the number of members decreased from a high of 513,000 (1996) to 457,000.

The graph below illustrates that the stations are primarily losing members the middle of the donor pool – the members giving 4-7 membership gifts across the 10 years of this study.

While traditional measures such as renewal rates and member counts are valuable, they are not the only indicators. A closer look at number of gifts, giving intervals, giving amounts, method of giving, and selected demographics helps us identify how members who lapse differ from those who don’t.

Snapshot of the File

Of the 1.5 million members who have given to our test stations in the past 10 years, all but 29 percent are lapsed off the file. About one-quarter of these members lapsed within the past two years (23%); another quarter lapsed three to five years ago (26%); and 22 percent have been gone six years or more.

Year of First Gift

Table 1 displays the status of the members according to when they first came on the file.

• Twenty-seven percent of pre-1995 members are still on the file; most have lapsed (73%).
• Nineteen percent of those joining 1995-1998 are still on the file; most have lapsed (81%).
• One-third of those joining since 1999 (excluding New Members) are still on the file (26%); the remainder are already lapsed (74%).

Table 1 Current Membership Status by Year of First Gift

*Excludes this year’s New Members.

It is the slippage between 1995 and 1998 that is most troubling, as it may foreshadow erosion of the post-1998 membership in the next few years. If time alone accounted for lapsing, we would have expected to have more 1995-1998 members still on the file than pre-1995. By 2005, we’ll know the lapsing rate of the 1999-2001 members.

Current Members

Current members are defined as those who have given a membership gift within the last 12 months. As we saw in the File Snapshot, current members constitute 29 percent of the 10-year file, a subset of the 1.5 million members, or about 457,000.

Current members are comprised of New Members (first-time givers), Renewers (those who also gave a membership gift last year), and Rejoins (those who have given a gift in the past, lapsed, but have reinstated their membership).

By definition, all of the rejoins in the graph above have lapsed in the past, since they weren’t on the file last year. But the other interesting point is that most of the renewals have also lapsed at one time or another. In fact, only 34% of current members (excluding those who are new this year) have never lapsed, and most of these are “new” multi-year members who came on the file after 1999. The longer you are on the file, the more likely you are to miss the 12--month window at least once.

The main point here is that the definition of yearly membership is a station definition, and not necessarily a member definition. Other evidence of this also came from our depth interviews, where many of those who were lapsed by the station’s definition insisted that they were current members given their definition of membership.

Snapshot of Rejoins

Rejoins are people who give a gift—or multiple gifts—and for some reason, fall off the file for 12 months or more and come back on later. During the time they’re off the file, they’re considered to be lapsed, because you don’t know for sure if they will ever renew.

As we noted in the section on Current Members, most of those currently on the file have lapsed and rejoined at least once along the way. Consider Table 2, which shows the percentage from each group who lapsed at least once and rejoined in the past 10 years.

Table 2 Lapsing and Rejoining Rates

*Excludes this year’s New Members.

It would appear that the norm would be that most members lapse at one point or another in their career at each station. Why they lapse can vary from being a little late in getting their renewal letter back into the mail to members waiting to pledge for a specific premium.

Since we looked at 10 years of membership data, we can gauge how many lapsed members will rejoin at some point in the future. The best indicator of this phenomenon comes from the middle years of the study when we are far enough in the past to find future rejoin activity. Looking at 1995 –1998, we find that on average future rejoins could increase the membership roles by 31%.

Note: Future Rejoins are shown on the base of the active membership. The blank slot in the graph is meant as an indicator that these are not active members.

Obviously, not all these future rejoins can be activated in any single year. However, this does illustrate that there are many members who are simply waiting to donate. They are either on their own timetable or that have not been approached with the proper motivator to spark a donation.

Are Lapsers Different from Current Members?

Table 3 shows the number of gifts by membership status. Current members include 21 percent who have given only one gift (New Members). Over half of current members have given four gifts or more.

Table 3 Number of Gifts by Membership Status

In two of the lapsed columns, we see evidence that stations are indeed losing multi-year givers. Sixty-one percent (26%+28%+7%) of those who are lapsed less than two years have given at least two gifts or more. And one-half (29%+21%) of three to five years’ lapsers gave two or more gifts.

Method of Giving

We examined all gifts across the 10-year window and categorized each member according to their most common method of giving.

• If a member gave more pledge gifts than mail gifts, they were considered pledge-dominant.
• Conversely, if a member gave more mail gifts than pledge gifts, they were considered mail-dominant.
• If their most common giving method was neither pledge nor mail, they were considered other-dominant (e.g., telemarketing, online, special events, white mail).
• If no single giving method was greater than the others (i.e., they were equal), they were considered equal mail or pledge renewers.

Table 4 shows the dominant giving methods according to membership status. Note that mail-dominant is the largest group among current members and among those lapsed two years or less.

Table 4 Percent of Giving Method by Membership Status

Regardless how a member was acquired they are expected to renew by mail after certain time on the file. Pledgers and other acquired members lapse at higher rates than the mail renewers.

Table 5 inverts the picture, showing the membership status according to method of giving. Mail-dominant givers are almost twice as likely to still be on the file as pledge-dominant; that is, mail givers are less likely to lapse.

Table 5 Current Membership Status by Renewal Method

Clearly, the mail model is better than other methods given the data above. Again note that the lapsing is the most common behavior, even for mail renewers.

Giving Interval

Stations usually define a membership year as 12 months, a calendar year. Members, though, may not agree. It could probably be argued that the 12-15 month interval givers are simply “sluggish.” Or perhaps the renewal mail doesn’t work; but when they miss the guide, will they renew?

Table 6 shows the categories of giving intervals (along with average giving intervals) according to membership status. Less than one-sixth of current members (15%) buy into the 12-month giving model; rather, one of the largest groups is inclined to give beyond the 15-month window (42%).

Table 6* Average Time Between Renewals

*Excludes member who gave only one gift.

Twenty-five percent of those who lapsed before 1995 (Lapsed 6+) were steady givers -- more inclined than any other group -- to accept the station’s 12-month renewal cycle. Now some of these “steady members” have moved away, died or as more common they are still living the market but have lapsed for whatever reason.

The next largest group of steady givers comes from those intermediate lapsers (Lapsed 3-5). This may signal that the “current file” now reflects a different kind of member, one for whom the 12-month membership cycle does not resonate. Those steady givers of the early and mid-1990s may have been those habitual givers from the GI Generation, who are slowly dying off the file.

There is a hint of good news in this table: The average giving interval suggests that some of the lapsed members—particularly those who have lapsed within the past two years—are simply on their own giving cycle, and they’ll be back on the rolls again.

Amount of Gift

Over the 10 years that we examined, there were a number of changes in average gift amounts, many of which were likely brought about by a growing emphasis on pricey pledge premiums. We wanted to look at gift amounts two different ways: average first gift amount and average subsequent gift amount. By definition, only those who gave more than one gift are included in the figures for average subsequent gift amount.

About half of current members gave an initial gift of $40 or less; slightly less than that proportion (43%) has averaged $40 or less in subsequent gifts.

When we look across the columns from recently to longer term lapsed members, the numbers suggest that longer term lapsed members gave at the basic level. For instance, 45% of members lapsed two years or less gave at $40 or less on subsequent gifts. That same number jumps to 59% among members lapsed six plus years. This may reinforce our assumption that the trends we see are those GI Generation basic level, steady givers who are indeed falling off the file.

Table 7 Gift Amount by Membership Status

When we look at gift amount (first gift and subsequent gifts) by members’ first year on the file, we find many people do respond to “mail model”. Members joining in recent years are giving higher initial gifts (pledge?), but are more likely to drop to basic member levels in subsequent years. This could change as they stay on the file and membership appeals encourage them to give larger gifts.

At the beginning of this section we asked whether or not lapsers are different from other members and indeed they are different. There appears to be some lapsers that are more pledge reliant than current members – however, many lapsers still gave by mail. We also find support that many lapsers were GI generation members, since a substantial number of lapsed members had given on a regular basis within a 12-month window. Similarly there is a large group of long-term lapsers that gave at the basic membership level.

Demographic Differences

The demographic data from Claritas were incomplete, precluding the development of a solid profile of lapsers, current members, or rejoins. This is common to all compiled databases that marketers use to estimate populations and identify potential markets. Claritas collects data from a variety of sources (syndicated surveys, grocery store scan cards, customer mailing lists, etc.), so naturally, all households are not included. In some cases, data are modeled based on zip code.

Table 8 shows the demographic differences we observed among our groups of current members and lapsers. Current members are somewhat better educated than lapsers, especially long-time lapsers. They also have lived longer at their current residence, and are less likely to have children, both suggesting that they may be older. Income does not vary across groups.