Don't Give Up the Day Job: An indie Windows Phone 7 post mortem

It's now more than 3 months since Brickbat, the first game from my one-man indie studio Jeffalisk, debuted on the Windows Phone Marketplace. Here I take an unflinching look at the few things that went right, the many things that went wrong and share my download figures for fellow devs to mull over.

Brickbat screenshot

What went right

Learned a lot

At the end of the day, this is all you're really looking for from a first project, especially when you're coming to game development from scratch as I was. Nine months in development might seem a bit over the top for a mobile phone Breakout clone, but included in that time was learning C# and XNA from scratch and a crash course in the fundamentals of game design and programming. The end result was never likely to be anything to set the world alight - my main aim all along was just to see a project through to release, and anything else would be a bonus.

Hey, you've got to pay your dues before you pay the rent.

Choice of platform

Now, there will be people - many people - who think this needs to go under "What went wrong" along with, well, virtually everything else. But I still think, despite Brickbat's poor download figures which we'll get to later, that Windows Phone was the best choice for someone starting out in mobile games at the time I did.

The App Store is saturated with so many polished titles already that trying to get a new game noticed is real needle-in-a-haystack stuff, plus I've never heard a good word said about Objective-C. Android is everywhere but the writing a game to perform well on all the various software and hardware combinations out there was a headache I really didn't need.

With Windows Phone, I took a gamble to get in, at an early stage, on a new platform that's just getting started. Many of the disadvantages for established devs didn't apply to me because I was just starting out. Microsoft's free dev tools are great to work with, and with tightly controlled hardware standards compatibility isn't really an issue.

It is a gamble, granted, but for now I still feel this was the right choice and I can always switch platforms later on if needed.

What went wrong

Attack of the Clones

When Brickbat started development, there were perhaps five Breakout games on the WP7 marketplace, and most of them were pretty rudimentary. By the time Brickbat was released, that number was well into the dozens - notably including a high-profile Xbox Live title (IonBall EX). It seems every dev and his dog chose a Breakout clone to test the water with.

Worse, I'd been coy about the game's influences in the original marketplace submission - neither "Breakout" nor "Arkanoid" were included as search keywords, and it was submitted in the hugely overcrowded catch-all "action + adventure" category, with the result that it was quickly buried off the front page and couldn't even be found by people specifically looking for that kind of game. Both these points were rectified for version 1.1, and had an immediate impact on downloads.

Being crowded out of the market was partly down to a long development time, but mostly down to a poor, uninspired choice of project to start with. Even the most ardent WP7-owning Breakout fan would have been getting sick of them by the time Brickbat came out, and this is something that should have been foreseen before the project ever got off the ground.

Lesson learned, though: future Jeffalisk games will bring something original to the table, or they won't get made.

Shifting scope

Perhaps inevitably for a first game design project, the scope of Brickbat shifted dramatically as development went on. At the height of my design mania, it was going to be a story-driven game using Breakout as a metaphor for psychotherapy, with the player breaking down "mental blocks" to help various patients. Eventually it was scoped back to a reasonably bare bones arcade game. But the amount of wasted effort that went into the wilder flights of fancy was ridiculous and could have been put to much better use adding polish to the more manageable game that we ended up with.

Underestimated development time

Another classic beginner's mistake, I started out with an unrealistic aim of completing development in 3 months. Looking back, this was never going to happen. The problem was compounded by announcing tentative release dates, twice. This is something that I can't stress enough for anyone starting out in indie dev, especially if you've also got a day job: everything is going to take much longer than you expect it to! And just when we thought we were done, we hit more delays – Brickbat failed certification twice because of bugs, and each time through the MS cert process took roughly 1 working week to complete. I'm not saying every indie project's release date should necessarily be “when it's finished”, but setting targets you can't keep kills motivation. I won't be doing it again.

Holiday launch

Through a combination of the factors I've already gone into, Brickbat ended up launching at the worst possible time imaginable – between Christmas and New Year. Take it from me, fellow indies, this is a terrible time to release an app. Nobody wants to buy a mobile app at that time of year, because they're all off work, off school and have access to “proper” games. Even those who do want to buy an app are much less likely to be spending time online or on social networks so they aren't going to see your marketing. And even if they somehow do, and end up buying your app, they're not going to be seeing anyone to spread the word. Worse, any press release you send out will be roundly ignored because the games sites aren't working either. But I was impatient. I'd just spent 9 months working on this thing, so it was going to be released as soon as humanly possible. The temptation was overwhelming. But in hindsight, this was a massive mistake and I should have waited until January for a more coherent launch.

Launched without a trial mode

Another huge mistake, and another one that first-time devs fall into all the time. Other WP7 devs warned me on Twitter that lots of people won't download without a trial, but I still went ahead and launched without one. The results were underwhelming – first week sales totalled a less-than-inspiring 14 copies worldwide, at the Marketplace's lowest available price point of 79p (99¢).

Realising my mistake, a version 1.1 update - which included a trial mode - was rushed out , and the impact on downloads was dramatic – I saw a big spike when the trial went live, and paid downloads started ticking upward again after having flatlined. Next time I launch a paid-for game on WP7, the trial will be in there from day zero. For the love of god, WP7 indies, learn from my mistake here. (There is, of course, the wider question of whether paid-for games are viable on WP7 at all, which I'll get to later.)

Low-key marketing effort

I'll be honest here, I had lots of good intentions about getting word out there but – mainly because of the timing issues mentioned above – they all fell by the wayside. The launch marketing consisted of a launch trailer on YouTube, a press release that got featured on one site, and continual plugs through our Twitter account. But there wasn't much real strategy and this will need looking at for future projects. Games, particularly generic ones, won't just sell themselves.

Don't get too excited: Sales figures

Fortunately for me, I don't need to earn a living through selling my games. I say fortunately because it honestly doesn't look like I'll see a cent from this game on its own. Let's take a look at my figures to date:

Here's the graph:

Brickbat post mortem downloads graph

The first huge spike is the 1.1 patch – which introduced trial versions – going live. The second, slightly higher spike? I'll be honest, I have no idea why that happened. (If any fellow devs have experienced similar, I'd love to hear about it in the comments). As far as I can tell the app's never been featured on the Marketplace, or reviewed anywhere, so I am genuinely at a loss as to where all those hits came from.

With the exception of those spikes, sales – and downloads – have been very flat. Taking into account Microsoft's cut of the sale price, those 50 sales have earned me about $35. There's a $200 minimum threshold to hit before I see a payout, so I'd need to sell about 5 times as many copies again before getting paid, and downloads have pretty much dried up. I'm not ordering the swimming pool just yet, put it that way.

Now, a lot of this is down to my own ineptitude. There were many mistakes made with this release. It was the wrong game, at the wrong time, launched badly and not promoted properly. But those are still some pretty wretched figures, and I wonder if they indicate a wider problem with paid apps on the WP7 market.

It seems to me that the WP7 marketplace, when it comes to games, is becoming a two-tier market that shares much in common with that for downloadable titles on Xbox 360. Xbox Live branded titles enjoy prominence in the storefront and anecdotal evidence suggests many gamers never look beyond the Xbox Live section of the Marketplace, leaving self-published games to fight it out for a smaller slice of what's already a very small pie when compared to iOS or Android.

Elbert Perez, to my knowledge more or less the only one-man-band dev to make a living out of developing for Windows Phone, gave up on paid-for apps early on in his career, and has provided plenty of figures over on his site occasionalgamer.com to back his theory that releasing games for free and supporting them with the revenue from in-app ads is the only way to earn reasonable money on WP7. The hit word game Wordament also follows the ad-supported model.

I'm reluctant to go down that road, and probably won't give up on paid-for just yet, but hopefully by releasing my figures I can at least provide some food for thought for anyone thinking of entering what's still an unproven marketplace.