SPSS - UK Customers

Help the Aged

Improving direct marketing response rates using software from SPSS Inc.

The situation

Help the Aged is one of UK’s largest and best‑known charities, raising over £75 million in annual contributions specifically to address issues that matter to older people. Their four main priorities are combating poverty, reducing isolation, defeating ageism, and challenging poor care standards.

Since the fight for charitable donations is more competitive than ever, the utmost importance is placed on using marketing resources and budgets as effectively as possible. One way to do this is to cut costs by reducing the number of recipients of their direct mail campaigns – without losing possible high value donors. Help the Aged are one of the leading pioneers in improving the targeting of fundraising activity.

The challenge

Help the Aged’s old system was to cull supporters from direct mailing programmes using the 'recency of last gift' principle – a common technique among direct marketers. Their mailshots went out to anyone who had donated within the last four years. The problem with this exclusion principle was that it was too gross. Stuart McCoy, Database Marketing Analyst at Help the Aged explains: “If we have two donors who last gave four years ago, we would want to cull the one who gave £5 once much quicker than the one who until then had given £50 every year.”

However, by including ‘frequency’ and ‘value’ variables in their analysis, Help the Aged expected to have a much better indication of a supporter's worth to the charity. They would be able to match this worth against their mailing cost at an individual level, and determine whether or not they wished to retain the donor for direct mailing.

A ‘Recency Frequency Value’ (RFV) model was therefore developed, with the explicit aim of retaining the vast majority of income while reducing mailing costs by an appreciable amount (this type of model is also known as a ‘Recency Frequency Monetary Value’ or RFM model).

The solution

By using software from SPSS Inc., Help the Aged segmented their existing pool of donors, contained in a file holding millions of records, by ranking them from highest to lowest by Recency, Frequency, and Value respectively. Each file was then split into five equal bands (quintiles), and a score from one to five was assigned to each band. Frequency was defined as the number of gifts a supporter had made in their lifetime, and Value was defined as the average (mean) gift over their lifetime (this proved to be a more discriminating factor than total value of gifts).

Each supporter was then assigned a three character RFV score, giving a possible 125 (5x5x5) combinations of individual scores. “Over several historical campaigns, the income was then calculated for each RFV score and then matched against the total mailing costs for particular groups of donors. This enabled us to generate values for net income and return-on-investment so that groups with a negative net income and poor returns could be identified and culled from the mailing programme,” explains Stuart McCoy.

Without the use of SPSS Inc. software, our tools were not flexible enough to allow us to manipulate transactional data to build the RFV scores, let alone compare costs and incomes. So it would have been impossible to duplicate this methodology. One mailing saw a doubling of both response rate and gross contribution per person mailed, when compared to a similar campaign run two years earlier.

Stuart McCoy
Database Marketing Analyst
Help the Aged

The results

By implementing the RFV model and tailoring their selection criteria, Help the Aged found that they could cut costs by mailing fewer people – but still manage to retain income levels from supporters. The savings that were realised from reduced mailings far outstripped the small loss in income from each campaign, so the exercise was deemed a great success.

As well as conferring the ability to perform segmentation exercises such as this, SPSS Inc.’s software proved to be an essential tool for exploratory data analysis. In addition to its modeling capabilities, some useful Key Performance Indicators were derived simply by looking at the boundaries between the RFV scores – for example, what proportion of the database had given more than once, or whose last gift was within the last eight months.

Stuart McCoy concludes that the project was highly successful, saying: “Without the use of SPSS Inc. software, our tools were not flexible enough to allow us to manipulate transactional data to build the RFV scores, let alone compare costs and incomes. So it would have been impossible to duplicate this methodology. One mailing saw a doubling of both response rate and gross contribution per person mailed, when compared to a similar campaign run two years earlier”.

Interested in using SPSS Inc. software to improving direct marketing response rates? Download the Help the Aged PDF here.