Why Blocking Services Sustains a Paid Connectivity Market

You might assume that a working block destroys the market for workarounds: if the ordinary service is unreachable, who would pay for a route that keeps changing? The assumption fails because blocking acts on both sides of the market at once. It raises what access is worth to a user and it raises what reliable access costs a supplier, and those two effects can support paid connectivity even while interference continues.

The myth: a successful block clears the market

A block changes the value of the blocked resource rather than deleting it. If people still need the service, some will pay for a path to it, plus setup help, a client that keeps working, and someone to call when it stops. That is a product description, not a conspiracy, and it explains why blocking and a market for workarounds coexist without either one disproving the other. That is the whole mechanism, and it needs no coordination to work.

The inverse claim is just as weak. A market existing does not mean the blocking is ineffective, and it does not mean every seller is tolerated. Demand and enforcement are different variables, and a single observation of either tells you little about the other.

Demand rises, and so does supply cost

A dated example shows why availability is a service rather than a feature. OONI's September 2022 Iran report documented blocking and outages during that month's protests. Some tested circumvention tools could bootstrap and fetch a small page, and the report explicitly noted it lacked evidence of sustained effectiveness over long periods.

That limitation is the interesting part for anyone reasoning about a market. Maintaining access can involve endpoint capacity, distribution, support and a response to each change in interference. Those are recurring costs, and recurring costs are what turn a free workaround into a paid offering without proving that the offering is good or safe.

A successful test is not a product

Bootstrapping and fetching one page exercises a narrow slice of the path: name resolution, a handshake, a request, a response. It does not exercise the hours in which interference changes, the client update that follows, or the support conversation when a user cannot connect at all. Treat a short test as evidence about that test, not as a forecast about a month.

This is also why a provider that rotates endpoints is not automatically reliable. Rotation is one operational response among several, and each response has its own cost and its own failure mode.

Encryption is not unblockability

Encryption protects content from an observer who cannot interfere with the path. It does not prevent interference with the path itself. A network can make an endpoint unreachable, disrupt protocol delivery, or degrade the underlying connection without reading a single byte of the payload, and none of those actions require breaking the cipher. Availability is therefore a separate property that has to be tested on its own.

That distinction matters when a seller markets encryption as if it were availability. The two properties are separate, and only one of them is what a censored user is usually buying.

What would actually count as evidence

Notice that each row needs a different kind of source. Reachability data cannot establish ownership, and ownership records cannot establish whether a service stayed usable. Mixing the rows produces confident conclusions that the underlying evidence never supported.

The defensible version is narrower than either headline: restrictions raise the value of reliable access, interference raises the cost of supplying it, and a market can persist in the gap. That explains the persistence without assuming an unblockable technology or accepting an allegation about every seller.