Part 6 of After the Feed, a field guide to how social media gets rebuilt. See all parts →
We built the thing in Part 5, and did it without an ad engine. Which raises the obvious objection: advertising exists because it answered a genuinely hard question — how do you pay for a free product used by millions of people? If you throw the ad model out, you need a real answer, not a wish. Here’s the answer, and why it’s not as fragile as it sounds.
Why the ads have to go — it’s structural, not moral
The point isn’t that ads are evil. It’s that the ad model is the feed model. You can’t keep one and drop the other.
Surveillance advertising pays out in proportion to attention and data, so it rewards whatever maximizes both: infinite scroll, autoplay, notification bait, a feed tuned to agitate. Every dark pattern from Part 1 is a rational response to that incentive. A network built on trust and calm can’t run on a revenue model that pays it to be neither. So the money has to come from somewhere that gets better when the product gets better for you.
There are three such places, and two of them are already proven (Part 3).
Stream 1: subscriptions for what you actually keep
Charge for the archive, not the access. Posting and viewing stay free; what you pay for is durable, high-quality storage of your originals — full-resolution, uncompressed, yours — plus the nicer creative tools and a custom handle.
This isn’t speculative. Retro already sells almost exactly this at around $36 a year and just raised a $21M round on the strength of it. People will pay a modest recurring fee for a photo product that respects them. The subscription both funds the storage it pays for and aligns the business with the user: you’re selling them a better version of their own stuff, not renting them to a third party.
Stream 2: turn pixels into objects
This is the one advertising literally cannot do: sell a physical good. The tactile-artifact opening from Part 4 is also the cleanest revenue line in the whole design.
Monthly print boxes, photo books, one-off postcards mailed anywhere — real products with real margins. Retro’s postcards (a couple of dollars a card) and Fujifilm’s billion-dollar Instax business both prove the demand is there and people pay happily, because they’re getting something they can hold. Every print sold is revenue that arrives because the user valued a memory enough to make it permanent. The incentive points the right way.
Stream 3: a creator market
Once the network is open, let professionals build on it — photographers selling color/development profiles, curators running verified feeds — and take a standard platform cut. It’s low-overhead (you’re running a marketplace, not an ad exchange) and it deepens the product without the company having to build every feature itself. Think of it as the Telegram-bots lesson from Part 2, monetized.
Does the math actually close?
Here’s the part worth being honest about: the following is an illustrative model, not a forecast. But the shape of it is what matters.
A rough model, at ~100,000 engaged users. An unbundled network is cheap to run compared to an incumbent. You’re indexing structured records off an event stream — not running billion-item recommendation models over every asset — and media is served from your own storage through ordinary caching. Blended infrastructure lands in the ballpark of a few cents per active user per month. Against that: even a few percent of users on a ~$4–5/month plan, plus print orders at healthy margins, plus a marketplace cut, comfortably clears the cost — with gross margins that look like software, not media.
The reason those numbers can work where an ad business would need scale in the hundreds of millions is exactly the unbundling from Part 5. The expensive thing about big social platforms — the giant, always-on recommender crunching everything for engagement — is the thing this architecture doesn’t have. Take out the attention engine and the cost base collapses to roughly “storage plus indexing,” which usage-based subscriptions and physical goods can cover directly.
The alignment is the moat
Step back and the real advantage isn’t any single stream. It’s that every revenue line gets healthier when the product serves you better. Better archives → more subscriptions. More meaningful moments → more prints. A richer ecosystem → more marketplace activity. There’s no point at which the business benefits from making the product worse — which is precisely the trap the ad-funded feed can never climb out of.
Is it proven at platform scale? Not yet. The closest live proof — Retro — is still measured in low millions of users, not hundreds of millions. But it’s real, it’s growing, and it’s profitable-shaped doing exactly this. The bet of this whole series is that “small, trusted, and sustainable” is a better starting point than “huge, exhausting, and cornered.”
So where does it all actually go? Last part.
Next in the series: Part 7 — where this goes next: the roadmap, and the honest hopeful, grim, and boring-but-likely futures. See all parts → · Get new parts by email
References
- Retro pricing and model: Engadget; $21M Series A
- Physical-goods demand: Fujifilm Instax sales to exceed $1 billion
- On why the ad model shapes the feed: Part 1 references
Comments