Commentary April 2024 8 min GDA Research

01 Play-to-earn economies funded rewards through emission of the same asset the reward was paid in — structurally self-referential.

02 Every such system depends on net new capital inflow to hold reward value. When inflow slows, the reward devalues and engagement collapses with it.

03 Denominating rewards in real-world assets, experiences, and services sources reward value externally, from partners with their own customer acquisition budgets.

04 This converts the reward from a dilution cost borne by holders into a marketing cost borne by merchants — which is how conventional loyalty has always worked.

05 The resulting business is a distribution channel, not a game, and should be valued as one.

The structure was consistent across the cycle. A game or application issues a native token. Users earn the token through participation. The token has value because it can be sold. It can be sold because new participants are buying in.

The reward and the funding mechanism were the same instrument. There was no external revenue underwriting the reward pool, only the issuance schedule. That is not a criticism of any particular design; it is an accounting observation that applied to nearly all of them.

Such a system is stable only while inflow exceeds outflow. The moment participation growth slows, reward value falls, which reduces the incentive to participate, which accelerates the decline. The failure was not gradual and it was not a matter of execution quality.

It is worth being precise about what went wrong, because the diagnosis determines the replacement.

The problem was not that users were speculating. Loyalty programmes have always had a speculative component, and airline miles are traded, hoarded, and arbitraged without destroying the programme.

The problem was that the issuer had no external source of value with which to honour the reward. An airline can honour a mile because it has a seat. A retailer can honour a point because it has inventory. A play-to-earn issuer honoured its reward by finding another buyer for the reward, which is a different arrangement entirely and one with a known endpoint.

Denominating a reward in real-world value — assets, experiences, services, and goods — restores the structure that makes conventional loyalty durable.

The reward inventory is supplied by merchants and partners who are acquiring customers. They fund it from marketing budgets, because that is what it is. The reward has a floor set by the cost of the underlying good, not by the current bid.

The economics invert accordingly. Under emission, the reward is a dilution cost borne by existing holders. Under external sourcing, it is a customer acquisition cost borne by the merchant. The first is extractive and finite. The second is the model behind every large loyalty programme in existence.

There is a second-order effect. Because the reward is redeemable for something specific, its perceived value is stable and legible to a user who has no interest in market prices. That widens the addressable audience well past the population willing to hold a volatile asset.

Where play-to-earn used gameplay as the earning mechanism, prediction is a materially better one, and the reason is engagement quality rather than novelty.

Prediction requires a view. A user who commits to an outcome has demonstrated attention and judgement, not merely time spent. That is a higher-value signal for any partner buying access to the audience, and it is more resistant to automation than gameplay, which was systematically farmed.

Prediction is also natively episodic. It attaches to scheduled real-world events — sporting seasons, elections, earnings — which supply a recurring calendar of engagement without the issuer manufacturing artificial scarcity.

The correct comparable set for a rewards network denominated in real-world value is not gaming. It is loyalty and performance marketing.

The revenue question is what merchants pay for qualified, attentive users. The cost question is what reward inventory costs to source at scale. The growth question is how many earning surfaces — play, predict, engage, attend, watch — feed a single balance.

None of those questions require a view on digital asset prices, which is the point.

Funding source Emission: token issuance · External: merchant marketing budget

Reward floor Emission: current bid · External: cost of underlying good

Cost borne by Emission: existing holders · External: acquiring merchant

Stability condition Emission: net inflow positive · External: merchant demand for users