The first five minutes of my game, rebuilt five times

TL;DR: Inselnova is a browser strategy game you can play with no download. The hardest part to get right has been the first five minutes. I’ve rebuilt the first session five times. Day-one retention sat around 25%, which means three out of four people who started never came back. Here’s what each rebuild was chasing, the levers that moved the number, and the one thing I got wrong about what onboarding even is.

The first session is the whole game

The first few minutes aren’t where the big stuff happens. You’re not running a war with your neighbours in your first five minutes, and you shouldn’t be. That comes later, once you’ve built up.

What the first session has to do is teach the early loop. Gather, build, wait, come back, do the next thing.

It also has to show you the shape of the rest of the game without handing it over. You can see there’s more. Raids, alliances, spying, a market. The hints are everywhere, you just know you can’t do those things yet, not until you’ve grown into them. That discovery, the game opening up the more you play, is the actual pull.

The first session’s job is to get people far enough to feel that. And too many weren’t getting there. Day-one retention was about 25%. People would land on their island, look around, and stall. Not rage-quit. Just close the tab and not come back. That’s the worst kind of churn because it’s silent. Nobody tells you why.

It’s a shop nobody walks into

Think of it like a shop. You can have the best thing in the world on the shelves, but if nobody walks through the door, none of it counts. The front has to pull people in. The first session is the shop front, and for a game that depends on strangers choosing to stay, it’s the most important thing you’ll build.

So I track every step of it. Not full session replay or mouse-movement logging like PostHog, I don’t have the time to sit and watch recordings right now. But every step a player takes through onboarding and the tutorial is logged like a sales funnel. Landed, first build, first action, finished the first set, came back the next day. I can see exactly where the drop-off is.

When a step’s numbers look bad, that’s where I spend time, and I mean serious time. A leak on one early step outweighs any feature I had planned, because that step is bleeding the players everything else depends on.

And I run the whole thing myself, over and over, on a local server. Cold, like I’ve never seen it. It doesn’t replace real players, but it catches the obvious problems before they ever reach one.

Where new players actually stall

The early game wasn’t too hard. It was too quiet, and it punished people for not knowing things yet.

• You’d land and not know what to do first. An island, some buttons, no obvious next move.
• A few early decisions quietly hurt you if you got them wrong. Setting a tax you didn’t understand could tank your own island before you knew what tax did.
• Nothing pulled you toward other people, and the other people are the actual game.
• The world looked empty. There’s a world chat, but it goes quiet for stretches, and a silent chat makes a live game look abandoned.

So the rebuilds weren’t really about a tutorial. They were about removing the ways the first session made people feel stupid, and adding a reason to still be there at the end of it.

The world felt dead, so I added a ticker

A live game that looks empty is worse than a small game that looks busy. New players were landing, seeing a quiet chat and a still map, and deciding nobody was here.

So I added a ticker. It collects the interesting things happening in the world right now, raids, colonisations, players making moves, ranks them by what’s worth seeing, and scrolls them across the screen.

When I first put it in I thought it was too much. Too busy, too noisy. Then I got used to it, and it changed how the whole thing felt. It looked like things were happening, because they were happening. Real raids, real players, real moves. Without the ticker none of that was visible, and a new player had no way to know the world was alive right now.

What each rebuild was chasing

I won’t pretend the five versions were a clean plan. They weren’t. Each one fixed the thing that was most obviously broken at the time. The themes that came out of it:

• Give them a first goal. A first set of achievements, the Settler’s Mark, so there’s an explicit “do this next” instead of a blank island. Finishing that first set turned out to be a strong signal that someone would stick around.
• Stop punishing ignorance. The early decisions that could quietly wreck a new player got locked or simplified until they’d learned enough to make them. Fewer ways to lose before you understand the game.
• Make the start feel competent, not crushed. New islands get an early advantage so the first session feels like you’re getting somewhere, not getting stamped on by someone who started a week ago.
• Get them to other people. The biggest one. The real anchor isn’t the tutorial at all. It’s joining an alliance. Players who get into an alliance early stick around at a completely different rate to players who stay solo.

That last point reframed the whole thing for me. I’d been treating onboarding as “teach the player the controls.” It’s actually “get the player to the other players before they leave.”

The thing I couldn’t see

Here’s the honest part. After five versions, I genuinely can’t tell anymore whether the first session is clear or whether I’ve just stared at it for so long that I’ve gone blind to it.

That’s why I started asking strangers to play it cold and tell me where they stalled. Not the kind feedback. The brutal kind. You cannot judge your own onboarding, because you already know where everything is. The person who’s never seen it is the only one who can tell you the truth.

If you’re building something similar

• The first session is your highest-leverage work. Not the mid-game, not the feature you’re excited about. The door.
• Watch for silent churn. People who leave in the first few minutes don’t complain. You have to go looking for them.
• Find the anchor. For me it was alliances. Whatever yours is, getting a new player to it fast beats any amount of tutorial polish.
• Get fresh eyes. You’ve gone blind to your own start. Someone else hasn’t.

Day one is still where I spend most of my time, and it’s still the number I watch first. If you want to see what the first five minutes feels like now, it’s free and there’s nothing to install: inselnova.com.