What happens to my business if a platform I rely on shuts down or changes the rules?

The cleanest case study in platform risk, worth knowing in its specifics. Google had offered small businesses free websites built from their Business Profiles: simple, convenient, and adopted by businesses across every category, exactly the kind of offer that feels safe because the vendor is enormous.

The sequence: Google announced the product's end, and in March 2024 the sites were turned off, with visitors redirected to the businesses' profiles. On June 10, 2024, the redirects stopped too, and the old addresses began returning errors. Businesses were advised to build new websites elsewhere, with Google suggesting a list of alternative builders, and Google Ads campaigns pointing at the dead sites stopped running until owners updated them.

Read the anatomy: no malice, no failure, just a strategy review that de-prioritized a free product, executed with corporate reasonableness, a notice period, a redirect window, migration suggestions. And for the businesses whose web presence lived there, the reasonable process was still the same event: the address on their business cards, their reviews' destination, their search equity, gone on a vendor's schedule. The lesson is not that Google is untrustworthy. It is that free-and-convenient surfaces are exactly the ones strategy reviews sacrifice.

Shutdowns make the news; the routine repricings and rule changes move more total value:

1. Repricing on captive customers. Subscription platforms raise prices precisely because leaving means rebuilding: the lock-in is the pricing power. Every renewal season brings examples, and the businesses that pay quietly are funding the leverage used against them.
2. Feature removals and forced migrations: the editor 'upgraded' out from under your workflows, the API deprecated, the plan restructured so your tier loses what you depended on. You adapt on their schedule, at your cost.
3. Policy and terms shifts: content rules, data terms, and acceptable-use changes that can strand a business model overnight, with appeal processes designed for their scale, not yours.
4. Algorithm repricing: the quietest and largest, reach and visibility reallocated by tuning decisions, feeds throttled, and an audience you built metered back to you at new rates.
5. Acquisition roulette: your vendor bought by someone with different plans, the product you chose becoming the product they inherit.

The common thread: each event is rational from the platform's seat and an unbudgeted tax from yours, and none of them requires the platform to fail. Healthy platforms do all of these.

Sort by a single question: if this surface vanished with ninety days' notice, what happens to revenue? The taxonomy that falls out:

Load-bearing, deserving owned ground:

1. The website: your primary machine-readable asset, the record engines and buyers verify. Owned files, portable hosting.
2. The list: the audience you can reach without permission. Exportable, backed up, on a provider you could leave in a week.
3. Client records and the business's captured knowledge: the method, the context, the history, in plain files you hold, whatever tools currently read them.
4. The method-instruments: diagnostics and tools that generate leads or deliver value, on your domain, in your code.

Legitimately rentable:

1. Echo surfaces: social profiles and distribution channels, valuable but replaceable, because the canonical version lives on your ground.
2. Commodity plumbing: payment processing, scheduling, email delivery, real dependencies, but swappable within weeks because they hold process, not assets.
3. Experiments: anything still proving itself deserves rented speed.

The failure mode is inversion: canonical content living in a feed, the only copy of client history inside one SaaS tool, the lead engine built in a platform's proprietary format. Rent where you can walk away. Own where you cannot.

One honest hour, four columns:

1. Inventory the surfaces. Every platform, tool, and service the business touches: website host or builder, email provider, social accounts, scheduling, payments, storage, the AI tools, the course platform, the CRM. Most owners list fifteen to thirty.
2. Score the blast radius: for each, write what breaks in the first month if it vanished with ninety days' notice, in revenue terms. Be literal: 'the website disappears' versus 'we switch schedulers with mild annoyance' are different universes.
3. Score the exit cost: could you leave in a week, a month, or only by rebuilding? The dangerous quadrant is high blast radius plus high exit cost, and everything in it is a standing emergency wearing a subscription.
4. Check the copies: for every asset in the dangerous quadrant, does an owned, current, usable copy exist? The list exported monthly, the content in your files, the client records in your storage, the site's code in your repository. A copy converts many dependencies from existential to inconvenient on its own.

The output is a migration list in priority order, and for most established businesses it is short: the site, the list's export habit, the knowledge capture, and one or two tools. A quarter of unhurried work retires the whole dangerous quadrant.

Owned core, rented edges, and copies of everything, which is cheaper and lazier than it sounds:

1. The core on your ground: the website as owned code on commodity hosting, the captured business knowledge in plain files, the list exportable and exported, the instruments in your repository. This is the layer where vendor decisions cannot reach revenue.
2. The edges rented cheerfully: distribution, plumbing, and experiments on whatever platforms serve best today, held loosely, because the canonical assets live elsewhere and every edge vendor is replaceable by design.
3. The copies as standing habit: exports on a calendar, the repository current, the knowledge files living where you and your AI work daily, so backup is a byproduct rather than a discipline.
4. The walk-away test as procurement policy: before adopting anything new, one question, how do we leave? Tools with clean exits get adopted freely; tools without them need to be worth the marriage.

The payoff is temperamental as much as financial: platform news stops being threatening, renewal negotiations happen from strength, and the next sunset announcement is a Tuesday. Getting the core stood up, your business captured into files you own, with an AI working on infrastructure you hold, is exactly what our AI Native Activation session establishes.