When Collecting Becomes Infrastructure

Why digital collectibles need an engine, not a destination

In earlier essays I argued that a brand can issue a Set the way a mint issues coins (Issue a Collectible), that a Card can be something a person keeps rather than something they scroll past (Kept Media), and that a Card is becoming a small piece of software (When a Card Becomes Software). Those essays were about why collecting matters.

This one is about where collecting lives, and what has to sit underneath it. The answer I have come to is not the one I started with. I began by assuming collecting needed a place. I now think it needs an engine.

1   The Case for One Destination

The obvious way to build digital collecting is to build one destination. Brands issue Sets; collectors come to one site to open packs, fill their Albums and trade duplicates. It is how the sticker album worked in every school playground: one album, many series, and one place where the swapping happened.

The destination has real advantages, and they deserve to be stated plainly before I argue against it.

  • One Album, many brands. A collector keeps Cards from a publisher, a bank, a museum and a football club in one place. The collection grows faster than any single brand could make it grow.
  • Virality across brands. A collector short of one Card asks a friend, who collects a different brand’s Set and discovers it in the process. Trading spreads the habit sideways.
  • People who come for one Set find another. Small issuers borrow an audience they could never build alone.
  • One habit, not twenty. Nobody wants twenty collecting apps. One daily visit is easier to form than many.
  • Network effects. Every new issuer makes the destination more valuable for every collector, and every collector makes it more valuable for the next issuer.

A shared destination turns many small collections into one large habit. That is its power, and it is not a small one.

2   Why Brands Still Say No

Ask a brand where it wants its collectible Cards to live and the answer is almost always the same: on our website, in our app, in our emails. Not somewhere else, however good that somewhere else is.

The reasons are sound.

  • The visit. The point of collecting, for a brand, is that people come back. A collector who returns every day to a third-party destination is returning to someone else. The brand has paid for attention and handed the return trip to another platform. This is the rented-attention problem I wrote about in From Rented to Owned, in a new form.
  • The relationship. A brand wants the collector’s activity, preferences and history inside its own relationship with the customer, not as a report from a partner.
  • The experience. Brands spend years on how their properties look and feel. A collection that lives elsewhere wears someone else’s design.
  • The content. Many brands do not own their best content outright. Characters, players, images and archives often come with licences that forbid hosting on another company’s consumer platform.
  • The company it keeps. On a shared destination, a brand’s Set sits next to its competitor’s.

Brands want collectibles to bring people home, not send them away.

That appears to leave a brand with two poor choices. It can join a destination and give up the visit, the relationship and some control of its content. Or it can build a complete collecting system itself, and become a software company for something that is not its core business.

White-label software offers a partial answer. It puts the brand’s logo and colours on a ready-made experience, and for many brands that will be enough. But issuers with strict licence terms, distinctive gameplay or established apps of their own want to design the whole experience, not reskin someone else’s, and still not rebuild the machinery underneath. For them, a third option is needed.

3   A Card Is More Than an Image

The third option starts with seeing what a digital Card is made of. It looks like a single object. It is three.

Figure 1. A Card has three layers. The issuer controls the first two; an engine can maintain the third.

  • Content is what the Card depicts: the artwork, the character, the story, the licence that governs its use. The issuer controls it.
  • Experience is where and how the Card appears: a website, an app, an email, a game, a QR code on a museum wall. The issuer designs it.
  • State is what the system remembers: which Card was issued to whom, who holds what, which are duplicates, what has been traded, which Sets are complete, and the history of all of it.

Most of the effort in building a collecting product goes into the third layer, and almost none of the brand’s identity lives there. The character on Card 17 belongs to the issuer, or to whoever licensed it to them. The fact that a particular person holds Card 17, received it on a Tuesday, traded a duplicate for it and completed the Set the following week is a record. It has to be kept accurately, for every collector, for years.

The artwork is what people see. The state is what makes it last.

Once the layers are separated, many licensing problems become manageable. The artwork never needs to leave the issuer’s own systems. The engine holds identifiers, rules and records. The issuer’s app asks the engine what a person holds, and shows the artwork itself.

Not every problem, though. Some licences forbid any third party from processing records about collectors, or restrict particular digital uses of a character altogether. An engine has to be able to run where the issuer requires, including inside the issuer’s own environment. Separating the layers is an architectural option, not an automatic clearance of rights.

One qualification matters. Holding a Card is an entitlement under the issuer’s terms, not ownership of a copyrighted character. A collecting engine must respect the limits of the licence: Sets that expire, artwork that must be withdrawn, Cards that cannot be traded outside a community. The record of a person’s progress can outlive the artwork, but only within the terms the issuer sets.

4   The Engine Underneath

Separate the layers, and a new kind of product appears: one collecting engine under many independent experiences. It matters who holds what, so it is worth being precise.

The issuer controls the content and the experience. The engine maintains the state and enforces the rules. The collector keeps the entitlements and history the issuer has granted.

The engine is the custodian of the record, not its owner. If a brand left the destination for an engine only to become dependent on a different vendor, nothing would have been gained. So exit has to be part of the design: an issuer must be able to take its complete records elsewhere and replace the engine without erasing a single collection. That holds even where Cards can never move between issuers.

Other categories have made this move. Few companies build their own payment processing; they call a payments API. Few build their own search; they add a search engine to their catalogue. Commerce has steadily separated the shop front from the machinery behind it. Each time, the brand kept control of what customers see and stopped maintaining what they do not.

Collecting is ready for the same separation. A publisher could offer its historic front pages as Cards inside its own app. A football club could issue player Cards on its website. A museum could let visitors collect artefacts by scanning codes in its galleries. A school could issue Cards that record what a student has learned. A consumer brand could build a Set around its category and run it inside its own emails. Five different experiences; one engine underneath.

The engine has to get several hard things right:

  • A claim that cannot create unauthorised extra Cards, even when the same request is repeated or replayed. Duplicates are part of collecting; accidental ones are not.
  • Trades that complete fully or not at all, so no Card is ever lost or duplicated in transit.
  • Odds that are fair, and that can be audited when a prize depends on them.
  • Rights that are enforced in the system, not just written in a contract: where a Card may appear, whether it may be traded, when a licence expires, what happens after a takedown.
  • A history that is durable and auditable through upgrades, traffic spikes and the occasional crash, kept for as long as the issuer’s terms and the law permit, and deleted when privacy rules or an expired licence require it.

Displaying a Card is easy. Keeping a million collections correct is not.

5   Why AI Makes the Engine More Valuable, Not Less

The obvious objection is that AI has made software cheap. A developer can ask a coding agent for a collectible-card app and have a working demo in a weekend. Why would anyone buy an engine?

Not because the code is hard to write. AI is making both the interface and the engine cheaper to build, and a capable team could generate the logic for claims, trades and progress quickly. The difficulty is what comes after: testing, securing, monitoring and operating that code, handling fraud and bots, recovering from failures, enforcing licence expiries and keeping up with changing rules on prizes and privacy, for as long as people keep their collections.

Code can be generated once. Ownership records, licence rules and collectors’ trust have to be maintained for years. The scarce capability is not writing the engine; it is accepting continuing responsibility for the correctness of millions of changes to what people hold.

The advantage of shared infrastructure is not that others cannot build it. It is that most issuers should not need to keep operating it.

There is a further point. Because AI makes the interface cheap, every issuer can afford a distinctive experience of its own. That is exactly what an engine needs: many different front ends, all relying on the same hard machinery underneath.

When the interface becomes cheap, the value moves to what sits underneath.

An engine of this kind needs two ways in. Its APIs serve live applications: issuing, claiming, trading, completing. The Model Context Protocol, an open standard for connecting AI agents to tools, serves something new: an issuer’s own agents, which can create Sets, schedule releases, check performance and help developers integrate. An issuer should be able to set up a collection by describing it to an agent, the way a developer today sets up payments by calling an API.

6   The Economics of an Engine

A destination and an engine are different businesses. A destination has to attract collectors and then earn from them. An engine serves issuers who already have audiences and content. The issuer brings the demand; the engine supplies the capability.

There are several reasonable ways to price it: a platform fee plus a charge when a collector first activates a Card, a price per active collector, annual volume commitments, or a licence for a deployment the issuer runs itself. Different issuers will prefer different units. Whatever the unit, one principle should hold.

The issuer should not pay more simply because a collector returns to something they already hold.

The whole point of collecting is that people come back to what they have kept. A price that rises with every reopened Card would punish the behaviour the issuer most wants.

The real economic test is not the price. It is whether each new issuer is cheaper to launch than the last. If every new customer needs almost as much engineering as the one before, the engine is a services business in disguise, a point I have made about software more broadly in the Software Foundry essays.

There is a second advantage that only an engine can build. Across many issuers it can learn which Set sizes people complete, which odds keep them interested, where collections stall and how long novelty lasts. That learning must be aggregated, privacy-safe and agreed with each issuer; no issuer’s data or content should ever be visible to another. Done properly, it is the loop I described in Martech: The Moat Is the Loop: learn across issuers, apply within each.

The hundredth issuer should launch faster, and do better, than the first.

7   The Shelf and the Album

The engine solves the brand’s problem. It does not, by itself, keep the advantages of the destination. A collection that lives only inside one brand’s app is a walled garden. There is no Album spanning brands, no trading across Sets, no discovery.

Figure 2. A destination gives a network but sends people away. An engine keeps people at home; a shared Album connects collections only where issuers permit it.

The way to reconcile the two is to separate two words that are usually treated as one. A Shelf is what a brand’s property shows: its own Sets, its own Cards, in its own design. An Album is the person’s collection: everything they have collected, from every issuer that allows it.

With an engine underneath, the Album becomes an option rather than a condition. An issuer can permit its Cards to appear in a collector’s personal Album. The collector can choose to bring them there. What travels is the record of what the person holds, not a copy of the artwork; the Album shows what each issuer permits, and links back to the issuer’s own property. The shared Album can send people home rather than away.

These are separate permissions, not one switch. An issuer might allow its Cards to be shown in a personal Album but not traded there, or allow trading within its own community but never across issuers. Some issuers will allow nothing at all, especially those with licensed characters. That is fine. The engine must be useful without the network. But where issuers and collectors agree, some of the destination’s advantages return: cross-brand collections, discovery and, where permitted, trading that spreads the habit.

The Shelf belongs to the issuer. The Album belongs to the person.

This changes the order in which collecting can grow. A destination has to attract issuers and collectors at the same time before it is worth much to either; fifty disconnected Sets and a thousand silent accounts do not make a network. An engine is useful from its first issuer. It can grow one issuer at a time, and connect collections later, wherever permission exists.

Figure 3. A destination needs both sides at once. An engine can grow one issuer at a time.

Collecting may need shared infrastructure before it can support a shared network. The engine can succeed without the network. The network, if it comes, may arrive through the engine.

8   What Would Prove This Wrong

The argument stands or falls on one test: can three very different issuers launch three very different collecting experiences on the same engine, without anyone writing three different products? If the third launch is much easier than the first, collecting is becoming infrastructure. If every launch needs its own code, it is a useful service, not a category.

Four things would prove the thesis wrong:

  • Issuers build simple versions themselves and never outgrow them. If a basic in-house system is good enough for most brands, the engine is a niche.
  • Nobody wants to collect. An engine makes collecting possible. Only the content, the Sets and the experience make it worthwhile. Excellent infrastructure under an experience nobody wants is still an experience nobody wants.
  • Licensors refuse even the record. If rights holders will not let anyone else hold even the state of who owns what, the separation of layers fails where it matters most.
  • Collecting is too varied to share an engine. If a card game, a learning journey and a museum trail need fundamentally different machinery, there is no common layer to build.

I would watch the first three issuers closely. The answer will come from them, not from argument.

Every Property a Place to Collect

Every company with stories, characters, knowledge or history can turn them into things people collect and keep. A publisher has its archive, a club its players, a museum its artefacts, a brand its category. The collectible should live where the company’s audience already is.

That does not mean every company should build and run a collecting system. The issuer controls the experience. The collector keeps the history of participation, within the issuer’s terms. The engine exists to make both possible, and to connect collections only when everyone involved agrees.

The future of collecting is not another destination. It is the engine that lets every brand’s own property become a place to collect.

Published by

Rajesh Jain

An Entrepreneur based in Mumbai, India.

Leave a Reply