Just One More Day

I started looking into this because I couldn't quite explain what I liked about Pokémon and Stardew Valley. They get grouped into different genres, but the appeal feels pretty similar: there is always something else worth finishing.

In Pokémon, that might be getting a team member to the next level. In Stardew, it's getting through the day. Except getting through the day gives you ideas for what to do tomorrow, which makes tomorrow a bad place to stop too.

Calling that “addictive” doesn't explain much. I wanted to know what makes the next little task feel worth doing, and why the same feeling shows up in apps that have nothing to do with games.

You finish one thing and inherit three more

Take a sprinkler in Stardew. Watering crops takes time, so building something that does it for you is an obvious improvement. You gather materials, make it, and put it down. Now you can plant more crops, rearrange the field, or spend the morning doing something else.

The upgrade changes your plans. You want to see it work.

That's what I find interesting about Stardew's connected systems. Farming, crafting, and exploration give each other a purpose. Materials are more interesting when you already know what you want to build. An upgrade matters because you remember the inconvenience it removes.

Pokémon does this with your team. A new move gives you something to try in the next battle. A new catch might change which creatures you bring along. Progress keeps introducing choices you didn't have before.

That is my read of the design, rather than an experimentally proven explanation for why people stay up playing. But it gets closer to the appeal than describing every reward as another little hit of dopamine.

Being nearly done changes things

One of the better studies I found involved coffee loyalty cards. Researchers studying the goal-gradient effect found that customers bought coffee more rapidly as they approached a reward.

They also compared an empty ten-stamp card with a twelve-stamp card that already had two stamps. Both required ten purchases, but people with the apparent head start completed the card faster. Showing people that they had already made progress changed what they did next.

It's easy to see how that could apply to games. An almost-full experience bar is an invitation to finish. And when a game has several things progressing at once, something is usually close. The next Gym might take a while, but another level is only a couple of battles away.

That doesn't mean a progress bar automatically makes something compelling. A meta-analysis of 32 survey experiments found no significant overall reduction in dropout from constant progress indicators. You still need a reason to care about reaching the end.

For games, some of that reason is simply enjoying what you're doing. Research on game motivation links enjoyment with feeling capable, having meaningful choices, and feeling connected. Getting better at something is satisfying. So is building a farm that works the way you want it to.

The streak can become the goal

The business examples get interesting when the thing being measured starts separating from the thing the user came to do.

Duolingo described an experiment where completing one lesson became enough to preserve a streak, instead of requiring the full daily goal. The company reported a 3.3% relative increase in day-14 retention. Fewer users completed their daily goals, although total learning time increased.

I wouldn't read that as a failed experiment. Making it easier to show up might be useful. But those results leave a question unanswered: did people learn more? Returning to an app, spending time in it, and learning a language are related, but you can't substitute one measurement for all three.

Autoplay works on a different decision. You have finished an episode and could leave, but the next one starts for you. In a controlled study of 76 Netflix users, disabling autoplay shortened playback sessions by an estimated 17.6 minutes relative to the comparison, accounting for baseline changes. It didn't significantly change how soon people returned. The study was short and measured playback rather than attention, but that distinction is useful. Keeping someone here and bringing them back are separate problems.

Stardew actually has a clear stopping point. You go to bed, see the day's results, and save. It just happens to leave you with things you're looking forward to doing tomorrow. I think that explains more of its appeal than treating the whole game as a trap.

What are we trying to get people to do?

If someone spends an enjoyable evening playing Stardew, the time can be the point. That doesn't transfer neatly to every product. I want my accounting software to help me finish and leave. A longer session there might mean something has gone wrong.

That's the part I'd borrow carefully from games. Give people a reason to care about their progress, make the next step understandable, and let them use what they've earned. Then check whether they're actually getting what they wanted.

The thing I like about Pokémon and Stardew seems to be that progress stays useful. Finishing something makes the world a little different, and I want to see what I can do with it.

Unfortunately, seeing what I can do with it usually takes another day.