Repost from Li's Newsletter: “Lessons on Points Programs for Crypto Apps” The full report and all related findings are available on the official website of Li's Newsletter.

[The below expands on a tweet thread (X essay?) I posted earlier this week. Let me know what you think below, and join the conversation on X!]

Seemingly overnight, points have entered the crypto zeitgeist as a new tool that app builders are leveraging to enhance user retention and engagement. 

Across the landscape, founders are adding offchain points programs to their applications, from Rainbow wallet giving users points for using Ethereum, to Friend.tech building its engagement loop around points, to NFT marketplace Blur’s new L2 Blast incentivizing users with Blast Points for bridging funds (over $800 million TVL since November). In some instances, these points allude to a future fungible token with real economic value; in other instances, users have made that leap of faith themselves. This trend comes amidst a broader search for product-market fit in crypto and ways to engage users in the middle of a bear market.

But outside of crypto, points programs have been a longtime fixture of consumer applications, from games to brand programs like Sephora Beauty Insider or Starbucks Rewards. At a high level, points programs let users earn points for completing various activities and redeem or use them in various ways, with the intention of motivating users to engage. 

Ten years ago, I worked on a web2 mobile shopping app, Shopkick, that had 3 million MAUs and partnerships with national retailers like Macy’s and Best Buy. The app rewarded users for actions like walking into physical retail stores, engaging with products in-store, and browsing in-app. We built a program where users earned in-app points for these activities, which were redeemable for gift cards at various merchants.

Some of my learnings from that experience can help guide web3 projects as they develop points programs:

1. Any sort of extrinsic motivation warps user behavior

2. Point programs change the types of users who decide to use your app

3. Maintaining ambiguity in points value gives you flexibility

Let’s drill into each.

Thanks for reading Li's Newsletter! Subscribe for free to receive new posts and support my work.

Subscribe

1. Under a points program, user behavior morphs to respond to incentives.

People respond to incentives, and inorganic activity arose as a result of our points system. Behaviors sprouted up like walking in and out of the store immediately just to collect points—things people would never do without extrinsic incentives. We dealt with this by capping the number of points users could earn from certain activities and by building a fraud detection effort.

Even among users who have organic interest in your product, the presence of points will distort their activity. Think of waiting for something to go on sale at a retailer that is known to frequently discount items; these stores have trained their customers never to buy at full price. Similarly, points can train users to hunt for earnings opportunities where they wouldn’t have otherwise done so. Short-term boosts in user activity can actually come at the expense of business health in the long run. These long-term adverse impacts can take years to appear and are tricky to unwind. An example is JCPenney’s failed “Everyday low prices” strategy after years of steep discounts and coupons.

Due to the distortionary impact of incentives, Shopkick also had to manage how we monetized and tracked metrics as a startup. Since our revenue came from retailers and brands who were interested in driving engagement, understanding the value of user actions was critical. Incentives altered what a store walk-in or a product engagement was worth, and so it was important to close the loop and track the end KPI of interest: conversion and revenue uplift from our user base. Founders building a points program should be careful to define KPIs that encompass their North Star goal, not simply the actions being incentivized.

2. Having points changes the types of users who use your app.

Points don’t just change users’ behavior on the margin—they actually alter the composition of your user base. Many builders intend for points to boost retention and engagement, but more fundamentally, a points system will change who decides to use your app in the first place. While we designed Shopkick with shopping enthusiasts in mind, the presence of points attracted bargain hunters and extreme couponing types—akin to real-life “airdrop farmers.” This mirrors research from the psychology world about how financial incentives crowd out intrinsic motivation.

This kind of shift in the user base is fine if the persona you’re attracting through extrinsic incentives has a reason to stick around (continued incentives and/or underlying PMF) or if they’re still net additive for your particular business model—think credit card points or airline miles maximizers. In our case, we had a sustainable business model that could fund the continued existence of the points program. But any app that’s funding rewards unsustainably from their balance sheet should do so with caution, and be cognizant that when subsidization ends, users who are simply responding to incentives will churn, especially since they were never the target user of the core product.

3. Maintaining ambiguity in points value gives you flexibility.

If you’re going to tie points to some real economic value, a best practice is to keep the precise value ambiguous. This gives you discretion to change the value of points to manage costs and test incentives, while also preserving the fun for users. For example, the restaurant loyalty and rewards network Blackbird gives users $FLY tokens (which are offchain) but keeps the specific value vague, writing in the app: “It will be redeemable for incredible stuff, like free cocktails and premium access.” 

Elucidating that a certain action = X dollars can diminish user motivation if the amount is too small. In Shopkick’s case, users earned points that could be converted to various gift card values, but the conversions varied across rewards. When users walked into a store or did other actions to earn points, they weren’t thinking about the dollar value of their action, which was nominal, but rather in terms of points, which felt more meaningful.

Points have a point

I’m hopeful that in the crypto world, implementing points onchain can unlock interesting experiences for builders and users. While the points program I worked on at Shopkick was limited to the confines of our app, using blockchains to track points can enable an ecosystem of applications to build around them. That can give rise to fascinating new user experiences. In the shopping context, it’s easy to imagine other brands and retailers wanting to see who the most loyal shoppers were at other stores and target their offers accordingly, which has precedent in the airline world with status match programs. For users, the added value of interoperable points across multiple apps could also make it more compelling to accrue them, reducing the onus on each individual app builder to bootstrap their own utility.

While I’ve outlined some key cautions to take with points programs, it’s important to note that I’ve also seen real benefits from them. At Shopkick, we deployed points in targeted ways to motivate users to move through the funnel and change their behaviors in the real world. For a nominal amount of money, we drove dramatic impacts on long-term retention and referrals. The devil is in the details, and implementing an effective points program requires continued experimentation and iteration, modeling of economic impacts, and rigorous tracking of KPIs.