Why deariary has no streaks
deariary has no streak counter. No badges, no levels, no weekly consistency score, no notification saying you have not written in three days. There is no number anywhere in the product that goes up when you show up and resets when you do not.
That is a deliberate decision rather than a missing feature, and the reasoning is specific enough to be worth writing down.
Streaks do work, which is the awkward part
It would be easier to argue against streaks if they were ineffective. They are not. A visible counter raises the psychological cost of stopping, converts an abstract intention into something with a running score, and reliably increases how often people open an app. Every product in this category ships one for good reasons.
We have written separately about why a system that needs a streak to survive has not solved its own fragility. This post is about three problems that remain even when the streak is working exactly as designed.
Problem one: the mechanic punishes the periods that matter most
Ask when a diary is most worth having later.
The week a parent went into hospital. The first month with a newborn. A move to another country. The stretch after a breakup. The two weeks of a genuine crisis at work. A long illness.
Now ask when a streak breaks.
It is the same list. Those are precisely the periods with no spare capacity for a nightly ritual, which means the mechanic imposes its penalty exactly when the record would have been most valuable. You lose the entries and you get a reset counter as a souvenir.
A diary that quietly stops working during the important parts of a life has an inverted design. Not slightly suboptimal: pointed the wrong way.
Problem two: the streak becomes the thing being optimised
This one shows up after the mechanic succeeds.
Once a counter exists, it is the most legible thing in the product, and legible metrics attract effort. So the goal shifts from having a record to keeping the number alive. And the cheapest way to keep the number alive is to write the minimum that counts.
“Good day. Tired. Long meeting.”
That is a streak-preserving entry. It took nine seconds, it satisfies the mechanic, and it contains nothing. Do it for six months and you have a 180-day streak and a diary made of nothing, which is a worse outcome than 40 real entries and no counter.
The mechanic did not fail here. It worked, and the record got worse because it worked. Any measure that is easy to satisfy in a degenerate way eventually gets satisfied in the degenerate way, and “did you write today” is trivially easy to satisfy degenerately.
There is a related effect we have written about: the act of tracking changes the behaviour being tracked. For a diary, that is a problem rather than a feature. A record you are performing for a counter is not a record of your life.
Problem three: in an automatic diary a streak would measure nothing
The third reason is almost arithmetic.
Entries in deariary are generated overnight from what your connected services logged. They appear whether or not you open anything. So what would a 400-day streak actually represent? That the pipeline ran. That the servers were up. That you continued to exist and use a calendar.
Displaying that as an achievement would be dishonest in a small but real way. It would take credit, on your behalf, for something you did not do, and then use the resulting number to make you feel invested in a product. Plenty of software does this. We would rather not.
The one place where not showing up has a consequence
There is an exception and it should be stated plainly, because a post like this loses its point if it hides one.
On the free plan, an account with no activity for 30 days is treated as dormant and generation stops until you come back. A warning email goes out before that, seven days ahead, once per inactive period.
Two things about it. First, what it measures is whether you opened the product, not whether you wrote anything, because there is nothing to write. Second, it exists because generating entries costs money on every one of those days, and generating them indefinitely for an account nobody is reading is not sustainable on a free tier. It is a cost control.
You could reasonably call that an engagement mechanic wearing a different hat. The distinction we would draw: it is not designed to make you feel you are losing something you built, and it does not display a number. But we are not going to pretend the email does not exist.
The metric we would actually want does not exist
The thing this product is for is the moment six months from now when you read an ordinary Tuesday and the day comes back to you.
That is not measurable. There is no event to log. So the choices are to ship a proxy for it or to ship nothing, and the available proxies are all wrong: entries written measures compliance, days opened measures habit, streak length measures neither. Each of them would drift the product toward optimising something other than the thing it is for, because whatever you display is what people and teams start improving.
So there is no number. The entries accumulate, nothing counts them, and if you do not open the app for two weeks nothing has been lost, because nothing was depending on you in the first place.
That is the whole feature. It is best understood as the absence of one.
