Fixing Retention - Battlescape's last chance

Retention is the percentage of your players who play your game again after they have installed it. It’s a metric that is essential for measuring the performance and monetization potential of freemium software in the mobile world. Most publishers will carefully monitor retention figures for day 1, day 7 and day 30. That is how many players are still playing after a day, a week or a month.

It is not usually as useful for traditional (premium) games where the practice is that the player pays for the game upfront. There is an important exception in the premium games world where the figure for day 1 retention becomes very important again. That is online multiplayer games. When you’re making an online multiplayer game you are essentially providing a software as a service (SAAS) product.

Having poor retention in an online multiplayer game results in empty servers for large periods of time. The more your game depends upon player interaction for gameplay, the greater your retention figures need to be. If the unique selling point of your game is “massive space battles with hundreds of players” then you’re going to need really good retention.

A healthy figure for day 1 retention is about 30-40%, day 7: 15-20% and day 30: 10-15%. With those figures we can run some numbers and work out how many new players we need to be acquiring every day to keep the concurrent player count at an average of 100 players. To make it easier we’ll assume the game has settled down after the initial launch spike. We’ll also assume a single server, so multiply any final results by the number of servers you are running.

First lets assume the average session length is two hours per day, that may be a bit generous, so bear that in mind and adjust the final figure based on your own estimation. That means that to get 100 concurrent users online at any particular time we will need at least 1200 (24 hours / 2 hour avg session * 100 users) daily active users (DAU). That’s very achievable for a game with decent retention. I’m skipping over the fact that there will be quieter periods on the server depending on geographic location. 1200 DAU is a good enough ball park figure.

A quick calculation for how many players we fail to retain on any given day is our day 30 attrition rate divided by 30. Attrition is simply the inverse of retention figures. So if day 30 retention is 10%, day 30 attrition is 90%. Divide that by thirty days and you get a daily attrition rate of 3%.

So we’ll be losing about 3% of our 1200 DAU each day which works out at 36 users leaving. However our daily active users are only a small portion of the total number of active users of our game. Depending on the game, the DAU figure is between 10-20% of the total active user base. Lets be generous again and say its 20%, which means we actually will be losing five times our DAU count (3% of 6000) or 180 users being lost every day.

That means to keep the game populated in the long term, using those fairly generous figures we need to be making 180 new sales of the game every day.

Now that’s probably possible if the retention figures are good enough. Even if the game’s price were reduced to $10 that would be $1800 of gross revenue per day (minus store cuts). That’s plenty of income to fund a paid user acquisition (PUA) campaign to keep the new players coming in. Somebody at INovae will need to become an expert in PUA because it’s unlikely that organic discovery will keep the game alive for very long.

But all that is dependent upon fixing retention, which is what my next post will be about.


So, how do we keep our attrition rate below 3% per day?

The most effective way is to look around at games that are doing it successfully and copy them. If you’re developing a combat orientated space sim, you should be playing all the other popular combat orientated sims and coldly assessing what makes them “sticky”.

This may be heretical but I’ve never played War Thunder or Rogue Squadron, but if I was creating a competitor I would be playing them, taking careful notes and discussing with my team members (and perhaps my developer level supporters) what makes those games so replayable. There is no need to reinvent the wheel, the I-Novae engine has plenty enough originality already.

Some things that are obvious:

An in-game tutorial. It’s 2020, this is expected now. Just a few waypoints, a bit of guidance and some targets to shoot at. There should be somewhere safe for this to happen, perhaps a local instance of the server?

A clear objective. It’s cool that there are all these battles going on all around Rethe Prime, but it just makes everything too chaotic. Copy the winning game modes from other multiplayer fps games. The simplest would be to make each Battlescape a control point, which have to be won in order to move to the next. As a secondary result that also fixes the issue of teams having nowhere to spawn a capital ship, as the last control point can always be able to spawn them.

Random interesting stuff. At the moment exploration is only needed to find enemy bases, but there are other simple ways that exploring gameplay could be added. I think there was supposed to be crates to be salvaged that were dropped by ships. How about crates just laying around the planet’s surface at the start of the game? Copy ED’s “unidentified signal source” mechanic so the player doesn’t know what they will find. Could be a crate, could be an enemy ship or base, or could even be a statue placed there by a high level kickstarter backer. It could also be combinations of the above.

Next, measure everything. Use whatever analytics you can get hold of to further understand what your players are doing. Sit in Spectator mode and just watch people play. If you can distribute slightly different clients to different segments of players do that and A/B test the heck out of any changes you make. You have multiple servers which is fantastic for A/B testing changes to game modes for example. Test different battle sizes on different servers, do players feel more engaged if there are less NPC ships in a battle and they feel they have more agency?

Finally, don’t be too afraid to pivot. If you are watching your players playing the game in a certain way, and it’s not what you intended, allow them to lead your decisions. They are the ones that will ultimately decide what will be successful for you.

Edited to add: Oh yes, add in some achievements as well for heaven’s sake. Yes, they are cheesy, but there is a substantial subset of your players that will keep on coming back just to get that next achievement level. There are hundreds of things you can add as achievements and they take very little time to code. If you need any ideas just ask in Discord.


while i dont see any errors in your post, player retention mechanics have been postponed to end of year.(daily missions, progression, customization, etc.)
so you are not wrong, just about a year early. :wink:

as far as i can tell, I:B will simmer in early access for a year or two and then make a second splash.

1 Like

You counted well. And they came to the conclusion that it is necessary to reduce the Price of the Game, but this will not save, and even if the Game is distributed by Shares or Free, people will come in and go. There is no game itself, no content. Many are not improvements. No one will play the game without content! This game is for 2-3 hours, then annoyed.


Somebody should really tell the devs to continue to work on the game then.


totaly! lol! :wink:

sometimes i think people have never seen an Early Access title before. if you cant handle unfinished and buggy games, you realy shouldnt play games 1-2 years before their release.

A year is way too long to wait.

1 Like

well, then dont wait, im not waiting either. im playing finished games and occationaly checking back on my Early Access games.(between the stars, factorio, below zero, wayward, starsector and probably a few others i cant think of at the moment ) some of them i bought 5 years ago, still not finished. there is realy no point in waiting around. lol

I’m not talking about me, I don’t have any trouble with waiting:

I’m talking about the dev team at INovae.


as I understand it, the developers swung at the global project, but did not calculate the forces. What they wanted, in principle, they did, gave us a seamless World, and a battle, large-scale battles. But the bottom line is that no one needs it in this form. And they sat down. Although it was possible to go by the sweat of least resistance. Make a game Free to Play, the same tanks are only more massive, what is missing? Content and development. And it was necessary to spend time earlier, now it will drag on for several years. Since there are no resources and labor. It is impossible to advertise the game in this form! The developers only pushed the players away, collected the negative.

1 Like

Call me crazy, but I think the daily attrition rate of 3% is … maybe a bit too simplified. If I am not mistaken, most people leave until after day 1, whereas less people leave until after day 7 / 30. In your example if a game lost 90% of its userbase after day 30, It is safe to assume that about 50% of the starting player base left day 1. Your daily rate of 3% (of the starting player base) leaving implies the exact opposite, that a higher and higher percentage of the remaining player base is leaving each day. Both attrition curves, although they both end at 10% of the starting userbase, call for different approaches to rectify the situation. One calls for improvements in the day 1-day7 experience, whereas the other calls for improvements of the late game experience.

1 Like

There is a way out, to sell the project, to recoup all the costs and earn money, Wargaming and Gaijin Entertainment, have vast experience in such projects, their Games are popular and cost-effective. Now it’s not difficult to create a Seamless World, all modern engines support it, ship models are created very quickly, so I don’t see any problems. But the developers will not go for it, they will also say, treat with understanding. We all understand that once in 10 years, maybe the game will come out of Beta, but it will be needed then when the Game Industry moves forward at a frantic pace.

1 Like

Modern engines support a coordinate system that enables solar system scale precision and have seamless planetary generation and travel without loading screens while also supporting multiplayer with 100s of players? As far as I know only the custom engine from star citizen comes close to these specific features, after they invested millions of dollars.

Why would they even need to sell anything? They made enough money already to continue the development until full release.

Your points have in fact been raised hundreds of times, they are already hard at work on the game to deliver more content, better retention mechanics and so on.
If you want to complain about something you should complain about their perfectionism that has them reworking stuff that isnt really broken lol.

You need to become more familiar with the capabilities of Game Engines. Do you think that there is no preload in Infinity? She is, she is simply not visible to the Player. What is it worth creating a large trainer and dividing it into sectors? With proper code, sectors will be generated.

1 Like

Let’s hope that the development will not be delayed.

1 Like

Please educate me what are these magical engines that have double precision floats except infinity and the star citizen engine while also featuring procedural planets and 100s of players?

Exactly correct, it is over-simplified and is a very rough estimate. I appreciate you noticing. As you pointed out, the attrition is heavily loaded towards day 1 with an exponential drop off curve as the days go on.

The reason I’m averaging it out over 30 days is because it assumes there will be players at all stages of the retention curve, people that have played for months and also people that have just downloaded the game. An attrition rate of 3% per day also looks about right given my own experience, although it may be as low as 1% if there are a lot of long-tail players at day 90 or more*, this is quite possible.

Generally speaking, if day 1 retention is below 30% you need to be focusing on the early game experience. If your day 1 figures are good then you can look at later game retention, based upon day 7, 30, 90 and beyond.

note* If day 90 retention is also 10% then your daily attrition by my rough measure is 90% / 90 days = 1%. To take that further if somehow your day 180 retention was also 10% (Your end game content must be really good!) then you can estimate your daily attrition as 0.5%.

The star system in Infinity is not as big as we see it. She flies in 10 minutes. So the card itself is small. It just consists of many sectors that are loaded. This elementary technology is in many Game engines. UE4, CU. In fact, this is an optical illusion. Fixed Planets? This is a model with a constant Texture, which is tied to the sector, attraction to planets and satellites, everything is written in the code. There is a random generation of planets, terrain. In today, Seamless World is not such a problem. About 15 years ago, but it was new, but not now.

1 Like

Sorry but you don’t know what you are talking about. The scale in Infinty or Star Citizen isn’t faked anywhere, things are really as big as they are presented. To have that possible the engine must support double precision floats which the star citizen devs have spend years and millions of dollar to implement into the former cry engine.
No other game has that and you can even see that in this trailer where the planet starts to jitter because to observer is too far away and the precision breaks: