The Channel Someone Owns
Nobody sits down and decides to ruin a mailing list. It happens one small addition at a time: the pottery class newsletter picks up a "partner spotlight," then a monthly roundup, then a survey. Nobody voted on this. There was no meeting where someone said let's extract more value from the people who trusted us. It just happened, the way water finds the low point in a yard without anyone digging a channel for it.
The reason it happens is boring and structural: the moment you opt into something, you've handed someone an asset. A list of people who said yes to X is not just a list of people who want X. It's inventory. It's reach. It's a thing with a market value the moment enough people are on it, and market value is gravity. Anyone holding that list, however good their original intentions, is now sitting on a resource that's worth more the harder it works. That pressure doesn't require a villain. It requires only that the channel outlive the specific promise it was built on, and channels almost always do.
Compare this to a plain resource nobody owns. A directory of links with no ad slots. A friend's personal mailing list they run for a dozen people because they feel like sending updates. An RSS feed with no monetization layer sitting between the writer and the reader. These stay clean, and they stay clean for a reason that has nothing to do with the operator's character. There's no lever to pull. You can't extract more from a channel that isn't an asset to begin with, because there's no buyer for a dozen friends' inboxes and no advertiser interested in one person's feed. The cleanliness isn't a virtue being practiced. It's a market failure that happens to protect the subscriber.
This is why every fix aimed at the symptom instead of the incentive fails on a delay. Opt-in checkboxes, granular permission toggles, "you can always unsubscribe" disclaimers: all of it assumes the problem is consent captured badly at the start, so better consent architecture will hold. But the drift doesn't come from bad consent capture. It comes from what happens after consent is captured, when the channel sits there being worth something and somebody's job is to grow what it's worth. You can architect the checkbox as carefully as you like. You can't architect away the fact that six months later there's a person whose quarter looks better if the list does more than it was built to do.
Put the asset in different hands and the pattern repeats regardless of who they are. A nonprofit's donor list drifts into a co-marketing list. A neighborhood app's opt-in alerts drift into local business promotions. An open source project's mailing list, once it gets big enough, drifts into a newsletter with a sponsor slot. None of these operators are unusually greedy. They're operators of an asset, and assets get worked. The honest version of "we would never sell your data" is "we haven't yet found a buyer we're comfortable with," which is a schedule, not a promise.
So the useful question, when you're handing over an opt-in, isn't whether you trust the people running the channel today. It's whether the channel is structurally capable of becoming valuable enough to be worth extracting from. A feed with no ad model and no subscriber count anyone's optimizing against will probably still be a feed in five years. A list that's growing, that has a name, that shows up in someone's quarterly deck, is already partway to becoming something else.
The exits are rare exactly because the internet keeps building better tools for turning attention into an asset, and an asset with nobody positioned to cash it in is now the unusual case rather than the default one. Worth noticing which of your opt-ins are still just lists, and which ones somebody's already counting.