We just ran a Kickstarter for Sunless Sea, our Fallen London spinoff game of exploration, survival, and loneliness. It got FUNDED... 100K worth of funded against a 60K funding goal. Now that the rainbow haze of delight has started to recede, we wanted to share some of the lessons we learnt - in the spirit of paying it forward to other devs who may run Kickstarter campaigns of their own.

There's a lot of very good advice out there already that we don't need to duplicate - so we're focusing on the lessons which are most specific and unusual to us. We hope you find it illuminating.

Why we went to Kickstarter

We're an existing studio, right? Not a big studio, but we've been around for a while. Why did we need outside funding?

Fallen London rarely makes enough to support a four-person team + rent + servers in any given month. In the past, we've plugged the gap with client work. To make Sunless Sea, we realised we'd need to commit the team for six months. The last year's been pretty lean - we had to lay off three employees at the end of 2012 - and there would have been a serious risk of not making payroll before we finished the game, let alone sold enough copies to support ourselves. Kickstarter was the obvious option, and we've run a successful KS campaign before.

There are other advantages too. One, although it's hella hard to get Kickstarter press, the KS site itself gives you exposure. Going to Kickstarter was a good way to, er, kickstart our marketing campaign. And two, we've always developed iteratively and in the open, and we've always listened to community feedback. Involving the community from the start was a good way to destruct-test our ideas, our game, our message.

But we knew it wasn't going to be easy. Our first, successful KS was 400% funded - but it was a much smaller project than Sunless Sea. (Funding goal £6.4K, final total 28K, Sunless Sea had a threshold of 60K). The other failed outright - it made 5K against a threshold of 15K before we cancelled it. Kickstarters are all or nothing. If we had reached the end of the month and found we had nothing, it'd have been a difficult situation.

We picked September for our Kickstarter month. Any earlier, and we wouldn't have had enough prep and prototyping time. Any later, and we might have stalled on the runway.

Primordially early user testing

I've just gone back through my emails to see when we started work on this, and it turns out in February I proposed a KS of about this size and shape as a 'genuinely left-field idea'. (Paul described the project as 'Hatline Miami'.)

We thought it over, tossed around ideas, but the failure of the Below Kickstarter - a good game with a good pitch - had left us nervous. Time is the life-blood of a small team, and we couldn't afford to waste it on a project that might not get funded.

Fortunately we have a secret weapon: the Enthusiastic Urchin! He's a character who lives on opt-in cards that offer very small amounts of in-game currency in exchange for answers. We've used it in the past to gather demographic information. This time, in June, we had him proffer an imaginary suitcase with four possible options: a collection of short stories, a deck-building digital card game, a comic, and a naval exploration game. Which of these, we asked, would Fallen London players pledge to as a Kickstarter?

We never really thought the short story collection was likely, and no-one seemed interested. The card game, rather to our surprise, also did very badly. The comic - and we'd been producing an experimental comic by then - and the game which would become Sunless Sea both did about equally well. But we ran the numbers and we just couldn't find a way to make a comic work as a Kickstarter. So Sunless Sea it was.

This meant we could greenlight the considerable internal effort needed for:

The prototype

The next step wasn't to run a Kickstarter! The next step, in June, was to build a very early prototype to get a feel for the technology and the core gameplay, and to get better numbers that I could feed into a project plan for estimates. The worst possible outcome wouldn't be a Kickstarter failure - it would be a Kickstarter success followed by a project failure which left us out of funds with our backers disappointed and our reputation ruined. At all costs we wanted to be sure we could deliver what we promised. This was new technology to us, and a new kind of project. We needed to do it right.

Early experiments were promising, and we had a halfway sane estimate of how much it would take to build the game. We thought - bearing in mind Kickstarter fees, dropped pledges and contingency - we needed about 90K. We'd have felt safer with 100K. (That's pounds sterling. Almost 160K US dollars.)

90K... sounded like a lot. It didn't sound like a sum of money we could raise, and it sounded - it was very hard to judge this - greedy for an indie game of this size and nature. That was really, really subjective. Almost no-one actually knows how much an indie game will cost to make, especially because indie games are so often so different. But we tried the number out on some informed guinea-pigs, and they agreed it sounded high. And funded KS game projects tended to be either cheaper, or much pricier, than that. And in any case, 100K sounded like we'd just picked a big number.

60K sounded like a number we could raise. It was about three times what we'd raised last time, and it didn't (subjective!) sound greedy. So I went back to my design document and I cut out about 1/3 of the game. It hurt, but it hurt a lot less than failing the KS or the delivery would have.

The pre-pre-Kickstarter campaign

In June/July, we built a promotional website with excerpts from the design document and with concept art. We thought it really nailed the concept of the game. We wanted to play it, just looking at the thing. But, that theme again - we wanted to test before lift-off.

So we mailed a few dozen superfans, trusted press and professional contacts - people who we reckoned might want to and be able to give us insightful feedback. (If you're reading this and thinking 'wait why didn't I get an email?' ... we didn't want the mailout to be big enough to be mistaken for a publicity drive, so we went through the initial list and ruthlessly deleted one in two names at random).

We're really glad we did that. There was loads of useful feedback, but two things above all:

(i) our initial pitch, '2D Elite with steamships' was mostly only meaningful to people who were born in the UK before about 1975

(ii) SCREENSHOTS!

The screenshots from the prototype were ugly as hell and we had been nervous about doctoring them, because we didn't want to give the impression of dishonesty. But everyone said they wanted to see screenshots to understand. So we knocked something up based loosely on the prototype, called it a 'screenshot concept', and put it on the site. And that was all we needed! Someone said later 'As soon as I saw the ship in the middle of the screen, I understood what was going on.'

Not coincidentally, almost every 'Journalist's Guide to Indie Devs About How Not To Fuck Up PR' says 'screenshots! Or video would be even better'.

The community

I've barely mentioned this so far, but our existing fan-base (THANK YOU, GUYS) was our single biggest advantage. We had agonised over why one of our Kickstarters succeeded hugely and the other failed - Chris wrote a very candid post-mortem here - but we still think one of the biggest reasons was, as Chris said, 'Below isn’t a Fallen London spin-off, and most of our players are Fallen London fans rather than Failbetter fans.' Sunless Sea was part of the Fallen London universe, and we needed to get our fans on board[1].

So this is where we launched the promotional website - and this is now firmly in familiar Kickstarter territory, early teasers, announcements, building momentum. We had about a month to go when we announced, and I think that timing was about right. It was soon enough that people didn't lose interest, far enough away that we could incorporate player feedback about the game and the pledge rewards into the campaign, and short enough that we didn't blow too much vital company time on something that might still fail. It gave us a sign-up list of about 3000 addresses of people who wanted to be told more about the game.

I'm really glad we got feedback on the pledge rewards. The second pass worked much better than the first, and it's hard to fix problems with pledge structures once a KS has launched.

Submitting the project

There are two things you should know about submitting a project to Kickstarter. One of them is repeated everywhere. One of them was a surprise to us.

(i) Submit your project WELL IN ADVANCE. You can't know how long KS will take to approve or throw back a project, and you don't know what changes you might need to make, and you don't need to hit the LAUNCH button until you're ready. I think we allowed a week. It was a good thing we did, because...

(ii) ...Kickstarter told us they don't allow rough drafts. We've read other blog posts that say they do, or they used to. We put in a rough cut of the video and some 'TBD' copy, and they said clearly, no, they want to review the final version, and could we pass that on to any other creators? So here we are, passing it on.

Looking ahead

This took us up to the actual campaign. The key lesson here is: we had already been working on the Kickstarter for four months at this point! Not all four of us full-time - we'd also been running Fallen London and two client projects - but I think by the time the KS launched, we had put six to seven person-months of salaried effort into tech development, game design, concept art, copy, pre-launch marketing, project management... if we hadn't got funded, we would have been left deeply in the red.

The day before the Kickstarter launch, I did a basic calculation based on our promotional site email list, on the very fudgy assumption that some of those 3000 people wouldn't pledge, but about the same number of newcomers would pledge. The average Silver Tree pledge had been about £28. So 3000 x £28 = £84,000. It was not a sophisticated analysis. But I'd much rather have a number we know is iffy than no numbers at all. You can always refine guesstimates into landmarks.

So how close was it, and where was it wrong? I'll talk about that in the next post. [SPOILERS: happy ending.]