Designing the cogitats economy

A free player of Cora's Atlas earns cogitats by solving puzzles — roughly thirty-nine per puzzle on the median — and spends them on hints when the answer will not come. An Atlas Pro subscriber gets a small monthly stipend on top and pays thirty percent less per hint. Hint costs are forty, eighty, and one hundred and twenty cogitats. They used to be ten, twenty, and thirty. This dispatch is about why we changed them, and why the price of a nudge is a UX problem before it is a monetisation one.

1. The premise

The job of a hint, in a narrative puzzle game, is to be the second-worst option. The best is to solve the puzzle. The worst is to put the device down and not come back. The hint sits in the middle, for the player who has stared at the screen for nine minutes and is about to close the tab.

If a hint is too dear, no one buys it, and we lose the player at minute ten. If a hint is too cheap, everyone buys it, and we lose the player at the end of the chapter, when they realise they never actually had a relationship with the puzzles. They had a relationship with the help button. The puzzle was wallpaper.

Hint pricing is a UX problem first. The monetisation is a consequence: a player whose hints feel earned keeps playing, and a player who keeps playing might one day choose to support the small team in Paris that draws the maps.

2. The first version: ten, twenty, thirty

The first cogitats economy we shipped priced hints at ten, twenty, and thirty. The earn rate, then as now, was about thirty-nine cogitats per puzzle. The arithmetic was inviting: one puzzle's earnings paid for a whole cascade of hints on the next one. A free player finishing Vellestria at full completion exited the nation with more than two thousand surplus cogitats unspent — about twenty-five puzzles' worth of dead currency, sitting on the balance sheet doing nothing.

We watched the per-puzzle hint-burn rate for several weeks. The pattern was uniform across categories. Hints were not being used at the moment of genuine confusion. They were being used at minute three, the way a fidgety reader peeks at the last page of a novel because they can. The hint had become a free help button. The button worked. The puzzles, increasingly, did not.

3. The audit

We ran the numbers in a small spreadsheet. At twelve percent hint usage — roughly the observed rate — a free completionist earned three thousand cogitats across the free nations and spent seven hundred. Surplus: about two thousand three hundred. At thirty percent usage, the surplus still sat above a thousand. A working economy ends each nation near zero, not in profit.

The diagnosis was straightforward. Earn rate was correct: thirty-nine cogitats per solve is the celebration moment, the small "+39c" flash that tells the player the puzzle was worth their time, and nothing in our playtests asked us to weaken that. The problem was sink. There was only one — hints — and it was priced like a free sample.

We considered three options. Decay the earn rate (worse celebration moment). Add more sinks (good, but slow). Raise the price of the existing sink (cheap, fast, surgical). We picked the third.

4. The new prices: forty, eighty, one hundred and twenty

Four times the cost. Same earn rate. We shipped it as part of v4.3 and watched.

The first thing that changed was the time between hints. Median delay from puzzle-start to first-hint-tap moved from about three minutes to about seven. The second was the re-read rate: players were noticeably more likely to re-open the puzzle prose, or the Carnet entry that gestured at the constraint, before reaching for the hint. The hint had stopped being the path of least resistance.

The arithmetic is now this. A free completionist at twelve percent hint usage ends the free regions with about two hundred cogitats in surplus — close enough to zero that the next region's first puzzle clears it. A player at thirty percent ends in the negative, which is exactly the moment the in-game low-balance promo fires and offers a discounted pack. We used to fire that promo at region completion, which was the worst possible moment because the player had just earned a thousand cogitats and felt rich.

The shift in player feel is harder to measure but easier to describe. A forty-cogitat hint is a choice. A ten-cogitat hint was a reflex.

5. The Pro discount: thirty percent

Pro subscribers pay seventy percent of the published price. That is the thirty percent hint discount on the pricing page. We chose thirty for a specific reason: the perk has to feel real, and the friction has to stay intact. At fifty percent, hints stop being a deliberate trade. At twenty, the perk reads as a rounding error. Thirty is the number where a Pro player still pauses before the tap, but pauses for slightly less time.

Pro also includes a monthly stipend of one hundred cogitats — enough to cover two or three first-tier hints, not enough to coast through a chapter on. The stipend exists to make the subscription feel ongoing, the way a magazine subscription drops a new issue through the door each month. It is the small recurring gesture; the discount is the daily one.

6. Lifetime, sixty euros

Atlas Pro is six euros a month. Lifetime Atlas Pro is sixty euros, one time. Ten months of monthly is the break-even; after that, the lifetime tier is the cheaper option in perpetuity. We picked sixty, not eighty and not forty, by asking the question we ask about every price on the site — who is this for?

The lifetime tier is for the player who has decided they care about the universe: who has read the Carnet entries that explain why Theodor's marginalia matter, and solved enough of Vellestria to know that there are four nations still to come, sketched on the wall behind the desk in Paris, waiting their turn. They are not buying a month of access. They are buying the whole atlas, in advance.

Sixty euros is a small bet. Ten months of monthly is the floor; the four future nations are the ceiling. The hint prices are forty, eighty, and one hundred and twenty because the puzzles deserved a real currency, and that currency had to mean something on the way in and on the way out.

Cora's note: the balance flickers up by thirty-nine each time you solve. I have watched players pause on that flicker. The number is a small celebration, but the pause is the second when the solver decides whether to bank the cogitats or spend them. That second is the one we are designing for.