Calculating Liability Reserves for Loyalty Programs

The finance department of a large hotel chain was concerned that it was understating the size of the financial liability incurred when it awarded “frequent guest” points that guests could redeem in the future for free room nights. The sales department found its loyalty program very effective in improving business hotel occupancy and revenue, and it wanted to expand the program. The problem was that it was unclear what the size of the hotel chain’s future liability was. Over the history of the program, frequent guests were redeeming free hotel nights, but their point balances were also increasing over time. Billions of points were outstanding, and it was unclear how many of these points would ultimately be redeemed and when they would be redeemed. Moreover, there was evidence that if the company attempted to dilute (or otherwise lower) the value of the points, people would accelerate their redemption rate. This result prevented simple cost control methods from limiting the liability easily.

Also at issue was the impact of the distribution of redemptions over time. More specifically, the existence of the frequent guest program was most valuable to large business and airport hotels where guests earned a lot of points but relatively infrequently had to redeem them. Conversely, resort hotels redeemed lots of points, but relatively few points were earned there. Since most of the hotels were franchises, the rules of the frequent guest program needed to be perceived as fair to all parties for the frequent guest program to be successful.

Ventana built a simulation model to study the liability and fairness issues. We discovered that the extant auditor model was flawed. While it seemed to perform correctly for the prevalent conditions, it failed a reality check if conditions changed and program members started drawing down their point balances. Nevertheless, the Ventana assessment found the companies liability reserves sufficient and that with appropriate business rules an expansion of the frequent guest program was financially responsible. Subsequently, the company expanded the program, and it became one of the most successful loyalty programs in any industry.

Ten years later a division of the hotel chain asked Ventana to reexamine the liability issue for one unique segment of its business. The sales department of this segment wanted to reduce the size of its liability reserve account to provide new funding to the sales department. The auditor’s model predicted what all company managers felt was an unreasonably high ultimate redemption rate for the points.

Ventana was gratified to learn that over the previous decade the auditor’s redemption model had changed to one very similar to what we had proposed a decade earlier. Moreover, the intuitive case the company managers built for reducing the liability research for the one unique business segment seemed sound.

Ventana built a new simulation model and examined the historical data covering the earning and redeeming of points for the business segment, including unique rules governing the operation of the segment’s loyalty program. This time the assessment was not favorable to the company’s sales and marketing staffs. Not only could we find no basis for reducing the liability, the model predicted that the ultimate fraction of points redeemed would continue to climb over time. This would necessitate continually increasing the liability reserve. These initially counterintuitive results devolved from the unique rules adopted for the loyalty program of this particular business segment a decade earlier. These rules produced no apparent serious consequences for several years after they were adopted, but the seeds of a long-term problem were sown. There were already early signs of the emerging trend in the liability, but the year-to-year changes were so small that managers did not notice the signs (sometimes called “boiled frog” syndrome). Moreover, there was no way to redress a serious imbalance in the corporate program without changing the redemption rules. This was deemed infeasible by the segment’s CFO, although the issue scheduled for periodic review. The business segment had to live with the situation, and the liability account was allowed to grow as required.