Make your loyalty rewards a gamble, not a certainty

People completed almost 6 more surveys on average (33% more) when they were randomly paid either HK$20 or HK$40 at the end of each survey, instead of a fixed HK$40

Jan 14, 2021

Imagine a mobile payment company that offers its customers a cash reward every time they pay $60 or more. Which reward would work better?

Uncertain incentive: The reward is either $2 or $4 with even chances, and the customer will find out which after each payment.
Certain high value incentive: The cash reward is always $4.

Economic theory and common sense point towards 2, a fixed $4 reward.

But is that so?

New around here? Subscribe below for two evidence-based tips each week.

Previous tip: If customers feel distant from your brand, use abstract language (All tips here)

Tip type: Existing research (July 2018)

Make your rewards variable and uncertain, not fixed

Impacted metrics: Customer lifetime value | Purchase frequency
Channels: Loyalty rewards | Promotions | UX

Recommendation

When you use incentives (e.g. rewards, discounts) to encourage repeat behavior or purchases (e.g. 10 stamps get you a free coffee, a gamified learning app, frequent flyer programs) make your rewards variable and uncertain, not fixed. They’ll be more effective and will save you budget.

For example, people would be much more likely to bring their own cup to Starbucks if they randomly received either $0.25 or $0.50 off, rather than a fixed $0.50 off.

Effects

People repeat a task more for an uncertain incentive than for a certain incentive, even when the uncertain incentive is financially worse. For example:
- Runners ran 87% further when they were randomly given either 3 or 5 points for each lap they ran, compared to a fixed 5 points for each lap (1 point = HK$1. Talk about “running the extra mile”)
- Survey-takers completed almost 6 more surveys (25.96 vs 20.31), over 21 days, when they were randomly paid HK$20 or HK$40 at the end of each survey, compared to a fixed HK$40
For the effect to work, the uncertain incentive needs to be revealed at the end of each task or purchase. Without the immediate ‘reveal’ the effect backfires (compared to a higher fixed amount) because people are averse to risks and smaller benefits.
When rewards are uncertain (and overall lower than the fixed reward), people are less likely to ‘enter’ (67% vs. 88% in one experiment). However, the boosting effect on repeat behavior becomes stronger and stronger over time, quickly offsetting the initial drop.
The effect is probably stronger if people don’t know what the probabilities of getting a certain reward are (e.g. 50% vs 50%, 10 vs 10 vs 80%), but this was not directly tested in the study.

Why it works

We like to resolve uncertainty, it’s in our nature. An uncertain incentive not only offers us a financial benefit, but also the fun experience of discovering the unknown through our action.

Limitations

It’s likely that the lowest possible incentive should be above a certain minimum (e.g. not 0), to avoid triggering loss aversion. However, this was not tested.
We don’t know what the ideal probability distribution of incentives should be (50% vs 50% as tested, or 30% vs 70% for example).

Companies using this

Some companies use this ‘surprise’ technique for their products (e.g. Kinder eggs, fortune cookies) but very few currently use it in their rewards and loyalty programs.

Steps to implement

Apply this technique to your offers and loyalty rewards or to influence physical and in-app behaviors. You are in effect adding an additional layer of gamification to your incentives.
Test different amounts and probability distributions to find what works best to optimize both the effectiveness of the reward and your budget savings (from a lower total cost of incentives, on average).
Think of how you can minimize the initial negative effect of uncertain incentives on ‘entering’. For example, you could tell your customers that the first incentive to sign up is fixed (e.g. $10), and gradually for each purchase rewards widen and become more random (e.g. between $1 - $15).

Study type

Lab and field experiments, Hong Kong and United States

Source

Shen, L., Hsee, C. K., & Talloen, J. H. (July 2018). The fun and function of uncertainty: Uncertain incentives reinforce repetition decisions. Journal of Consumer Research.

[Link to paper]

Affiliations

CUHK Business School, Chinese University of Hong Kong; Booth School of Business, University of Chicago; and Carnegie Mellon University. Hong Kong and United States

Remember: This research could be disproven in the future (although this is rare). It also may not be generalizable to your situation. If it’s a risky change, always test it on a small scale before rolling it out widely.

Want to reproduce this, share feedback, or ask a question? -> Reach out at thomas@ariyh.com

New to Ariyh? -> Subscribe below or read 37 other evidence-based marketing tips here