You already paid for your trigger insights. In data.

The lock icon over your own patterns

Ninety days in, the app finally had enough logs to say something. You open it on a Tuesday morning, still photophobic from Sunday's attack, neck stiff, running on the kind of foggy half-sleep that follows a bad one. And there it is: a chart of your triggers, blurred behind a lock. Fifty dollars a year to read what your own tracking built.

You did the work. You logged through the aura, through the postdrome, on the days when looking at a screen was its own small punishment. The correlation the app is now selling back to you was assembled entirely from data you entered by hand. The question worth asking before you tap Subscribe is a plain one: what, exactly, are you being charged for?

The deposit-and-paywall pattern

Here is the shape, named so you can spot it in any app you try. The tracking is free while you deposit data. The insight arrives right when the deposit is large enough to be useful, and that is the moment it locks. You are not paying for a feature. You are paying to unlock the value of your own accumulated logging, at the point of maximum sunk cost.

This isn't intelligence you're renting. It's arithmetic you already own. Finding which foods, weather shifts, sleep gaps, or hormonal windows line up with your attacks is a correlation problem, and correlation is one of the oldest, cheapest computations in statistics. It does not need a language model. It does not need a data center. It runs on the phone already in your hand, in the time it takes the screen to redraw.

Why 'premium insights' is the wrong frame

The subscription framing leans on a quiet implication: that turning your logs into patterns is heavy, ongoing, valuable machinery, the sort of thing that reasonably costs money every month to keep running. That implication is doing the persuading, and it does not survive contact with how the math actually works.

A trigger correlation is computed once against a fixed dataset, then re-run cheaply when new entries arrive. There is no recurring cost that scales with a monthly fee, because there is no server doing anything a subscription pays to keep alive. When an app charges annually for insights, the annual part is a business model laid over a computation that has no annual cost. The recurrence is chosen, not incurred.

What the analysis actually is

Strip away the packaging and trigger detection is a handful of well-understood operations. Count how often a candidate trigger and an attack co-occur. Compare that against how often each appears on its own, so coincidence doesn't masquerade as cause. For a phase-aware app it means widening the window: a trigger on Monday can surface as an attack Tuesday and a postdrome tail into Thursday, so the correlation has to look across lag, not just same-day overlap.

None of that is exotic. It is the same statistical toolkit a spreadsheet runs, applied to a timeline you built. The honest version of an insights screen is not a black box you rent access to. It is a calculation you could, in principle, check yourself, performed on data that never has to leave your device.

How it plays out over ninety days

Watch the timeline and the leverage becomes obvious. Weeks one through four, you log diligently, because tracking through a chronic condition is a habit you have to earn. Weeks five through eight, the app has enough to hint at patterns but shows you a teaser, sparse, tantalizing. By week twelve it declares your data ready and drops the paywall, and by then walking away means abandoning three months of hand-entered history that only reads inside that app.

That last part is the hinge. If your ninety days of logs exported cleanly and opened somewhere else, the fifty dollars would be optional. The fee is only enforceable because leaving is expensive. The paywall and the lock-in are the same mechanism seen from two angles.

The fee is only enforceable because leaving is expensive.

What changes once you see the cost for what it is

If the computation is free and the data is yours, then everything about the recurring charge is a decision the company made, and decisions can be made differently. A trigger analysis can be a one-time purchase, because it is a one-time cost. Your logs can export in a format another tool can read, because there is no technical reason to trap them. Neither of those is a concession. They are what the underlying facts already imply, once nobody is dressing up arithmetic as ongoing machinery.

So the useful test for any app you're considering isn't how polished the insights screen looks. It's two questions the marketing rarely answers on its own: does the analysis run on my device, and can I take my data with me when I leave.

The version that charges once

We built Postdrome on the other side of that math. The trigger analysis runs on your device, against your own continuous timeline, headache through the 24-to-72-hour postdrome tail and not just the attack itself. It is a single one-time purchase rather than a meter that keeps running, because the computation does not keep costing us anything after you've paid. And your data exports in a portable format, so the decision to stay is never enforced by the cost of leaving.

That is not a generosity we're advertising. It is the pricing that honestly matches what the work is. A recurring fee for on-device correlation would require us to pretend the arithmetic is more expensive than it is, and we would rather show you the patterns you built than charge you rent on them.

What we're watching next

The pricing question is downstream of a data-ownership question, and that one is getting louder. As more of this class of app moves computation on-device by default, the justification for charging monthly for local analysis gets thinner every release. We're tracking how export standards evolve too, because the day migraine logs move as freely as a calendar file is the day insight paywalls lose their grip entirely.