We present pacts - a new way of connecting crypto projects to power users and, ultimately, a new onchain marketing primitive. Through utilizing user onchain history-based screening and streamed milestone-locked rewards, pacts can offer higher longer-term user engagement and protection from sybillers and bots, saving money for protocols and making sure that marketing incentive budgets go to real users. The design of pacts is rooted in behavioral research and analysis of advantages and pitfalls of current crypto marketing toolkits.

The lay of the crypto marketing land

Uniswap has undoubtedly popularized airdrops back in 2020 by providing a stimulus check to the larger crypto community during the defi summer. $UNI was a retroactive airdrop, which makes total sense - the project managed to gain a critical mass of users and escape the early growth stage, and wants to give equity to all people who used it previously. After all, early users are instrumental in creating network effects and stress-testing the product in the initial stages.

Retroactive airdrops, however, work only for projects that have already managed to find product-market fit with a reasonably sized audience. Most crypto startups nowadays do not have the luxury of PMF yet and need to use marketing budgets to find an audience for their solution. An easy example of a solid distribution - Dymension, being a modular blockchain, airdropping tokens to Celestia stakers (modular infra early users/investors), active Ethereum addresses (certified power users) and certain NFT holders (powerful signal of being a power user, more on that later).

The paradox is that the mere possibility of retroactive airdrop incentivizes sybil user activity trying to mimic as organic, with the goal of getting (and dumping) tokens later. A hilarious example of that perverse incentivization is that after Starknet rewarded people for Github contributions, even small ones such as correcting typos, Githubs of other yet tokenless L1s and L2s started to get flooded with people correcting typos (or correcting what they perceive as typos). Clear application of the old adage “Show me the incentives, I’ll show you the result”.

Liquidity mining, previously a dominant tactic of protocol bootstrapping, was also proven inefficient in terms of marketing spend per user. According to Sixdegree research, liquidity mining is not effective as a means of increasing long-term liquidity nor as a means of user acquisition (in Uniswap’s LM case, cost per acquired user was 7k$+ USD).

Another popular distribution/brand awareness tactic are quests, with a lot of platforms having relatively similar offerings - Galxe, Intract, Zealy, Layer3, and so on. Questing usually involves setting up a reward pool for mostly offchain tasks. Since the number of participants is quite large, rewards are usually given away through a raffle - the purpose of the raffle is to ensure that the amounts per winner are not extremely low.

Airdrops and questing are dominant tactics in crypto marketing. However, they both suffer from the same problem - they attract the wrong kind of user, mainly the ones that are using the chain/product only because the quest was offered, and have no intention of actually using the product after receiving the reward. The root of the problem lies in adverse selection.

Adverse selection

Adverse selection is a quite old concept from economics, or, more precisely, contract theory, which explains situations where buyers and sellers have information asymmetry, which leads to worsening the quality of goods in the market for everyone.

The original theory (Akerlof, 1978) uses used cars as an example. When buyers go to purchase a used car, they mostly see normal, working cars. However, some of the problems associated with used cars (for example, suspension problems) are not easily detectable on initial checkups and will be discovered by buyers only after purchasing and using it for a bit. These cars with hidden defects are called lemons (apparently, it’s old British slang). The visualization is provided below:

The problem is that with lemons on the market, the average price that the buyer is willing to pay decreases. Assume the value of a normal car is 5000$ and the value of a lemon is 2000$. Therefore, if buyers expect 2 out of 5 cars to be lemons, they will be willing to pay (5000 * 3 + 2000 * 2)/5 = 3800$. Which, in turn, will make sellers of good cars quite mad - why should they get 3800$ for a car worth 5000$? Lemon sellers, on the other hand, are happy and content, as they still get 1800$ more for what their used car is worth. Owners of good cars will refuse to sell, which means that % of lemons on the market will increase, which makes the average price down more, failure of the free market, everything breaks down, nobody is trading cars anymore

I hear you all saying “uh, can we get back to crypto pls?”, so back to incentivization mechanisms. In our case, “lemons” are users who never intend to use the protocol after the airdrop program or quest ends. In essence, marketing budgets are wasted on these users. “Valuable” users are ones who were attracted to the chain or product because of the promotion but are potentially interested in using it organically. The situation looks the following way:

Instead of coming up with random numbers to illustrate the example, let’s look at the real ones taken from the Sixdegree blog (check them out, they are an incredible team of data scientists in crypto space). And the numbers are pretty damning.

Cost per acquired user for airdrops is in the 4-5 figure range, which is very high compared to other financial services (200-1000$ range). The user retention also drops in more or less similar patterns for protocols that airdrop tokens. Sixdegree themselves provide the following conclusion:

• Given the increasingly fierce competition, crypto Protocols need to move away from broad-brush user growth strategies. Instead, they should leverage token incentives for more refined user growth and operations.

• Data-driven business growth is an inevitable path, and this approach has already been proven effective in traditional internet companies.

Increasing the efficiency of the marketing budgets can be achieved by minimizing the amount of sybillers in the pool, allowing not only to reduce unnecessary marketing churn but also to reward the “true” users more.

Anti-sybil slightly chubby flightless birds

Preventing sybils ain’t easy, but there are a few hacks around it (which is why Pudgy Penguins are getting all these airdrops). But don’t take our word for it and rather listen to the GOAT:

We have already defined “valuable” users as those who will continue using the product even after the promotion ends. Allowing everyone to participate in a promotion makes real users lost in the crowd of sybillers. This effect is quite well-observed even in the meatspace, with a lot of bargain-hunting customers, up to 75%, never repurchasing from the same store again (Reichheld & Schefter, 2000). Targeting only S-tier NFT holders filters for the following:

• User is most definitely real

• User is crypto-native

• Likely a defi/crypto power user

Even though it is not a given that these users will continue using the product after the promotion ends (maybe they don’t think value added is big enough without incentives, or they simply like the competitor more), eliminating fake users results in the higher part of the reward pool going to the valuable users, as shown below:

Pacts

As an alternative to quests, which are essentially sybil farms, we propose pacts - a new model of targeting and engaging with crypto users. Pacts change the way promotions are discovered and the way rewards are distributed. The solutions are supported by findings in academic and customer research literature as well as our own observations in crypto.

Questing reward model is fundamentally broken

Questing platforms have the following reward model - do an offchain/onchain action and claim the reward once confirmed. This system does not incentivize longer-term retention, as the user does not have a reason to stay past the reward claim. This contradicts the literature, which states that reward systems should be terminated gradually (Melnyk & Bijmolt, 2015).

Pacts propose a different reward system - streamed milestone-based rewards. In this context, streamed rewards mean that once the user completes the initial milestone, they receive rewards every second (accumulated rewards claimable at any given time). As pacts are aimed at fostering longer-term engagement, with this reward system, we aim to provide longer-term engagement and habit-forming of users while slowly getting them off incentivized usage and transferring to organic.

Secondly, to incentivize desired actions (such as trading volume for a derivatives protocol or trying out new features) Pacts introduce a milestone-based reward system that boost the streamed rewards provided that the user meets the conditions for the milestone. Below 3 different pathways of rewards are depicted depending on user meeting the criteria for subsequent milestones and performing relevant onchain actions. Once a milestone has been reached, user receives a temporary increase in the rate of the reward acquisition.

Habit formation is another critical area that is completely underused in current questing product offerings. Since most current questing platforms have a “claim reward once” model, in our view it does not lead to habit formation, which is a key for successful loyalty programs (Henderson, Beck and Palmatier, 2011). Since pacts are focused on longer-term retention and actually creating a habit, we rely on three pillars of habit formation - intention, repetition, and context stability (Wood and Neal, 2009). Pacts’ reward systems incentivize the target audience to use the product for a longer period of time, thus reinforcing the association in their minds.

Sixdegree’s research says the following about airdrop strategy:

“We recommend protocols not to do one major airdrop and instead to have many microairdrops using data-driven strategy to build a strong holder base”

Replace “airdrops” with “rewards”; the following result is our core thinking for Pacts. Instead of one reward claim, we stream rewards and use milestones to guide user behavior.

Fighting adverse selection

Current questing models do not protect against adverse selection or sybil users since they are open for everyone - every user “applying” for a quest is approved. Sybillers know they are a lemon; therefore, quests are a positive expected value for them.

10k reward pool, 100 x 100$ rewards, 20000 participants

Value of sybil user - 0$

Expected Value (EV) of questing - 0.5$

Pacts will offer screening and whitelisting wallets depending on user types and onchain history. The screening conditions include a wide array of variables:

• Raw account onchain activity data

• Third-party reputation score software

• Usage of protocols

• Past pacts completion history

• User type based on past activity

Projects will be able to set the conditions for their pacts according to the audience they want to target. This means that the projects will be able to target lower quality users if they specifically want to, without having to pile in both high-value and low-value users in the same funnel. Siegenthaler, 2017, argues that allowing costless and non-binding communication in decentralized markets can result in much higher efficiency of market function. In our case, costless communication reveals the user type before the start of the promotional campaign. The paper also argues that there exists an equilibrium in which low-quality sellers truthfully reveal their type rather than try and mimic a high-quality seller. This is consistent with our hypothesis that separating different user types will result in higher efficiency of onchain marketing - projects willing to target lower quality accounts enter a sub-market of social boosting and lower level tasks. In contrast, projects willing to get power users will enter a lemon-free high-quality sub-market.

Changing the discovery

Pacts allow users to see what promotions are offered specifically to them and which ones are the closest % match based on their onchain history. This is a stark difference to the way the discovery works currently on questing platforms, where users are presented with thousands of similar-looking promotional campaigns that are open to everyone. Pacts are designed to connect projects to relevant audiences and show users promotions specific to their area of expertise and onchain profile.

Referrals

High-value users usually hang out with other high value users, whereas sybillers either don’t hang out (as they are bots and don’t understand the concept of “hanging out”) or are disincentivized to share quests with other real users, as more participants for the same reward pool decrease questers’ expected value.

Pacts utilize the power of referrals not only on the platform level, but on a pact level as well. Users will be able to refer their friends to Pacts platform and get lifetime share of all rewards that they acquire. Besides that, users will also be able to refer friends with similar onchain profile types to individual pacts as well, increasing their rate of reward acquisition.

Pacts market

Ultimately, we see Pacts platform as a marketplace for projects to bid for valuable users and for users to monetize their onchain history and activity. Current systems do not allow projects to adjust the desired reward depending on the value of the user account. However, there is no reason why the attention and time of users should be valued the same.

Letting lemons/sybillers dominate the market leads to fewer and fewer power users actually offering their time and activity to projects. To observe this effect, one can study any of the heavily sybilled airdrops that did not reward the actual early users of the protocol and the enormous sentiment drop following right after.

Besides being utilized as a marketing instrument, Pacts can also be used for customer development - pre-selecting the wallets and sending personal invitations to participate in product building. With all the data collected from the wallets, Pacts helps select strong users for targeted campaigns. After the campaign is launched, the statistics are displayed in an accessible dashboard, providing a no-code solution for marketers to analyze the output of the campaigns.

Pacts will also finally crypto marketers to do A/B testing, offering slightly different promotions to user groups to see which ones convert better or have higher retention rates. Pacts will also have in-built data dashboards, allowing to see and analyze pact performance without any coding work required from the marketer.

We are building Pacts to bring efficiency to this market, connect projects with power users, and increase the overall productive output of marketing budget spent.

Thanks for reading! If you like the way of our thinking, note that we are currently raising - reach out to hello@pacts.market to know more.

Follow us:

Twitter - @pactsmarket

Website - pacts.market

Citations

Akerlof, G. A. (1978). The market for “lemons”: Quality uncertainty and the market mechanism. In Uncertainty in economics (pp. 235-251). Academic Press.

Henderson, C. M., Beck, J. T., & Palmatier, R. W. (2011). Review of the theoretical underpinnings of loyalty programs. Journal of Consumer Psychology, 21(3), 256-276.

Melnyk, V., & Bijmolt, T. (2015). The effects of introducing and terminating loyalty programs. European Journal of Marketing, 49(3/4), 398-419.

Reichheld, F. F., & Schefter, P. (2000). E-loyalty: your secret weapon on the web. Harvard business review, 78(4), 105-113.

Siegenthaler, S. (2017). Meet the lemons: An experiment on how cheap-talk overcomes adverse selection in decentralized markets. Games and Economic Behavior, 102, 147-161.

Sixdegree, B. (2023). Airdrop & Liquidity Mining 03 - Liquidity https://blog.sixdegree.xyz/airdrop-liquidity-mining-03-liquidity/

Wood, W., & Neal, D. T. (2009). The habitual consumer. Journal of Consumer Psychology, 19(4), 579-592.