It’s been just over 4 months since the end of Sub Terra’s Kickstarter campaign, which propelled Inside the Box from a one-person team to a thriving creative studio – and it’s now high time we went back over our numbers with the beauty of hindsight to learn from that experience. What went well? What went badly? And of course, why?

For a bit of context, I’m a massive data nerd. When we designed our first game, Molecular, back in 2015, it was to avoid the mind numbing ‘coder’s block’ I was getting writing my thesis on a piece of nerdy software that did fancy things with data (If you’re really interested, to optimise the spatial organisation of UNHCR-run forced migrant settlements). Since I’ve been a part of the tabletop industry, I’ve been fascinated by all the interesting data that huge, international projects like Kickstarter campaigns produce, as well as all the extra insights that we can draw from a huge worldwide community willing to help us make the best possible games.

Additionally, as many of you know, I’m a massive advocate of being open and transparent with the way we do business, and in helping others independents in the industry grow. I hope that the rest of this unreasonably in-depth statistical waffle provides some insight into how we approached the Sub Terra campaign, and the ways in which you can make your Kickstarter campaign more effective at communicating your project. We’d love to see more small publishers of tabletop games in the UK, and across the world enjoying more success with their backers!

Hold onto your optimised variables, we’re going in!

Project Timing

Season

There’s been a lot of discussion on a range of advice groups and blogs over the last couple of years about the optimal time of day, day of week, point in the calendar month and point during the year to launch, and run, your tabletop kickstarter project. Working from the top down, I honestly don’t think that it makes much difference, given the piecemeal information I’ve found so far – although there is some commentary to suggest it’s harder to do well over the summer (where average pledge per backer decreases), though this may be due to the increase in first time project creators starting up hobby projects over the summer months.

Date

For Sub Terra we worked out when we’d need to launch so that we would definitely be able to have everything ready for the campaign, but would also have plenty of time to get to Essen Spiel with copies of the game and have everyone’s games out well before that. That came out at the start of January. We decided to launch on the 10th, as that gave a decent buffer after the madness of the winter season/new year period.

Day of Week

The specific reason we launched on a Tuesday? Most people that work Monday-Friday jobs are probably catching up after the weekend on stuff in their inbox or things that need to get a good start on at the beginning of the week, rather than scrolling through their phone or procrastinating on Kickstarter – so Monday seemed like a bad day. Fri/Sat/Sun, the gamers will probably be too busy with family/friends playing their favourite games or doing other fun stuff, so probably don’t have much time to be on Kickstarter looking for the next cool thing. For that reason we chose Tuesday, close enough to the start of the week to mean the first few days will be strong (no weekend to worry about straight away), but also not on ‘oh no my inbox is full’ Monday.

Time

Now for specific time of day. We’re based in London, UK, but the largest backer location for us by far is the US. This meant we had to balance our immediate core supporter base in the UK, as well as all those potential backers in the US in terms of when we tried to get everyone’s attention to say ‘Hey we’ve got a thing!’. We decided that 8am our time was a good bet, as that equated to between midnight and 3am across the USA, meaning our first day would register on Kickstarter (New York), Kicktraq (Mid-West USA) and our (UK) analysis platforms as the best possible crossover without meaning getting up at a silly time or staying up too late. This meant we could get the best possible first day on the most viewed stat-site, Kicktraq, without making the back end of the Kickstarter dashboard for us look weird or making it seem like we had a bad first day when we didn’t.

Funding over Time

So, now we’ve launched the campaign – what more can we learn from timing? The first thing I did was plugged the Kickstarter data into a Google Sheet (online version of excel, essentially), and made a chart showing cumulative pledges over the time (using the precise backing times registered on Kickstarter, to the minute).

As is the case with the majority of successful Kickstarters, the first bit of the campaign rapidly grows, as expected, as lots of people jump on the newness of the campaign, and after about 48 hours slows down significantly. In the last 48 hours, as expected again there was a significant rise, as excitement builds over the stretch goal content and Kickstarter sends out a 48-hour warning to all those who ‘starred’ the campaign but have not yet backed the project. The mid-section of a campaign, is usually what creators lament about, that necessary but tense period where the excitement and engagement with a project slumps, along with backers and pledges. You hope that things keep going, but you know they might not.

Many creators, us included, find it really hard to work out what to do during this period. It’s entirely necessary for a new publisher like us to maintain a long-ish campaign so that people have the chance to come across us organically through social media or good old word of mouth – but equally we don’t want to use up all the initial energy generated in the first few days. Basically, we need to work out how to either spread that excitement out, or create additional excitement during this dryer period to engage existing backers, and keep the conversation flowing.

If you look super close at the curve, you can see a few small bumps upwards, virtually all of these directly correlate to the launch of new competitions, new add-ons other other exciting bits we introduced, we’ll come back to that later, though. For now, let’s have a look at the time of day people joined the campaign.

Pledges By Time

Why’s is time important? Knowing when new people have joined the campaign is crucial to understanding when your backers are most active. If most of your backers joined at, say, 6pm, the chances are when they come back to read an update or scroll through the comments that it’s also near enough 6pm. Obviously the issue of timezones hits us hard here. When you’re running an international project like a Kickstarter campaign say goodbye to sleep! As said above most of our customers are not in the UK, and are significantly further afield in East Asia, North America etc. This means that although we can understand when our backers are most active (or at least get a good approximation for that), this might be at times outside of our normal office hours, which puts a strain on the team. That said, understanding this in advance would have significantly helped us to plan our days more effectively so we didn’t ware ourselves out in the early days of the campaign.

So here’s a pretty graph

What I did here was get all the Kickstarter back data (which you get from the dashboard as a creator, split into pledge levels), aggregated them all together on one spreadsheet, sorted the data by time of day pledged (in Greenwich Mean Time), then grouped them into 30 minute chunks. It spat out this delightfully wiggly graph, which gives us a sense of when people are joining the project, in British time. This should help us work out when the best time is to make an announcement, be hot on the comments section, or just generally be available.

As you can see, the biggest group is coming online to  pledge between 16:00 and 23:30 – i.e. not ideal timing for keeping a healthy work-life balance. This is likely because of the huge contingent of North American backers, who would be coming online at a reasonable time for them! This is fairly intuitive, though not something we prepared for particularly well – for future campaigns I’m thinking that we start late finish late, to match the patterns of backers more closely, rather than having a slow morning and having to stay at the desk for 10 more hours than expected!

So now let’s dive in – is there any difference in pledge distribution (i.e. the times that people join the campaign – a proxy for activity) throughout the week? Answers below in wiggle form.

Pledges By Day of Week

 

Kind of unsurprisingly, the weekend distribution is lower (even taking into account number of days in weekend versus weekdays), and with a shifted pattern to take into account people’s tendency to stay up past their bedtime to play Mariokart and a cheeky bit of one-player Scythe while they scroll through Kickstarter.

The difficulty with interpreting the data above however, is that it’s a global picture, not a localised one giving us a sense of how different regions are interacting with the campaign. What would be useful would be to have a look at two different regions that have a lot of backers, and see what we can deduce from the wiggles.

Cue ‘one I prepared earlier’…

Pledges by Day of Week (Grouped) – UK & USA

 

As we had a lot of data points for US backers (over 2,500!) the graphs give us a lot of information on when the average US-based backer joins in the campaign (and we assume, is generally active). The biggest times are late afternoon/evening British time, which makes a lot of sense as this correlates to lunchtime procrastination and back-home-from-work-pre-dinner scrolling. There’s obviously a dip while most people are asleep, and a trail off in the late evening, with a sort-of unsurprising jump up again in the small hours. (That blue light really doesn’t help with sleep, people!)

The same rough pattern can be seen on the UK graphs, though there’s a pretty distinct dinner-time drop in activity (impossible to see that in the US one unless we split down further by timezone).

The weekend graphs however, are completely mental. The US one follows a *roughly* similar pattern to the weekday one, with a later bedtime, but the UK one is an erratic mess. I’m going to put this down to lack of data to produce a sufficient discernable pattern, but oh my, insomnia much!

So what can we learn from this?

From these and the other pledge-timing graphs we can get a sense of when it is backers are active, when they first join the project and when (presumably) they are active after the fact (we can probably make a general assumption that when people back is the same *rough* time as when they’ll come online again, overall). This helps us to plan our work better during the campaign so we don’t lose our minds/sleep, work out when’s best to put out more information, get a vague sense of when to avoid posting crucial info, and more generally the state of Kickstarter backers’ awful sleep cycles. What we don’t know, though, is who these people are or where they are. So let’s dive into to EVEN MORE DATA.

Backer Demographics

Where our meeples are

So where are our backers from? this is a seriously crucial bit of information for the logistics of generally large and heavy items like board games where the cost of shipping is the majority of cost to publisher for each game. It’s also worth having a closer look at where your backers are from, or at least based when they back, to understand your backers better as a community. Are translations something that would be worth investing in? Is there enough of an interest from a particular country to be worth investing time in approaching partner publishers there? Is it worth going to a convention there? When’s best to post information? What cultural peculiarities might you need to be mindful of? Knowing your backers is a really important aspect of being a communicative and successful creator, especially as Kickstarter as a platform is growing faster and faster outside North America.

One of the most striking things we drew from this information was that although the population of the UK is 65.14 Million, and the USA 321.4 Million (about 5 times the size) there were only 2.5x as many US backers as UK. I’m unsure how much this can be attributed to the fact we’re based in the UK (and some of the most effective word-spreading we do is here, through conventions and events), although I suspect it’s significant. After a bit of research it turns out that in the US in 2016, 78% of adults had a laptop computer and 77% of adults had a smartphone, whereas in the UK smartphone takeup in 2016 was 81%, trends in both reports seemed to indicate similar growth patterns, so potentially the slightly higher take up of smartphones in the UK, coupled with the home field advantage made enough of a difference to swing the numbers in this direction. Although this campaign was truly international, it seems that where you’re based still has a huge effect on your business.

So what about pledges? Did people in some countries pledge more? Short answer – yeah some did.

Interestingly, the UK accounted for a smaller share of pledge revenue than the share of backers, whereas the US accounted for a higher share of revenue versus backers. Canada was about the same, as was Germany. This is really interesting, as a first guess might have led us to believe that backers in the UK may be more likely to pledge higher, given that many backers in the UK have met us, played a real version of the game in person with the ITB team, and hopefully as a result have a lot of trust in us. That said, there does seem to be a bigger culture of expecting low/free shipping in the USA than in the UK, so it might be the case that US-based backers are more likely to want to make the most of the shipping rates by bundling lots of extras together. And there are, of course, questions about the relative wealth of people in different countries that we sadly aren’t (yet) equipped to answer!

Now it’s time for some fun maps.

 

So here’s a map of the world, with each country colour coded by number of backers (darker the blue, the more backers), with those that had zero backers in green. Kind of obviously Western Europe and North America were the big hitters, with Australia not doing badly. Not particularly surprisingly either, most of the parts of the world with fewer cultural and liguistic links to the UK/Europe and lower uptake of smart devices had little to no backers. Interestingly, though, take up in South America, East Asia and Eastern Europe had a surprisingly high number of backers all things considered!

Regional Brand Awareness

What might be more useful however, would be something that gives us a sense of the number of backers relative to the population of the country. What I did was divide the number of backers in each country by the total population of that country, then divided that by a million to get a ‘backers per million population’ value. This helps get a sense of ‘market penetration’ in that region, essentially how well known is our product in that place (via the platform Kickstarter). There are however a huge range of factors affecting this, such as internet access, prevalence of use of Kickstarter, and the strength of the tabletop games market in that local area. However, as a first stab it’s not a bad way to identify what the key areas to focus on might be for next campaign!

The lighter the purple, the higher the proportion of backers in that country versus population.

Unsurprisingly, Canada, the US, Australia, and the UK come out high. Interestingly Ireland comes out horrendously, as does a lot of continental Europe, with the notable exceptions of Germany and France. Those in the darkest purple had zero backers. From this we can safely assume that Can/USA/Aus/UK/Germany will stay big hitters, but it’s interesting to see the Scandinavian countries doing well, especially Sweden (which has a reasonably small population of 10 million). Similarly, I would imagine that as our appeal grows in France, Germany, Italy etc, our backers may be more likely to come from Eastern Europe as word spreads organically through local gamer networks and the retail pipeline. Similarly, as organic growth occurs in the Americas, and if a distribution or licensing deal is reached in say Brazil, we could see significant growth there.

What’s also worth considering is East Asia, where there have been a lot of backers as well as significant retail interest (particularly with Sub Terra), which could indicate a rapid rise in interest for future campaigns once we have ITB products in the retail pipeline there too.

Future Regional Opportunity

What this picture doesn’t tell us however, is what the relative opportunity ‘value’ of each of these regions is. It’s hard to tell precisely, as the tabletop industry isn’t known for widely available analytical commentary on the size of markets across the world, other than hand-wavey figures across an entire continent, and Kickstarter is hardly going to hand over their user data to establish user numbers in regions, but what we can do is make a good guess. What could we use as a proxy for opportunity? Size of population is something, but the current market penetration may also be an indicator of potential threshold for that overall chance of a backer popping up there. Similarly, how many people are already in on the campaigns we run?

 

For this one I used the exact same colour scale as the market penetration value, but introduced the ‘opportunity’ idea through a bit of a fudge of an equation (any of you who, like me, studied Engineering will appreciate the value of the ‘fudge factor’). The lighter the purple, the higher opportunity based on current market penetration and probability of further backer growth. What this graph has essentially provided us is the semi-obvious answer – focus on local and established markets. While it’d be great to break into more African and Asian markets, it’s going to be very difficult to do that versus growing the local European and North American markets and letting them diffuse across borders.

Here’s a top 20 on the opportunity value table. Bear in mind this is some well rough maths, and there are clearly some outliers, like Aruba and Gibraltar.

Country Opportunity Quotient
United States 6.503748443
United Kingdom 5.07259841
Canada 3.412664087
Australia 3.032799839
Germany 3.024029085
France 2.290264123
Sweden 2.23743373
Netherlands 1.956777307
Spain 1.880923715
Gibraltar 0.49961883
Bermuda 0.4792105529
Andorra 0.4718029814
Aruba 0.461187393
Iceland 0.4277927908
Belize 0.4246325522
Malta 0.4211912137
Luxembourg 0.4123287091
Macau 0.4097774371
Cyprus 0.4024975102
Estonia 0.3908498856

 

Pledge Levels & Add Ons

So now we’ve had a look at some timey-wimey stuff, here’s a look at how people interacted with pledge levels. In case you’ve forgotten or weren’t on the campaign, the main pledge levels were as below. Apologies to those with exchange rate anxieties – these prices reflected only part of the death of the Pound Sterling back in January ’17.

So these were the main options, there was also a couple of options for custom content and other nice bits, but these encompass the vast majority of backers.

I believe it’s important to understand how backers interact with pledge levels so we as creators can best create value for them, rather than looking at is exclusively as creating value for ourselves. The worst position, in my opinion, you can be in during a successful campaign is where your backers would love more stuff, but you don’t have the stuff to give them. Anticipating this demand is difficult, and expensive at times, but crucial if you want to maximise the success of your campaign while retaining a fantastic value proposition for your backers. Ultimately, backers want to see you as a creator recognising them for their trust and support of your as-yet unrealised project.

When we came up with the various options for Sub Terra, we wanted to ensure that backers could choose from an easily understood, limited selection of options that gave them plenty of entry points, depending on their interest and budget. We made the core game as cheap as humanly possible, with the expansions and extras available either as one big set or as separate add-ons, meaning backers could jump on at any level they liked, without having to sift through endless options.

Here’s the distribution of pledges (No reward, £1 thank you, Caver’s, Veteran’s, Collector’s, Deluxe, Personal, Immersive)

As you can see, the most popular pledge level by well over 1000 backers was the Veteran’s Edition pledge at well over 2700 backers. This option gave backers every bit of game content without any of the extras (Graphic Novel, Miniatures etc), so it was understandable that this was very popular. The surprise to us was that the Collector’s Edition (big box with all the expansions like Veteran’s Edition, but with added graphic novel, development diaries book and UV torch) was relatively unpopular, with the Deluxe Edition (same as Collector’s but with minis range) very popular considering the high price tag (although the value of the bundle was very high).

By presenting backers with a range of options that enabled those with the budget and enthusiasm for the project to invest in us, and giving those with a limited budget or those who preferred to not to go all in a way to get the game without breaking the bank, we created a very large crowd of supporters while *nearly* satisfying the demand of our backers for Sub Terra content.

Below we’ve also broken down the main levels by country – with the interesting revelation that US backers were significantly more likely to get the Veteran’s Edition than the Caver’s Edition versus UK backers, who were roughly equal numbers on both. US backers also formed the vast majority of backers of the Deluxe Edition (the all-in package).

Pledge Level Popularity by Country

Add-Ons

So now we’ve looked at what levels backers pledged at – what about add-ons? Did people like add-ons, or did backers stay away? If they did add-on, what was it for?

As many creators (and backers) know, there is no formal system to administrate add-ons in Kickstarter, which has precipitated the growth of ‘pledge manager’ survey-style systems integrated with payment platforms like PledgeManager and Backerkit to solve this problem for creators who wish to have a range of options available to backers.

From the graph below, you can see that the most popular add-on totals were £9, £16 and £28, which correlates pretty directly to the combination of one or more of the most popular add-ons (miniatures). There’s not a huge amount of insight to be had here, other than the relatively low number of those backers who decided to take advantage of the ads-ons facility. I believe this was due to the fact that the pledge level structure directly disincentivised add-ons, by offering significant value adds of content in each progressive level, with no add-ons not also available in a pledge level.

Now that we roughly know what people added on, what about in relation to pledge levels? Given the structure of the pledge levels and add-ons, there wasn’t a huge incentive to add-on, but below we’ve got a nicely colourful graph on how much people added on sorted by pledge level, as well as total pledge by pledge level. What’s weird is that those who chose the Veteran’s Edition pretty much matched the Collector’s Edition for total spend, by adding add-ons (mostly minis) whereas the Collector’s lot just didn’t bother. Overall the miniatures were a more popular addition to the core gameplay experience than the graphic novel and development diaries, though a significant number wanted both.

 

Updates

After spending about 5 hours scraping data from updates during the run of the campaign, I’ve come up with some interesting insights on how to make sure updates are as effective as possible in keeping backers informed, and importantly, engaged.

First off I got hold of the exact times that each update was sent out (thanks Jen Freeman for helping me on this one!), then coupled that with the pledge data to create a graph of the average increase in pledge total and backer number in the 12 hours after each update time – to see if there was a correlation between when I posted an update (regardless of content) and the amount it helped the campaign grow. Problem with this was the first and last day skewed the whole thing out of control.

So here’s an adjusted one with all the updates that were outliers stripped out, which shows some really interesting insights.

From the (admittedly limited) data, it looks as though posting updates late afternoon (4-6pm GMT) was a good time to optimise for general excitement (i.e. backer number growth), with late evening a good time to encourage existing backers to increase their pledge (either by upgrading pledge level or opting for add-ons). I would imagine this is because people are more inclined to use social media in the late afternoon, whereas already engaged existing backers are more likely to check in on a campaign later in the day.

The other metric which I thought would be valuable was seeing if there was a correlation between growth in backers or pledge total and the time since the last update – is there a sweet spot? Similarly for this one I stripped out outlier data for a more interesting result. There seems to be a weak correlation between longer intervals between updates and lower increases in backers (though it’s hard to tell from the data, it’s a bit meh). There’s a slightly stronger correlation between average increase in pledge total and interval length, implying that by keeping up the ‘momentum’ of a campaign you’re keeping the excitement and engagement level at a level which encourages sharing, backing and general engagement with the campaign.

Referrals

So, where did all these wonderful backers come from in the vast expanse of the internet?

What particular bit of stuff we said, or someone else said about Sub Terra spurred someone into action and gave them enough confidence to back our project? Was it something on Facebook? Twitter? A blog? A review?

Here’s a trio of pie charts, that shows the majority of referrals came from Facebook, with a significant amount coming from external sources, with a similarly sized contingent being custom external sources – i.e. customised tracked links from things like paid adverts that we set up or links we set up with partner organisations, rather than someone else posting the link independently of us. We got slightly more pledge revenue from Kickstarter search pledges (people who just came across it) and from custom pledges rather than external referrers. This might be because people browsing Kickstarter are people that back a lot on KS already and are more trusting of creators, and similarly those places where we’ve had a direct communication line with potential backers has engendered more trust, as opposed to other sources.

In terms of those Kickstarter pledges, the vast majority, over 70%, were from searches of some description on Kickstarter, with 20% from earmarked notifications/pages, and a few percent from other KS internal referrals. What’s interesting here is that this suggests that a huge chunk of backers are just browsing Kickstarter, rather than coming via a mailing list, Facebook page or other platform. This could well be because people have a look then come back later via search, but even so this is pretty incredible that seemingly so many people are constantly browsing the tabletop games section for new, interesting things!

 

Let’s deep dive into that Kickstarter referrals pile, given that’s the biggest chunk.

MORE PIE!!

 

Perhaps surprisingly the biggest chunk is from people literally just scrolling through and going ‘ooh that looks cool’, which really serves to communicate the value of a compelling thumbnail image and effective communication on the tagline/title. A decent amount as well was Kickstarter’s internal recommendations and notifications system, nudging people towards projects that they think they might like.

The various ways to search Kickstarter projects, from ‘Projects we Love’ to searching by location of creators, Kickstarter is a really versatile platform which allows backers to connect with projects they might love easily.

Here’s one that shows the total amount pledged from backers who were referred by some platform/link where we paid for some kind of advert or contributed a prize etc. Some of these are sponsored posts on Facebook, others are giveaways and newsletter features. As you can see – Kicktraq was incredibly effective as a referrer, though this did cost a reasonable amount of dollah. BGG looks as though it was incredibly effective in sheer numbers, but overall was quite bad on bang for buck, unlike Kicktraq, or Facebook (by quite a wide margin). That said, it’s worth doing still if you have enough in the budget, just to ensure people don’t miss out on your campaign because they tend to use one platform rather than another.

 

Towards the end of our campaign we were contacted by a marketing agency called Jellop, who specialise in assisting crowdfunding projects reach a wider audience. At first we were pretty skeptical, as we’d be approached by many dodgy companies with similarly grand promises. We decided, however to do a trial run, to see if they could deliver – which to our surprise, they did. After some negotiation with their team and consultation with our backers we came to a deal which gave them a slice of revenue from pledges they brought our way, which in the end made up a significant portion of the final tally. Although we didn’t make much from those extra backers, it swelled the ranks of people who are able to enjoy Sub Terra, earned Tim, the designer, a bit more on royalties, and helped to make the end of the campaign VERY noisy.

 

Here’s a closer look at those external referrers (not including the Jellop referrals or those with no referral information), with the staggering amounts of referrals from Facebook, Kicktraq, Reddit and BGG with progressively smaller and smaller amounts for smaller blogs or less directly related sites. It might seems like it’s not worth bothering engaging with the smaller sites or blogs, when compared with bang-for-buck of Facebook etc, but the reality is that without the noise generated by the super enthusiastic users of those smaller sites, across the world, the engagement on platforms like Facebook would likely be significantly lower. Additionally, those little bars on the chart add up, especially when you saturate the amount you can use a platform like Facebook.

Below is another tasty pie to show what the proportion of those pledges looks like, with a breakdown of the major platforms we use.

 

Conclusion

So after all of that, what have we learned to help us, and hopefully others run more effective campaigns?

  1. Know your audience!
  2. Plan your time and work across a campaign
  3. Use your campaign’s resources efficiently
  4. Talk to backers directly, often
  5. Don’t just do digital marketing
  6. Provide a range of clear pledge options
  7. Spread the word of your campaign wide
  8. Give your backers opportunities to actively engage

Although I had a lot of fun crunching these numbers and making pretty graphs, I don’t have all the time I’d like to delve deeper into the numbers, so if you’d like to – chuck your email below and we’ll automatically email you a copy of the raw data for you to play with. Enjoy!

Have any questions about our Kickstarter campaign? You can get in touch on Twitter, Facebook or email us directly.