# Reward Programs: Interval Schedules

|  July 15, 2005   |  Comments

Extremely successful rewards based on time rather than number of transactions.

This is the fifth column in series addressing customer loyalty and retention. Though real customer loyalty doesn't come from a program, "loyalty programs" are often used to retain customers through rewards or other incentives. We've looked at all the major reward types, including continuous rewards and ratio rewards. Today we'll talk about interval rewards.

Interval Schedules: Based on Time

Unlike continuous rewards (which occur every time) or ratio rewards (which occur every X times), interval rewards aren't based on any number of transactions. Instead, they're based on time. In B.F. Skinner's experiments, an interval schedule would result in giving food to the mouse every X minutes, regardless of how many times the lever was pressed. Like ratio rewards, interval rewards come in two flavors: fixed and variable. In the fixed version, the reward is given every X minutes. In the variable version, the reward is given at various intervals, all averaging to X minutes.

Real-world examples include any kind of time-based sale. For years, GNC had "Super Tuesdays." Members got a discount on the second Tuesday of each month. Macy's and other retailers often have huge annual sales. Because these sales occur on a fixed schedule (and are independent of how many times you shop there), they're fixed interval schedules.

Variable interval rewards are harder to find in the real world. One example is radio station giveaways. The DJ might say, "Sometime this hour, we'll give away two tickets." In another promotion, the audience has to listen for a certain singer whose song will be played at some point in the morning. Then at some point in the afternoon, another song by the same singer is played. The 10th caller who can identify both songs wins the prize. Both are examples of variable interval rewards because the timing isn't exact but averages to every hour or every four hours.

Charting Interval Reward Effectiveness

Interval rewards are extremely successful. The variable interval schedule is proven to be the most effective one on animals. The fixed interval version is a close second. Here are the ratings according to our three measures:

• Learning curve: medium-high

• Frequency: high and consistent

• Decay: low

The learning curve is pretty slow for this reward because takes a while to get accustomed to it. Though tests with animals show these rewards (and ratio rewards) have relatively slow learning curves, we humans have an advantage: we understand the reward schedules because we're told what they are. While a mouse has to figure out the food comes every hour, we understand because those are the promotion's rules. That doesn't entirely mitigate the learning curve, however. We're still only being rewarded every so often, not constantly. So the learning curve, while not as slow as it is with animals, is still slow.

The real magic of these two rewards schedules is their frequency and decay. Both fixed and variable versions show a very high frequency, without the scalloping of ratio rewards. Decay rates of interval rewards are extremely slow. In fact, these rewards have the best decay rates of the group. In lab tests, the variable version has a better decay rate, but in tests with humans the decay rates are closer to each other. This is good news. It's hard to imagine a lot of different variable schedules (other than the radio example).

Think Interval Rewards Through Carefully

Interval rewards can seem like a magic bullet, as they outperform all other reward schedules. But they must be implemented with caution, and with a real understanding of when they are appropriate.

Let's consider GNC's "Super Tuesdays," which has now been expanded to the first 10 days of each month. At first glance, this is a classic fixed interval schedule. Regardless of how many times you shop, discounts only occur on the second Tuesday of the month. A closer look at GNC's business model shows this reward schedule may actually be a mistake.

GNC sells vitamins, exercise supplements, and the like. Its products tend to be expensive (designer vitamins can cost over \$30). They're replenishable goods; once the bottle's empty, you need more.

Given these business rules, is its interval reward schedule effective?

Yes and no. By having a recurring discount schedule, GNC understands its customers aren't buying supplements just once. Therefore, a first-time buyer discount would be ineffective. It also realizes most of its products come in one-month supplies. If GNC's high-value customer actually buys only every two or three months, this monthly incentive isn't a bad idea. If its high-value customer normally buys monthly, however, the reward schedule coincides with his normal buying habits. This is bad.

If the reward schedule matches GNC's high-value customers' normal buying schedule, the retailer would actually create a continuous reward schedule masquerading as an interval schedule. The people most likely to use the program are very serious buyers (they must be GNC club members). Because they replenish their supplies monthly and every time they buy they get a discount, the schedule actually works as a continuous reward system for these customers. They get a discount each time they buy.

Because high-value customers stand to save a lot of money on discount days, they'll wait until the discounts are in effect to buy. A low-value customer (whose savings aren't as great) won't pay much attention to the sale day. This means the frequency level for high-value customers is low: they'll only shop once a month. The frequency rate for low-value customers is unaffected.

This isn't a good business model. Training the most valuable customers to only shop when they get deep discounts effectively turns them into a liability. Moreover, because GNC's high-value customers only shop during sale days, removing the program would show a high decay rate, mirroring that of a continuous reward schedule.

Interval reward programs only work if they don't affect user behavior the rest of the time. Macy's annual sale doesn't affect consumer behavior the rest of the year. Unfortunately, it's so infrequent it doesn't help build customer retention, either, and can't strictly be classified as a retention program.

There's a happy middle ground. An effective interval schedule can be mixed with other schedules. Airlines do this effectively with double-point weeks. This promotion works in conjunction with regular promotions to increase business without becoming a business rule (as many continuous rewards would). My gym runs promotions for discounted training sessions during certain weeks. This doesn't deter me from buying sessions at other times because promotions are less frequent than my normal buying pattern.

Although interval rewards are the most effective, they have specific (and sometimes limited) uses in the real world. I'd argue a more effective program for GNC would be a simple "bottomless bowl" program. This would reward high-value customers with the convenience of automatic product refills. There wouldn't have to be a deep discount, as savings would be in convenience, not money. An interval reward schedule for replenishable goods is a tricky matter, especially if the reward is timed with a customer's normal purchase behavior.

Coming Up

Next is the final installment in this series: how to use all these schedules competitively. We'll address what to do when your competition launched a promotion and how to combat it.

Until next time...

Jack

Get the ClickZ Analytics newsletter delivered to you. Subscribe today!

### Featured White Papers

2015 Holiday Email Guide
The holidays are just around the corner. Download this whitepaper to find out how to create successful holiday email campaigns that drive engagement and revenue.

Three Ways to Make Your Big Data More Valuable
Big data holds a lot of promise for marketers, but are marketers ready to make the most of it to drive better business decisions and improve ROI? This study looks at the hidden challenges modern marketers face when trying to put big data to use.