The Good, the Bad, and the Unlikely: EVN Kickstarter Analysis

Report 1: General Data

Data. It’s big these days, right?

Crowdfunding. It’s a tricky business, isn’t it?

Visual novels. You know what those are if you’re reading this, don’t you?

I’m a researcher who’s spent an awful lot of time scrutinizing English visual novel Kickstarters, and this series aims to share what I’ve learned.

Methodology

After watching me painstakingly enter campaigns into a spreadsheet, Perihelion (@apogeedwell) gathered data from almost 500 projects on kickstarter with a ruby script that incorporated the libraries kickscraper, nokogiri, and ocr_space in addition to her own code.(1)

We excluded 37 projects (as of January 2019) considered false positives, such as graphic novels, and 47 projects with locations set to Japan or for Japanese games were separated into their own group. As the text search was based on the English keywords “visual” and “novel”, after these exclusions we consider the remaining 407 projects to be EVN campaigns, dated from 2011 to 2018.

Some EVNs may have slipped through the initial search, but these 407 projects are likely to represent trends within English-language visual novel kickstarter campaigns. Projects in the JP group were primarily localization campaigns for existing VNs, which will be interesting to look at in the future, but don’t represent what your average EVN devs are getting up to.

After excluding 9 projects considered outliers based on their funding, we’re left with 398 campaigns.(2) For ease of discussion, values are rounded.

The big caveat: I may be a data analyst, but I’m an armchair Kickstarter analyst. I have no direct experience running a campaign, so take direct advice with some grains from your salt shaker.

Now with all that covered, for this article we’ll be looking at project outcomes, funding, backer counts, and goals.

Project Outcomes

Number of campaigns each year. Includes all 407 campaigns.

Though they had a slow start, the number of EVN campaigns took off after 2013. The points represent campaigns for that particular year, not a cumulative total, which means the total number of campaigns between 2011 to 2014 (56), is less than the number of campaigns in 2015 (65) alone! Keep this in mind when viewing data for these years, which is less stable than later years.

Bar chart of campaign outcomes.

At this time, about half (52%) of the finished campaigns met their funding goal, 31% did not, and 16% were canceled. In four cases, projects were canceled despite reaching their funding goal.

Changes in campaign success rates over the years.

Looking at success rates by year, remember that the earlier years have fewer projects. The results of just one or two campaigns could dramatically alter the success rate for that year. For example, 2011 had only one success…out of a whopping three projects. After 2014 the success rate hovers around the 50% mark, which is likely to continue for 2019, though it may dip closer to the 40% mark.

Even if it does, that would still be impressive compared to EVN Kickstarter neighbors!

EVN success rates compared to video games, games, and all categories on Kickstarter.

There is of course more to the concept of success than a campaign meeting its funding goal, but by this metric EVNs campaigns are quite successful when compared to relevant categories.(3)

When you look at this in terms of dollars pledged to successful projects, however, EVNs are a glass of water next to an ocean. 2017 was the most recent year I could find complete data on, and the total amount pledged to successful video games that year was $17.25 million. Successful EVNs, on the other hand, earned a total around $530,000. When looking at the total amount that successful EVNs have earned since 2011, that sum becomes a whopping $1.24 million.

EVN campaigns may be more likely to succeed, but they aren’t earning nearly as much in comparison to other categories.

So how much do they typically earn, anyway?

Funding

Histogram of EVN Kickstarter Funding. This and all following graphs exclude outliers.
  • Mean: $6,795
  • Median: $3,414
  • Standard Deviation: $9,395
  • Minimum: $0
  • Maximum: $47,993

A brief rundown of these concepts:

  • Mean: The standard way to discuss the idea of an average. However, very large and very small values can dramatically change its value.
  • Median: Another way to discuss averages. Though it changes less than the mean, it’s more conservative of an estimate. The data is quite skewed, so I use the median to discuss averages despite this.
  • Standard Deviation: This measures the amount of variation between campaigns. It’s gigantic for all of the data, as the campaigns are spread out over a large range of values.
  • Minimum/Maximum: The smallest and largest values.

Despite removing outliers, the distribution of funding (regardless of campaign outcome) remains skewed towards the right due to a handful of projects that receive far more funding than most. As a result the median ($3,000) is a more accurate description for the average amount of funding rather than the mean ($7,000), though both are included for the sake of completeness.

Histogram of funding with certain ranges highlighted.

Even when using the mean to describe the average, 67% of projects earn within the range of 0–$6,000. Campaigns are generally more likely to earn under $6,000 rather than over it, and only 19% of projects earn more than $10,000. Tuck that $10k point into your memory for later discussion!

Another thing to remember, yet again, are the less stable averages for years prior to 2014. The average amount of funding as generally gone between $3,000 and $4,000, and 2018 had a slump similar to 2015. Whether 2019 recovers or dips further remains to be seen, though there appears to be an overall downward trend.

Backers

  • Mean: 206
  • Median: 114
  • Standard Deviation: 262
  • Minimum: 0
  • Maximum: 1721

For backers, the change over time stood out to me more than the overall distribution, but the histogram is included for completion’s sake. One obvious reason for the downward trend in funding is the decrease in backers. Even when using the median as the average, 2018 saw more of a decrease than the 2015 slump.

Goals

  • Mean: $7,522
  • Median: $5,000
  • Standard Deviation: $9,002
  • Minimum: $72
  • Maximum: $62,777

The distribution of goals provides some context for funding amounts. One reason most campaigns don’t earn more than $10k is because only 18% of projects ask above that in the first place! The average goal remains higher than the average amount of funding no matter which value you use, suggesting that projects have a general tendency to ask for more than they’re likely to earn.

You see this tendency when comparing the average goal over time — despite the downward trend in funding, there’s an upward trend in goals. Even so, the average goal for a year has yet to make it over $9,000.

When looking at the goals of successful kickstarters, the cutoff after $10k is so steep that it’s why I highlight that number! Not many projects ask for more than that, but this is especially true of successful projects, where 92% of them have goals under $10,000.

Once a project reaches its goal, over half (58%) don’t earn between an additional 25–50% of their goal, and even fewer (28%) earn more than twice their original goal. Shocking, I know. Most of the absurdly high values stem from projects with unusually small goals.

Summary

Have the golden years of EVN Kickstarters passed, with glory forever out of reach?

Doubtful, though 2018 was a slump. 2015 was a slump as well, but 2016 saw a jump in funding, backers, and success. It’s too early in 2019 to see if we’ll pull back out, but there will likely be a large number of projects, and thus a lot of competition.

Does all this mean you shouldn’t ask for more than $10,000 USD?

No.

Well, probably not.

When setting a campaign goal, remember to always ask for your minimum budget. If you ask for more than this, the campaign will be riskier, but if you ask for less of this, it will be harder to actually finish the project. (Discussion of how many projects release will be a later post.) When budgeting a project, however, these numbers are worth consideration if it seems crowdfunding on Kickstarter is likely.

Overall, the general data can’t tell you much about your specific project. There’s so much variability in the data and there’s only so much the project page can tell you. Kickstarters are like trying to get other people to put lightning in your bottle, but learning from others gives you something of a lightning rod. It’s far more useful to look at projects from creators with similar background and audience sizes, which is beyond the scope of this article. (But check out this book by comics creator Greg Pak for advice from people who have actually run Kickstarter campaigns. Most of my 2 cents have been donated by such people.)

Caveats aside, the general data can help with general discussion. If you have no idea where to start or what to expect, these numbers can help adjust expectations.

Future goals include an analysis of reward tiers as well as commonly offered rewards, looking at projects by genre, repeat campaigns, and critical considerations like how to price your dakimakura.

Until then, I hope you enjoy these hors d’oeuvres of thought!


1: Special thanks to Nai’s (@RemortStudios) bot, Lemonchan (@VN_Devtalk), whose antics were one of the things that inspired me to do this. I’d also like to mention bunnyAdvocate’s (@BunnyAdvocate) Steam and VNDB analyses as something else that got me interested in this. Go check her articles out!

2: Why do we do exclude these outliers? Well…

You may have heard about Spiders Georg, who lives in a cave and eats 10,000 spiders per day. When trying to find out how many spiders the average person eats, Spiders Georg will make your results say some really weird things. If you eat none, I eat 10, and Spiders Georg eats 10,000, then our average spider consumption will be 3,336 spiders.

Not exactly an accurate picture, is it?

Take him out, and we’re down to 5 spiders. That’s a bit easier to swallow.

In our data, we have a small number of campaigns whose success is leaps and bounds beyond the average. The Starfighter visual novel is a prime example, asking for $70,000 and receiving $143,183 from 2,646 backers. For most analysis, we exclude Starfighter Georg and similar projects. The highest-earning outlier was Megatokyo ($299,184 USD), while the lowest was Dysfunctional Systems ($51,262 USD).

Funding Outliers

Crystalline
Dysfunctional Systems
Exogenesis ~Perils of Rebirth~
Full Service
Harem Protagonist
Love Esquire
Megatokyo Visual Novel Game
Shining Song Starnova
Starfighter Visual Novel

3: The discrepancy in success rates between video games and games as a whole is likely due to the large number of successful tabletop gaming Kickstarters. There’s more about this divide in this report by Charlie Hall over at Polygon, while the 2018 numbers for video games (as well as some interesting commentary) are from this post on by Thomas Bidaux, though this only covers the first half of 2018. The rates for the games and all categories come from Kickstarter’s official stats.