Weather Factory

Welcome to this year’s Giant Summary Post! This is almost twice the length of the post-2018 State of the Factory, because a lot happened. Come for the kitten pictures, stay for the data.

As I said last time, these posts serves two purposes. One, it’s likely to be interesting both to other indie devs and to our community, but two, it’s useful to us. It’s a useful exercise to do a transparent retrospective on our year, and also I end up  summarising information here that I never actually write down otherwise (I’ve gone back and looked at our Year 1 post for reference a dozen times).

The story so far: I’m Alexis Kennedy, and I’m one half of a microstudio called Weather Factory. The other half is Lottie Bevan. In our first year of operation, we made a game, Cultist Simulator, that sold well and won some awards. By the standards of a small creative start-up, Year 1 was very good indeed.

Here’s what I said last time, at the end of 2018, about our plans going into Year 2:

“What are our priorities now?

In the short term, there are three projects fighting for our attention: the one we’ve codenamed Ophir, the one we’ve codenamed Procopius, and the one we’ve codenamed That Damn Library Game. You can probably expect to see us announce, and likely run a Kickstarter for, one of those this year.

Our next project will probably be a notch higher-budget than Cultist – same sort of scope, but slightly more adventurous UI and a few months of polish. I’ve never really made a game with polish. So we’ll be growing the team. I expect we’ll be four full-timers at the end of 2019, which might mean we’re five, because once you open the gates, head-count tends to tick inexorably upwards.

In the longer term…

In the longer term, I want to be a two-project studio sooner rather than later. I don’t like us putting all our eggs in one basket, and eventually a project will fail. In particular, as I said above, I don’t like being dependent on one storefront (sorry, Valve! but you know how it is). So I think, in 2020, I’d like us to be doing something ambitious and unusual on a larger scale.”

I’ll talk more below about how all that went. Let’s take it as it comes.

Cultist Simulator: the Story Continues

We started the year with no deadlines except those we imposed on ourselves, which was a nice relief after last year’s ferociously disciplined march to an aggressive release date. The one commitment we had was this: we wanted and also needed to release at least two pieces of DLC. ‘Wanted’, because there was a fervent community demand for it and we thought it was likely to make us some money, but ‘needed’ because of Perpetual Edition.

Anyone who backed the Cultist Simulator Kickstarter – or bought the beta when we were selling it before launch on itch.io – or bought in launch week – got the Perpetual Edition, which meant we guaranteed all DLC free forever to those purchasers. In order to make that promise mean something, we needed actually to release DLC. Lottie and I reckoned that three pieces of DLC was a decent number for people to feel like they’d got something worth talking about. We’d already released one piece, the Dancer, in October 2018. We planned two more pieces of DLC to release in May 2019.

Why May? Because that was the anniversary of the release of Cultist the previous year, and that was when we were going to make the ‘Anthology Edition’ available. The ‘Anthology Edition’ was a bundle of the game, with all the DLC, with the soundtrack, with a discount on each. Here’s why, and here’s where, in terms of community relations and audience expectation, things get fiddly and interesting.

The Perpetual Edition seems to have been a success. ‘Free lifetime DLC’ felt like an honourable thing to offer early backers; we think it pumped our sales in launch week; and, critically, it didn’t cost us anything extra before launch. (Well, I had to fiddle about for a couple of days with the Steam and GOG DLC ecosystems so I could make the Perpetual Edition banner show up right for people who’d bought it on those storefronts. Nothing’s ever really zero-cost.)

We’d probably do it again.

But you can’t please all of the people all of the time, and we had some grumbles from people who liked the lifetime DLC deal and regretted only hearing about the game on day 8, or month 8, when Perpetual Edition was gone for good. If we ever made Perpetual Edition available again, we might upset our early backers,  because we’d assured them it was a one-time deal. If we didn’t, we’d still get those grumbles, and miss a sales opportunity.

So we found a middle way. We announced that when we’d released all the DLC, we’d bundle it as an ‘Anthology Edition’. It was effectively the same as Perpetual Edition at a higher price and a year later. Everyone seemed happy with that. The early backers still felt they’d got a good deal, and the latecomers no longer felt they’d missed out entirely.

When we actually released the Anthology Edition, though, we ran into some other grumbles. None of these were serious or widespread, but I reckon they’re interesting because they shed light on how easy it is to make well-intended commitments that limit your options later.

For example, we released the first DLC, Dancer, at $2.99 / £2.50. We’re pretty certain now that this was too cheap for the amount of content. Even at the time we thought it was probably too cheap. Our players have certainly told us they think it’s super cheap. But we really didn’t know how to price a smallish piece of DLC. We havered over worrying about going too low and too high, and what decided us in the end wasn’t, honestly, generosity: it was that I’d got carried away and spent way too long working on the Dancer. We reasoned that if we set a low price, people wouldn’t expect too much content in the future, and that would allow us to do small pieces of DLC quickly rather than feeling we had to really push the boat out.

That worked out fine. The Ghoul and the Priest are noticeably smaller than the Dancer, but they’ve sold just as well and no-one’s complained they were light on content. And I wrote them both in about the same amount of time as I wrote the Dancer. This is a big deal when you’re a small team.

But… we had a couple of folk complaining that the Perpetual Edition was a bad deal, because you only got three pieces of $2.99 DLC. They were a small minority, and the community response to the grumblers was still, eh, it’s still free content. But I don’t think it was a wildly insane complaint.

Similarly, we got this:

 

I mean, it ain’t so. There’s huuuge quantities of content and story in the base game, and the DLC is bolt-ons. But some customers see three pieces of DLC at the bottom of the store page and decide that we’re gouging our players. I could go in and add a developer reply to this review and explain the whole Perpetual Edition context and talk about all the free updates we did…but when you need to explain your reasoning like that, then you’ve probably already lost.

Or take the soundtrack. We agreed early on to split revenues 50/50 with our usual collaborator, Mickymar. We released the soundtrack on Steam, GOG and Humble – and also on Bandcamp, Spotify, the other usual places. It wasn’t included in the Perpetual Edition. If you ask the next habitual gamer you meet if a soundtrack is ‘DLC’, they’ll probably say ‘no!’ or at least ‘no?’ DLC is new story content, expansions, horse armour. And anyway we couldn’t have given the soundtrack away for free without stiffing Maribeth and her compatriots.

But on Steam, in 2019, soundtracks were sold in the ‘DLC’ section of the game store page. This is just the way their distribution system was organised, and on GOG and on Humble and obviously on Bandcamp and what-all, soundtracks weren’t identified as DLC. In fact, even Steam have just changed it this year and soundtracks are no longer labelled DLC.

But what this meant was that a number of players, some of whom had supported us since the beginning, logged into Steam to get their free Ghoul and Priest DLC and noticed that there was something else labelled ‘DLC’ that they weren’t being offered free. Again, this was a minority, again, people basically understood the issue, and again, it wasn’t a big deal, but again, it wasn’t an insane complaint and we had to spend some time dealing with it.

The thing is this. When you make a commitment – whether it’s a Kickstarter stretch goal, a release date, or a promise of all DLC free forever to early purchasers – you limit your future options. If you’re a big company, your future options are surprisingly limited anyway, because there are bills you have to pay and things you have to do. If you’re a small company, your agility is your advantage. Anything that locks you in is a risk. And sometimes, with Perpetual Edition and DLC and our previous commitments, we felt like we were most of the way through a game of Twister.

It was fine! But it was fiddly. Next time, we’ll know about this stuff in advance. (And we’ll probably get caught out by something completely different.)

Here are the unit sales for Cultist, Cultist Perpetual Edition, DLC and the soundtrack to date (as of Jan 31^st 2020, twenty months after release). Don’t worry, I’ll share the revenue numbers further down, too, but I want to focus on the proportions for now. These are sales on Steam: they don’t include keys redeemed from Kickstarter, or from other storefronts. This also means that the DLC numbers you see there are people who’ve bought the DLC, not people who got it free from Perpetual Edition.

 

So ONE, yup, 22% of all our Cultist Simulator sales to date there are Perpetual Edition.

One way of looking at that is: there are 34K players out there who we can’t generate revenue from by selling them DLC. That sounds like bad news.

Another way of looking at it is: we sold almost 34K copies in launch week. Doing well in launch week – charting at #1, as we did – is really important for the long tail sales of your game. You can recover from a bad launch. People do. But you don’t want to have to. Perpetual Edition helped us do well in launch week. But we’ll never know how much difference it made.

And a third way of looking at it is this: we can still make money by selling any future games to those 34Kish early adopters. It helped us build a core of people who know they can trust us.

But here’s one final way of looking at it. there are at least 90 million active Steam users, and something over a billion PC gamers, in the world. We’ve sold Cultist to about a tenth of one percent of the Steam userbase, and a hundredth of one percent of the PC gamers in the world. Of course the vast majority of those people will never even hear the name ‘Cultist Simulator’, and most of the rest are probably busy playing CS:GO or Fortnite or something. But to quote Lottie: “The next time you hear anyone say they’ve tapped out their audience, kick them in the shins. No indie has ever tapped their entire potential audience. We simply do not have the money or resources.”

If you’re a small lifestyle business, it pays to keep an eye on the long tail and the long term.

TWO. Dancer has been on sale since Oct 2018, i.e. 15 months. Ghoul and Priest have both been on sale since May 2019, i.e. eight months, i.e. 53% as much. But they’ve both sold about 70% as much. I think this is because we made as big a splash on the anniversary of Cultist Simulator as we could.

THREE. Ghoul has sold very, very slightly better than Priest. Ghoul has 12 user reviews, 100% positive. Priest has 11 user reviews, 41% positive. I think the writing in Ghoul is a little better than the writing in Priest, and the design in Priest drew some criticism for being grindy (an Alexis Kennedy game? Grindy? Say it ain’t so). But it doesn’t look like players are very strongly influenced by the user reviews for a piece of DLC… at least one priced at £2.50. (Or very inclined to leave reviews on DLC. The main game is at 2,760 reviews, 80% positive.)

FOUR. The soundtrack has sold much better than we expected… although our expectations were extremely low. This goes against the received wisdom we keep hearing, that soundtracks don’t sell on Steam. We did sell it as part of a bundle with DLC, which must have helped. And it is, I think, an exceptional soundtrack that does a lot for the atmosphere of a game where atmosphere is more than averagely important. But we were still surprised. Make of that what you will.

Cultist Simulator: Life on Mobile

On April 2^nd, we released Cultist Simulator on iOS and Android.

Back in June, Lottie already did a gigantic data post on how that went, so I’m just going to re-summarise her main points, and then talk about what’s happened since.

What happened then

• We partnered with mobile publisher/porting house Playdigious. ‘Partnered with’ means they took care of porting to iOS/Android, the release, and post-release support, in exchange for a 50% revenue split. It’s been a very good experience and we would unhesitatingly recommend them.
• Sales have been decently good. We’ve proven to ourselves that premium/pay-once can still make enough money for a distinctive game from a small studio to be worth the effort. F2P would almost certainly have made more money, but isn’t our scene and would have required a much bigger ongoing commitment.
• We launched in Simplified Chinese as well as English (see Localisation, below). This proved to be a very good decision.
• We co-ordinated with Apple ahead of time and benefited from featuring at launch… but didn’t get featuring in the US. We suspect that’s because ‘Cultist’ suggests guns and Waco in the US, in a way it doesn’t in other parts of the world.

 

What’s happened since

Post-launch sales. As of now, 11 months since launch, Cultist has sold 120K+ units across iOS and Android, generating €350k+ net (split 50/50 between us and our partner Playdigious). This is, like most of our numbers, very respectable for a micro-indie, but still only crumbs from the big mobile studios’ tables. Lottie’s best-case estimate for Year 1 was 100K units, so we’re pleased with how it’s gone.

Reminder: always, always record your own estimate of your sales, even when you don’t really have a clue. Educated guesses get more educated every time you have data to compare them to. If you share your guesses internally, that helps keep you accountable and allows you to apply a bit of wisdom-of-the-crowd.

User response. Our reviews stabilised quite quickly around 4.7 on the App Store and 4.6 on Google Play. This is higher than I expected (and feared) because Cultist is notoriously divisive, partly because of its deliberate lack of a tutorial, and mobile audiences are generally expected to require more hand-holding than PC audiences. We do get a handful of one-star reviews saying ‘lol what the hell’ but they’re surprisingly rare. Here’s my hypotheses about why:

• We did raise this concern when we talked to Apple. They surprised us by saying it wasn’t too much of a concern, because customers self-select, but we should be clear in the app description that it wasn’t an easy game. So we were.
• The price (£6.99 / $6.99) is high for a mobile game. A few years ago it’d have been too high (and we applied a hefty launch discount) but higher prices are more common than they used to be for quality games on mobile. So a lot of people buying CS were looking for something unusual, and/or buying on the basis of word of mouth, and went in ready to give it a proper try.
• We’ve realised since launch that there really is almost nothing like Cultist Simulator on the mobile stores. In fact, there is a real shortage of thoughtful single-player premium games – they exist, but you have to dig to find them. This is obviously because the real money is in F2P multiplayer. It does mean there’s probably still an opportunity for smaller studios like ours.

 

China. Chinese players have responded extremely well to Cultist. We actually have a 4.9 star average on the Chinese App Store (Google Play isn’t, officially at least, available in China.) We’ve talked to Chinese fans and developers to try to get a sense of why the reviews are so good; but honestly the thoughts we have are so speculative and subjective, and I’m so wary of generalising about another culture, that I’m wary about sharing them.

However, here’s what our contact at Indienova (who worked with us on the localisation, and with Playdigious as the Chinese PR agency) thought might be reasons for our Chinese success in general:

“0. That’s the most important one, the game itself is content-rich and good enough.

1. Next, China has a larger population than other countries, obviously.

2. It’s very different to the other games on the ranking list you may have noticed.

3. We reached the core community who are interested in Cthulhu-like things very much accurately, they help us to build a very good public praise. (So we have a good base at the very first time)

4. Apple featured it and some biggest influencers recommend it once the mobile version comes out. (Then we have a fast growing spread)

5. For me, the Chinese localization is not bad but far from perfect at this moment (we are still working on it). [Note: this was back in March, and the localisation has gone through two rounds of improvement from our excellent volunteers.] However, I’m confident that it’s much better than average and it should help a little. (So we should be able to keep a high scores for a longer time.)

6. We have a very energetic Chinese players community now, we got 800+ followers on Weibo and 300+ players on our QQ group. We got popular in the timeline of Weibo.com too if you try to search “密教模拟器” on it, ppl talked and shared this game and it will help to bring some more new players.“

( I mentioned earlier that the title ‘Cultist Simulator’ doesn’t seem to have gone down too well with Apple in the US. Our very sensible volunteer loc team were concerned about the connotations of ‘cult’ in China, so they used 密教. We’re told that 密 is something like ‘secret, mysterious, inmost’ and 密教 is a term for the esoteric Buddhist traditions. I like ‘Esotericism Simulator’, actually. Apparently the original Sanskrit term for these traditions means something like ‘Diamond Vehicle’, and I think I like ‘Diamond Vehicle Simulator’ even more. )

A wrinkle. We did have one real headache with the Chinese mobile release, and I think it’s worth talking about – because it’s another really good example of how if you’re a small dev, then sensible commitments honestly made can box you in.

We began localising Cultist to Simplified Chinese in 2018. Meanwhile, I kept adding free updates – including changes to existing content – to the game. We branched off a stable version for Playdigious to work on porting to mobile. The translators stopped work and then started work later on that version, because it’s not practical to translate constantly changing text. Once I’d finally finished updating the PC version of the game in May 2019, the translators started translating the PC version… which still took months to do, because there was loads of new content.

This meant that the mobile game launched with Chinese loc, and without any of the updates. It also meant that the PC version got DLC and updates, but didn’t have Chinese loc until six months after the mobile launch.

This in turn meant that mobile players were unhappy that they had an older version of the game, and Chinese players were unhappy that they weren’t getting a translation on PC. A minority took to Steam to voice their displeasure, posting negative reviews accusing of us of reneging on our promises to translate the game into Chinese, or of employing an incompetent loc team. We got enough negative reviews that it had a noticeable impact on our review score.

And it was almost impossible to communicate this complicated bundle of reasons for the delays usefully, across the language barrier, to an audience that wasn’t so used to developers communicating openly. I tried posting developer responses in badly Google-translated Chinese, and it didn’t seem to help much. I got grumpy and posted some more defensive responses, and that definitely didn’t help.

That was a silly thing to do. It was frustrating that people thought we were deliberately holding back a Chinese translation for our own nefarious reasons. But it’s never a good idea to let your emotions get the better of you when talking to your community.

(I should add that Indienova warned us that releasing earlier on Chinese on mobile might cause this issue, and I failed to take them seriously enough.)

In summary, we’re very glad we did the mobile port… but it wasn’t zero-risk or free money. Playdigious took care of the majority of the work, but it ate up a lot of Lottie’s time and a little of mine too. And it had other consequences – from the wrinkle with the Chinese loc that I mentioned above, to the complexities of scheduling releases on multiple platforms. When you’re a small team, complexity takes time.

And I think it’s worth saying that, from the start, I’d designed Cultist with an eye to porting it to touchscreen devices. I didn’t know enough to do a particularly good job of that, but I actually got a Cultist build running on Android as long ago as the alpha, just as a basic sanity check for whether it worked. This is why, for instance, Cultist doesn’t use tooltips. When I built Fallen London, I put tooltips everywhere, because although it’s a web app I was a PC gamer and tooltips are a good way to drill down into information. When FL was ported to mobile, this was a colossal pain in the bum, because there’s no way to mouse over and get a tooltip, and tap or tap-and-hold are different kinds of interaction.

So if you’re thinking about distributing your PC game on mobile – and if you’re developing on Unity, that’s much easier to do – think about it early.

Localisation

In 2019, we released Chinese and Russian localisations of Cultist Simulator.

This was also something I thought about early. If you’ve been following our work, you might recall that I tried to impose a maximum word count on myself for Cultist Simulator of about 70K – or 25% of the length at launch of Sunless Sea. This was in part because it had been challenging to localise Sunless Sea (the initiative was abandoned after I left) and impossible to localise Fallen London, and I wanted to keep the count low. In fact, with free updates and DLC, Cultist is almost double its launch size… but that was still practical to localise, though not exactly quick.

I’m using ‘localisation’ and ‘translation’ pretty interchangeably in this post, which will probably upset some experts, sorry! More properly, ‘localisation’ refers to the whole process of making a game suitable for an audience in a different culture.  That can also mean changing UI to suit local preferences, or to accommodate linguistic quirks – or it can mean a more all-encompassing process of culturalisation.

Offworld Trading Company, a game by US-based Mohawk Games, courteously offers a UK English localisation. I believe the only change in the UK-based localisation is that ‘aluminium’ is spelt with the additional ‘i’. God save the Queen.

But for Cultist Simulator, it was almost all about the translation. We did have to make some button sizes a bit more flexible, because Russian tends to run a little longer than English (while Chinese runs shorter). I’d also hardcoded a lot more of the UI text than I realised – because I’m so used to operating in a monolingual environment – and we had to put significant effort into sorting that out (or rather get Chris Payne to sort it out for us).  But mostly, it was the difficulty of translating a large body of deliberately allusive and elliptical text into very different languages with very different cultural references.

Which languages?

We wanted to begin with two or three languages. That would limit our risk, but if we just picked one, we’d probably not learn as much or get as good a sense of how worthwhile localisation might be. This was Lottie’s project, so she made the decision about which. Here’s how.

She considered EFIGS (English + French Italian German Spanish). The other four are relatively easy to localise from English, and my work’s often been popular in Germany.  EFIGS is also a requirement for some distribution deals. But four languages was a lot; and crucially, their home countries all have a high percentage of good English speakers.

She considered Brazilian Portuguese. Brazil is often talked about as an interesting emerging market. Portuguese is the most widely spoken European language by native speakers after English and Spanish, largely because Brazil is so big. And Brazil has a low percentage of English speakers. But we hadn’t seen much interest in my work in Brazil; retail prices are on the low side because incomes are low; and there’s historically been issues with piracy there (because games were banned, then heavily taxed).

She briefly considered Japanese, because we’d heard that there’d been a significant revival of Lovecraft-esque culture there recently, and because it’s a populous country. But translating to Japanese is very expensive, and the market there is of course very mature and competitive.

But when she looked at our sales data, she noticed that there were a disproportionate number of sales in Russia and China compared to other countries with few English speakers. ‘Disproportionate’ was only a little over 1% for Russia and 3% for China, but that still stood out. She remembered a similar effect in the data for Fallen London and Sunless Sea, too. It made sense. My previous work seems to have resonated with audiences in Russia (and I think in some other Slavic countries, though it’s very hard to generalise here). And China is just immense. I constantly have to struggle to remind myself that the population of China is larger than the USA, Russia and all of Europe put together.

Of course, Russian is relatively difficult and expensive to translate to from English. And Chinese is about as difficult and expensive as it gets.

Lottie, however, asked around and was recommended this volunteer initiative by Tanya Short of Kitfox. They’re selective about who they work with, but it offers the possibility of a Chinese translation for the irreproachable price of 0 USD.

I’ve mentioned Indienova before, and the relationship is slightly complex, so to clarify:

• Indienova put Weather Factory in touch with volunteer translators who localised the whole game to a good standard. Indienova handled project management, and the volunteers did the bulk of the actual translation. As far as I can tell this is one part enthusiasm and one part business development.
• Playdigious, our mobile publisher, contracted Indienova to provide PR support for the mobile launch of Cultist Simulator. This was a happy coincidence (except, I guess, that their loc program was successfully generating relationships!)

 

We also, ultimately, contracted QLOC, a Polish company with a good rep, to localise to Russian, for about 20K USD, of which our then-publishers, Humble, paid half.

So, translating to Chinese. That was a journey, because not only is Cultist allusive and enigmatic and poetic and all that jazz, but the references are generally to European culture and mythology.

Is “pine” in “pine and knife” refering the plant or the emotion?

Pine: can you let me know the context on that to check? But if it’s the Pine-Knight, or the Pine and the Knife, then it’s the tree (which is sacred to the Phrygian goddess Cybele, if that’s relevant)

And then, of course, we had the usual problems with words with multiple meanings in translation.

What is a “Host” in Grail’s ascension? The apostle or the sacred bread?

‘Host’ here references the following meanings: – sacred bread – large number of people – someone who extends hospitality. The Grail ascension requires a feast with a large number of attendants where the distinction between ‘host’ ‘guest’ and ‘meal’ is blurred. Grail lore also often references Christian lore. In this case, then, the Host refers to the increasing crowd of guests, hosts and victims coming to the Feast (‘the Vitulation’)

Then we had questions that I couldn’t answer without sounding like I was high…

By ‘The Cross is imaginary‘, do you mean ‘The Carapace Cross never existed at all’ or ‘Now the Carapace Cross can only live in one’s imagination’?

‘The Cross is imaginary; the change is not’ is a repetition of Teresa’s ”The wine is imaginary,’ she explained, ‘but the cup is not. To navigate the Mansus one must understand the distinction.” … so the same terms should be used to translate this line. (What it means, more or less, is that ‘real’ vs ‘imaginary’ is a less simple distinction than is commonly understood, and though nothing in the Mansus is physically real, some things have a more primary and current reality.)

…the questions I couldn’t answer without information that I realised halfway through would be basically useless…

What is ‘the Labhite’?

‘Labhite’ is an invented word with a Hebrew derivation. To speakers of European languages, the Hebrew derivation will tend to make it sound ancient and there are cognates in multiple European languages with ‘Lion’. But it’s an invented word that most people won’t get references from.

…and the questions I couldn’t actually answer at all. Watch me trying to cover my embarrassment by sounding fancy.

Does the name John Sonne refer to Sun (the Sun-in-Splendour), Maevelin to Wanderer (the Vagabond), Leo to Lion (the Lionsmith), Corvino to Crow (the Beachcrow)? Any other references like that? We could try to reproduce those references in our Chinese translation.

Yes to all of these, except that I can’t remember whether I intended Maevelin as the Vagabond or not. Let’s assume I did 🙂 John Sonne is also a reference to Ben Jonson, the seventeenth century English playwright. The Humours of a Gentleman is a reference to Jonson’s play Every Gentleman in his Humour, but the characters are based more closely on the ones in Jonson’s Volpone. This obviously doesn’t need to be in the translation but I thought I’d mention in case it was useful for context.

 

We had a shared spreadsheet in which, ultimately, I answered three hundred and twenty seven of these questions over a period of six months.

Russian was a less arduous process. This was because there’s less of a cultural gap, and because we’d learnt a lot about how to work with loc partners during the Chinese translation (a lot of the questions we did get had already come up in the Chinese translation).

Just as with the mobile port, then, someone else did the bulk of the work, but it wasn’t zero cost or risk at our end. It soaked up weeks of both my and Lottie’s time. I guess it’s worth mentioning in passing that I had to fight to suppress my prima  donna instincts, too. I chose the words in Cultist very carefully, and seeing them translated into a language I couldn’t speak, with no real sense of how they came across, was painful. Talking to people on both loc teams who were attentive, thoughtful, and fans of my previous work – that helped a lot. But I basically had to nut up and get over myself.

So how did it go?

We released beta Chinese loc in September 2019, and full loc in October 2019. In the fifteen months after the release of Cultist and before the release of Chinese loc, 3% of our sales were in China. In the five months since then, 36% of our sales have been in China. It’s currently the country we sell the most units in (#2 is the US with 33%, #3 is the UK with 5%).

We released Russian loc in November 2019 (beta at the beginning,full release on the 25^th). In the seventeen months before then, Russia accounted for 1% of our sales. In the two months since then, it accounts for 3% of our sales. That’s a lot less dramatic. But there are a couple of giant provisos there. First, it’s a very short period of time which includes two major Steam sales. Second, we released the Russian version with almost no fanfare – no PR, no Daily Deal on Steam as we did with China, very little marketing. (This wasn’t intentional – it was for reasons I’ll talk about later in this post.) So far the additional sales have probably made us back about 50% of the cost of localisation into Russian, but I’d expect we’ll have recouped the whole cost by the end of this year.

All told, then, we’re glad we put the effort into localising Cultist, and we’ll likely do it again – with future projects, and with Cultist in more languages. We’ll probably start with the ones at the top of this section!

Cultist on Amazon Twitch Prime

This is a quick one. We distributed Cultist on Twitch Prime. It was a straightforward deal negotiated by our then-publisher, Humble: a flat fee in exchange for a DRM-free build of the game that they could distribute to Amazon Prime users who logged into Twitch that month. I don’t think I’m allowed to say exactly that the fee was. I can say that it was by no means game-changing, but was more than we make from Cultist Simulator sales in an average month, even after the 30% publisher cut.

Like everything else here it wasn’t quite zero-cost. We had to prepare a unique build for Twitch with some specific requirements. Unfortunately there was a bug in that version. We sent Twitch an updated build, but it doesn’t seem like there’s really a process for distributing updates. So we still occasionally get support requests for help with that build. We’ve put up a manual patcher here.

As hassle goes, though, this was minor, and it was basically a straightforward deal that we’re glad we took.

Awards and Recognition

This is pure puffery, so I’m going to keep it short.

In March, Cultist Simulator was nominated for two BAFTAs – Innovation, and Debut (we were shortlisted for Narrative, but didn’t get a nomination). Did we win anything? Good Lord no. These were the BAFTAs. Debut went to Yoku’s Island Express (from another small studio, but brilliantly executed, with Overwhelmingly Positive reviews on Steam), and Innovation went to Nintendo Labo. (Narrative went to God of War, obviously).

In July, we won three Develop Star awards – for Best Innovation, Best Game Design, and Best Microstudio. Lottie and I had actually gone out for a quiet dinner instead of turning up to the ceremony, so poor Claire Sharkey, who was handling our marketing at that point, had to give not one but three impromptu speeches for Weather Factory.

Also in March, this happened.

Hiring and Growing

We wanted our next project to be a slightly more ambitious and polished game than Cultist; and we wanted to run two projects simultaneously, so we didn’t have all our eggs in one basket. That meant growing the team.

We did that with some trepidation. It’s a big step to go from ‘romantic also business partners in a flat’ to ‘first actual employee’; but we had a good war chest and far more work than we could handle. So we went very carefully. Lottie and I wanted to make sure we hired competent people with a high degree of integrity whose skills complemented ours, and who could cope well with a high degree of autonomy. We were also keen to hire with an eye to diversity.

The most urgent pressure was to find someone to take over PR & marketing. Lottie was handling that, but she had too much to do already, and we wanted some outside expertise. So our first hire was Claire Sharkey, who we knew and respected already, to handle that side of things. We also needed some Unity expertise and UI expertise – I’m a software developer, but it’s not my strongest skill, and my UI skills are pretty dreadful. We hired Hannah Rose, a smart and versatile Unity developer with a good portfolio, for that. And finally we needed someone to support internal tools, so we could develop content faster, and to work on the unannounced second project. Marc Gagné, a community stalwart who’d built a number of fan projects, was a good fit for that. (We were hiring for a writer, too, when events overtook us – see below for more about that). With those hires made, we were that rare thing in game development, a majority-female development studio.

We got our shortlisted Unity candidates to spend a half-day building prototypes to demonstrate their skills. We paid them their (half) day rate to do that. This gave them an early signal that they could trust us, but it also meant we got code written in earnest that we could use to compare candidates properly. I recommend it. It’s not especially cheap, but it’s a lot cheaper than hiring the wrong candidate.

We also made some unusual commitments – motivated in part by altruism, but also by pragmatism. I think there are good reasons to do all these things, at least at small company size. I know they may not scale.

• We guaranteed a no-crunch policy, and were explicit about how. I’ve been outspoken about this for years, at this and my previous studio. I think the evidence is pretty clear that prolonged habitual overtime damages productivity.
• We committed to being internally transparent about salaries. This was a slightly alarming step, but people always know in the end through the grapevine – or worse, think they know. Publishing salaries internally is a sign of trust, and ensured that if there was an unfair imbalance, we’d be accountable.
• We committed to a profit pool for 2020 – something I’d instituted at my previous studio, that had worked well. Again, the moral argument for profit pools is pretty straightforward, if not very sophisticated. It seems right that employees, as well as shareholders, should benefit if the company does well. But pragmatically, my understanding and experience is that it’s a better way to motivate knowledge workers than either high salaries or performance-related pay. People stop being motivated by high salaries quite quickly after a hire or a  raise –  it doesn’t feel significant any more. And if anyone’s found a good way to run performance-related pay fairly in a small business with highly subjective metrics, I’ve yet to hear about it.
• And we added five zero-notice days off per year (‘duvet days’, ‘mental health days’). I’ve always thought this was just a good idea for workplaces, at least where it’s practical.

 

Here’s the Staff Handbook we ran with as an addendum to our contracts, which discusses the points above and more. I’ve added a least-restrictive Creative Commons licence to it, so if any part of it is useful, feel free to use it.

Industry Contributions (and things you might want to do)

Lottie and I both were very conscious that we had had a fantastic first year. We wanted to pay it forward.

Mentoring. My thing was a mentoring scheme, because I had no clue what I was doing when I started out in this biz, and I wish I’d had something like that to help me learn more quickly. We had an informal mentoring scheme available in 2018, but the problem with an informal mentoring scheme is that no-one really has any idea how much time to expect or what to ask for, and from the mentoring side, it’s hard to budget time. So in 2019, we formalised it and spent some time carefully selecting five mentees. The scheme is now defunct, unfortunately (see later in this post for why) but here’re the original details, in case you want to run something similar.

Coven Club. Lottie’s thing was Coven Club, which actually began in 2018 but hit its stride in 2019. Lottie wanted to provide a supportive environment for women working in the games industry – something more welcoming and special than a corner in a pub – so once a month she rented out an unusually nice space and provided snacks and prosecco. Coven Club ran for a year and then, unfortunately, we had to put it on hold (again, see later in this post for why) but the original details are here if you’re interested in doing something similar and find it useful. It got a really warm response, and it’s a shame it’s gone.

Wings Fund. Lottie also spent some time volunteering for Wings Fund, a venture looking to provide funding for projects for diverse teams. If you’re looking to help with something like that, or if you think you might qualify for funding, they’re here.

Salaries Spreadsheet. Lottie also set up and maintains a spreadsheet of UK games industry salaries. If you’re hiring, or trying to get hired, you might find it useful. If you’d like to help, do please add your own details via the linked Google Form in there.

We also got involved with Ubisoft’s Open Innovation partnership programme for indies – a really interesting initiative in which indies share their unique experiences and Ubisoft offers expert advice, or other assistance, in return. I gave a day-long workshop at Ubi Berlin, in exchange for expert assistance with our second project, Procopius. In the event, the expert assistance was indefinitely postponed (see later in this post) but my impressions of the programme remain broadly very positive, and I’d recommend it.

Finally, we gave a bunch of talks (GDC, Develop, Disco Montreal,  GameDev.lv in Riga, Devspace London, yadda yadda). We posted the slides for our talks here.

The games industry is still new, chaotic and uncertain, especially at the indie end. An ounce of action is worth a ton of talk, but a lot of good intentions in our industry get diffused into discussion, and peter out. If you’re enthused by anything I mentioned above, please consider doing something like it yourself. You don’t need permission from the Twitterati or the establishment to do something like this – it’s not presumptuous to try. Good luck.

BOOK OF HOURS, and the Kickstarter

The next game we decided to make after Cultist Simulator was not the one we’d expected.

In January 2019, I was finishing off the extremely savage New Game+ endings (‘not just difficult, but unfair’) for Cultist, and I found myself hankering to work on something more relaxed. So I tweeted:

 

It got an astonishing response. ‘This is what’s been missing from my life,’ people constantly said. In a follow-up tweet, I committed to trying to convince Lottie to let us make the game if we got 1000 RTs. It took about 24 hours to get there.

“….we have a road map until April for Cultist”, I wrote in a follow-up blog post, “and we had one and a half other projects we were going to do pre-production on real soon. So assuming we make That Damn Librarian Game, then we will need to rejig our planning, and that’s not something we want to promise on a whim. BUT after we get Christopher’s Build / the Major Victories (free update on Jan 22nd, folks!) out the door, Lottie and I are going to sit down and look seriously at how we could make this work. It’s not out of the question that you might see a Kickstarter this year.”

It did mean reworking a lot of planning. But that level of enthusiasm isn’t something a creative business readily ignores. And besides, we really liked the idea.

So we contracted Adrien Deggan for some initial concept art and Catherine Unger to provide us with a visual direction. I started work on the design. Lottie and Claire began spinning up a Kickstarter, scheduled for September. Hannah and Marc began working on the UI and on the tooling for a new content framework. We had a promising initial prototype which we reckoned would be demo-ready before the end of the campaign. And after months of referring to it as That Damned Library Game, we chose BOOK OF HOURS as a title.

We wanted to be a bit more ambitious than we had with Cultist Simulator –  probably 1.5x the budget, maybe 2x by the time we’d finished prototyping. That meant about 200-250K GBP. We chose 100K GBP as a Kickstarter funding threshold – not a small ask, but I’d raised almost as much for Cultist Simulator as a solo dev. If we got overfunded to the point where we covered the budget, great, if – more likely – we only raised 100K, we were confident covering the rest with our Cultist monies. And if we failed to get funded, we’d take that as a sign that the pitch wasn’t as strong as we had thought, and we’d rework it.

 

We were very jittery, but about as confident as you can be on the eve of a Kickstarter. Lottie and I between us had worked on five Kickstarters, four of them successful. We had a strong community and a lot of buzz. We’d shared the pitch with hundreds of potential backers and got buckets of helpful feedback. Lottie and Claire were working through a rich schedule of publicity  and promotional stuf for the two weeks before launch, to get the buzz building towards launch day. I started to hope that it might be our most successful KS ever.

 

Why the Kickstarter Didn’t Happen

On August 27^th and 28^th, a larger competitor went public with a smear campaign against both me and Lottie. It hit the week before our Kickstarter, and on the one-year anniversary meeting of Coven Club (Lottie’s feminist initiative, for which she’d just got funding after a year of Weather Factory paying for it).

Most of you reading this will already be aware of these events, in outline if not in detail. I want to talk about it dispassionately and usefully. The story of this year makes no sense without it, but also, it’s extremely rare for a small business facing catastrophic PR to talk openly about the detailed effects. Posterity, you’re welcome.

It might be difficult to understand what follows without understanding the mechanism of operation of this kind of attack. That mechanism is this: a minority of people will believe the smears, think badly of you, and act accordingly, but the majority will mostly be worried that other people will think badly of you, and distance themselves accordingly.

Commercial effects

The BOOK OF HOURS Kickstarter never ran. We had to cancel it two days before launch. The team’s last two months’ work was rendered obsolete.

All three of our new employees left. When the extent of the damage to the company’s reputation became clear, we offered them either or both of (i) a week’s paid leave to think things over (ii) a bonus payment of one month’s salary in addition to their month’s paid notice if they chose to resign. They were caught in a situation that they hadn’t expected, and we wanted to give them all the time and space we could.

In the event, all three left. They liked their jobs (“Working here is honestly a dream come true,” I’d been told in a one-on-one the week before the attack) but it was too traumatic and miserable, and the potential damage to their careers was too great. (“I thought I could weather it, but I can’t,” said the last to leave.) We don’t blame them, and they left with our regretful blessing.

BOOK OF HOURS was indefinitely postponed. We couldn’t make the game as originally envisaged without the two developers I’d hired to make it, and I was incapable of working for some months in any case. (We’ve since resurrected the idea in a different form with a much lower budget.)

‘Procopius’, our more ambitious second project, was permanently cancelled. There’s no way we can make it now.

Our community was devastated. We were particularly vulnerable to this form of attack because we’d invested so much in a friendly and transparent relationship with our fan base, many of whom now didn’t know what to believe. A civil war erupted inside our community. I had to leave the fan Discord after one of the mods joined in the personal attacks. All told, our mailing list and social media followings – the result of two-three years’ promotional work – were reduced by on average around 25% in the space of a month. (If 25% doesn’t sound as bad as you expected, imagine losing 25% of your home or your salary.)

Around half of our business partners terminated their contracts and/or their upcoming deals with us. There are legal reasons why I can’t be specific about which and how, here.

Revenue from Cultist Simulator went into free fall. Here’s our Steam revenue in the two weeks before the smear campaign hit, and the two weeks after. Any indie looking at this will experience a violent lurch in the pit of their stomach. Sorry about that.

 

Valve are still a little nervous about sharing sales figures at this level of granularity, so I’ve had to cut off the labelling on the Y-axis, but the delta should be clear: in the course of a week our daily sales plummeted by 75%, and hit their lowest day ever. We had no idea whether that was rock bottom, or whether they’d continue to drop.

If we hadn’t had an exceptional previous year, Weather Factory would have gone out of business rather quickly. (Of course, if we hadn’t had an exceptional previous year, we wouldn’t have registered as a threat to our competitor’s prestige.) In fact, without one stroke of unexpected good timing, we would still probably have gone out of business. I was largely unable to work by now, and would remain so for some months. Lottie was better, but demoralised and overwhelmed nevertheless.

This was the stroke of unexpected good timing:

 

On September 12^th  2019 we released the Simplified Chinese localisation of Cultist Simulator on the beta branch. This is what it did to our sales graph. We saw another uptick the following month, when we released the full public localisation, by which point Lottie was in good enough shape to do some promotional work.

Eventually, as the months passed, it gradually became clearer what was going on, or people decided they didn’t really give a stuff either way, and our sales recovered to about where they had been. But it was a close thing. The loc release had been scheduled for months. I had often been quietly sceptical of the benefits of a Chinese release, given the time and effort involved in translating the text. It had been Lottie’s initiative from the start; she was right; and it saved the studio.

 

Effects on industry contributions

Mentoring. All but one of our mentees withdrew from our mentoring scheme. We had to shut it down.

Coven Club. Coven Club was the subject of some particularly vitriolic online personal attacks. Lottie had to put it on indefinite hold.

Wings Fund. Lottie was asked to step down from her volunteer work at Wings Fund on the basis that (unspecified) allegations had been made against her character.

Talks. All our speaking invitations were withdrawn.

I’m happy to say, however, that the salary spreadsheet is still going as of the time of posting.

 

Personal effects, a.k.a. health and wellbeing Year 2

The most significant of the effects of this kind of experience is, unmistakably, the toll it takes on your state of mind, and that shaped the last four months of 2019 for us. It’s especially difficult to talk about this part dispassionately. So I am going to try for a tone of dry flippancy. But I don’t want the tone to suggest that this is a trivial thing, so I’d like you to be  aware that as I type this, I am remembering last August and September, and my hands are shaking, and yes, I did indeed have a full-on couldn’t-look-after-my-child, emergency-mental-health-referral, Lottie-had-to-explain-to-the-GP-because-I-couldn’t-speak-coherently, Internet-access-restricted breakdown from which I am still recovering. I don’t recommend it. I’ve been through a rancorous divorce, and I’ve been through my brother killing himself. This was worse than either.

I was largely out of action until the end of the year, I still find it difficult to communicate with people unless I know them very well (if you’ve emailed me, sorry, I have to work up to responses!) and I’m unlikely ever to speak publicly again, but I’m now capable of working at my usual pace on light medication.  Lottie’s morale and health have suffered (she has a problem at the moment with recurring nightmares), but she was less dramatically affected and she’s mostly recovered. Some days we’re quite cross, most days we’re pretty cheerful. We’ve mentioned previously that Lottie’s mother was also targeted for some unpleasantness, but she’s a tough lady, she was on the periphery of it, and she’s fine.

On the flip side, the work-life balance issues that I mentioned in last year’s post have been less of an issue lately.

 

More about this sort of thing

• https://www.youtube.com/watch?v=OjMPJVmXxV8
• https://en.wikipedia.org/wiki/So_You%27ve_Been_Publicly_Shamed

 

Year 2: Where We Are Now

In summary, in 2019, we released Cultist on iOS and Android. We released two more pieces of DLC, the soundtrack, and the Anthology Edition. We released Chinese and Russian localisations. We won some more awards. And for a few months, we were that rare bird, a studio of five with a majority female team and no gender differential in pay.

Here’s where we were at the end of 2018 compared to where we are now.

End of 2018	End of 2019
two full-time employees	we were five, but now we’re two again and we have no plans to grow
two regular freelancers	we’re not working with anyone right now, but probably will
two cats	two cats, though one now has a limp
three awards	eight awards
enough revenue from CS sales to cover our costs and then some, most months	enough revenue from CS PC and mobile sales to cover our costs, most months
a merch store generating a bit of revenue on the side	a merch store generating a bit of revenue on the side
one moderately exciting announcement you should see this month	(that was the mobile port)
One secret guest writing project we hope to announce this year	Crikey, that! That got cancelled for NDA’d reasons unrelated to anything else I’ve mentioned. A pity, it was cool
a next game semi-pre-announced with a lot of buzz	This was BOOK OF HOURS: we’re working on a smaller version with a tentative launch date of 2021
a soft deadline for more Cultist DLC in April	Yup, we released that! (in May). We have more to say on the topic of DLC, but haven’t announced anything officially yet.
……..a completely bonkers long-term plan that I have yet to convince Lottie is wise.	This was ‘Procopius’, which we can’t make with two people, so it’s dead for good, alas.
…	A podcast on odd narrative stuff
…	35K words of a book about ten years in indie gamedev (“A world that’s as glamorous as software development, as well-regulated as jazz, and as stable as a balloon full of frightened cats”)

Here’s a similar 2018-2019 comparison, but in money.

 

And here’s our revenue breakdown on Steam for all products to date. (NB this is before Steam’s 30% cut). Revenue breakdown on other platforms is in similar proportion, and in total runs about 10% of the Steam revenue.

 

Those are numbers since Cultist’s launch, i.e. 20 months of sales. My prediction for year 1 PC sales, based on Jake Birkett’s formula  and our week 1 sales of 34K units, was 157K units. Actual Year 1 sales on PC, all storefronts, was around 110K units. So 157K was much too optimistic, but still illustrative. The Birkett number still looks useful as a rough estimate, and it’ll help me refine next time.

Mobile revenue: here’s Lottie’s detailed breakdown. The bottom line for ‘units shipped on mobile’ was 59K when she wrote that post, four months after launch. It’s now been eleven months since launch, and the total of units shipped on mobile – as I mentioned above – is now 120K, or €350K + net (of which 50% has gone to our partner Playdigious).

We’re ten thousand words in, and nearly done. Here is the traditional kitten picture:

 

This is Sulochana, one of Lottie’s two beloved ragdolls. What isn’t visible in the picture is the damage to Sulochana’s left front wrist. On 31^st of May 2019, the anniversary of Cultist’s launch, when Lottie and I were drinking prosecco and marvelling at how well everything seemed to be going, the wind blew open a carelessly closed door and Sulo got on to the balcony and leapt from the sixth floor. She survived, and is mostly recovered, but she’ll walk with a limp for the rest of her life.

That’s pretty much where Weather Factory is now. We were grievously wounded and almost didn’t survive, and we’ll never be what we were, but we go on, and life’s not so bad.

Shorn of metaphor, that means that Weather Factory is just Lottie and me, probably for good. We won’t ever be a two-project studio, and we won’t ever work on the ‘ambitious and unusual [thing] on a larger scale’ that we had planned. But we will continue to work on small, carefully scoped, experimental narrative games. I can’t imagine giving talks or mentoring or hanging with the scenesters any more. But we hope to keep sharing data like this; I can probably manage some more design streams; and I intend to keep making games until someone breaks into my house and forces me to stop.

Our plans for 2020, then, are much more domestic than last time:

1. Spend some of that Cultist money on a house (I’m 48 and renting)
2. Get married
3. Get BOOK OF HOURS to beta. We’re very wary of running another Kickstarter, but we haven’t quite ruled it out.
4. More book, more podcast, more fun merch.

 

Thanks for reading to the end! I hope it’s been useful. And if you’ve stayed with us through the rough times, we really appreciate it, both of us. A heartfelt thank you. It’s probably meant more this year than any time before or since.

Here’s our mailing list.

Here’s our podcast.

Here’s our YouTubes.

Here’s Cultist Simulator.

And here’s BOOK OF HOURS.