Bundles & lock-in have lost their teeth
I just cancelled a bunch of services with enough data lock-in that I never thought I’d get away.
Leaving them was always possible, just never worth it.
But yesterday, I spent four hours on DIY data exodus to cut my monthly subscriptions in ~half (“hey claude, can you figure out how to do a lossless export of everything I’m using this platform for, with no downtime and safe rollback?”).
If others are doing this sort of thing — and I have to imagine they are, since the ROI was splendid — then it’s a major hit to the biz model concepts of data lock-in and feature bundling.
For example: I’m no longer using Circle. Although this post is published via Circle as a final heads-up for anyone subscribed, it’ll land on a static Jekyll site containing all historical content at the same URLs, plus RSS. I used to benefit from Circle’s whole “bundle” and was happy to pay. But now I’m just using it out of inertia as a pricey CMS. Migrating would have cost me a week, which never felt worth it. But now it is. Not because they made it easy (it still required a data dump + API + scraping), but because Claude made it easy.
I’ve also noticed myself picking apart bundles. If 80% of a bundle is trivially re-creatable, I’m finding myself no longer valuing that 80% in my mental math. I’m still happy to pay for the killer features, but the nice-to-haves have stopped seeming so nice. (For example, I’m currently in the process of unbundling my accountants.)
Webapps are feeling increasingly like friction that I’m working around in an attempt to get at the one thing the service does uniquely well. More and more, I’d rather pay a premium for narrow tools with clean APIs and good integrations. Give me an itemized invoice of services I can mix and match, not a monolithic bundle.
Yeah, that’s supposed to be suboptimal from a business model perspective. But is it? Now that I’m off Circle, I’m not going back. Whereas when I went to do the same cleanup to my convoluted registrar/dns setup, DNSimple gave me a lovely itemized invoice, and I was able to cut my bill by 70% while remaining a happy customer. When I’m ready to scale my spend back up, it’ll be with them.
Unbundle yourself, or your customers will do it for you.
A buddy of mine who runs a ~20m/y bootstrapped software business said their customer feedback has started becoming a lot more demanding. It’s a shift in tone and expectation. Customers are waking up to the fact that they can now DIY a lot of what they used to have to accept as an annoyance of the platform. The baseline used to be that your product could do something valuable. Now it has to be able to do that thing better than Claude.
It’s not just webapps. I spent ages tweaking my OpenClaw to be safe and useful, so much so that I thought I was going to be locked into it forever. But I had its full state already sitting in a git repo, and it turned out to be a ten-second job to losslessly hibernate it and cancel the VPS.
The whole assumption behind subscriptions is that folks aren’t gonna unsubscribe on a whim. Subscriptions are the beneficiaries of inertia, friction, and convenience.
It’s always been far easier to keep paying than to stop.
What happens when it isn’t?