Part of what makes it possible is factoring out time and just focussing on what the project would look like at 100% funded.
Yup, totally agree and that's what I did.
I should clarify that when I talked about using what a project looks like at 100% funded I am actually talking about how to shape and understand a lot of aspects about a campaign, such as how many backers are going to be needed and how they are distribution throughout the rewards. It is actually a perspective that puts backers first instead of the amount pledged. I should have provided more context in my previous post. It is a concept that is difficult to articulate this close towards bedtime. It is about being able to see where backers will gather in the reward tiers and how a campaign should behave, but much more than that it is a way to see the scale of the project itself from what needs to be achieved to what the costs should be.
When I said "What makes it possible" I mean that I had to find ways to factor out time in the problems because the type of mathematical box models I am playing with have the warning from the statisticians that created them that they are expected to produce invalid results if there is a trend in the data fed into them. It is because the probabilities that keep shifting based on past outcomes are throwing the results off. An example where those models won't work is the annual expansion of rabbit population on an island because the previous outcomes of rabbits being born have a big impact on the next year. Then the trend can start from the rabbits destroying the ecosystem so that food sources collapse. The models are much better suited for situations like casino games where the dice are not going to change during the game. Kickstarter campaigns do have trends from many perspectives, so I have to think of ways to process the observations so the trend doesn't matter/appear from other perspectives. Increases and decreases in project momentum can be viewed as just increased frequency of rolling an imaginary set of dice that decide what tier a backer will pick.
The
graphs for That Dragon Cancer are useful because the point I want to illustrate is more easily viewed in the big campaigns. These big campaigns get more imaginary dice rolls so effect of
The Law of Large Numbers can help behaviours become more visible.
There is that S-shape to the $15 tier in the backers-by-tier-over-time graph. It is the result of less y-axis gains when the campaign slowed down in the middle. I view that as a lot of rolls happening early in the campaign, fewer rolls of the dice in the middle and then more rolls happening during the end. I do not see it as a change in the preferences of the backers out there. What if every roll was exactly artificially space out with the same amount of time in-between? There wouldn't be a trend of the campaign speeding up or slowing down. A simulation can just roll the dice over and over at the same pace until the campaign hits 100% funded. I use 100% funded as the stopping condition. This can work because of what can be seen in the percentage-of-the-total-number-of-backers-by-tier-over-time graph. Look how the $15 tier keeps hovering around 60%. It shows how the imaginary dice are weighted.
Because I focus on the percentage of backers by reward tier, it actually makes it easier to
compare campaigns of potentially very different magnitudes. I can see if a $20,000 project has a healthier distribution of backers than a $100,000 campaign. If I just compared pledged dollar amounts, the performance of the $100,000 campaign could have the illusion of looking like a better performer. I'm looking for how well the rewards work at up-selling backers.
This approach I've taken may be dead-end. It doesn't really help to predict how much in total a campaign will raise (The burning question many project creators want answered) because its purpose is not to project for how much overall funding, but to point out what that funding would look like. To some it may look fancy, but really it is just watching the percentages at the reward tiers. I don't know if it is a dead-end yet, so I've kept exploring and refining it. So far it is working to help better understand the popular campaigns where press is not a problem. I'm not done all the work needed to see if it is wrong yet. Using a
Markov chains like some others are betting on may be an better route.
It could be important not to get trapped in the train-of-thought that just because other projects had 10% to 20% go into their rewards fulfilment, it means
Super Toaster Guy should also plan to spend 10% to 20%. A purely digital strategy can work for small projects if the creator is willing to give up some creative freedom by focussing a lot on backer decided content like minibosses and if time is valued less than money. The approach you were using is very useful in situations like building construction if the firms are just building standard homes. Where it can fall apart is not realizing that a project is more different than it appears. Labour is often one of the big or the biggest slice of the pie chart. The cost of material can be a larger slice than labour when a construction firm is building a house with gold-plated fixtures. There are some post-mortems about video game projects actually spent, but they really should be more frequent. It boils down to finding campaigns to compare your rewards to that have a similar scope in physical rewards instead of comparing to a campaign that had art books and working arcade machines as physical rewards.
Another way to fall into a trap is to try to emulate the incredibly successful campaigns from famous developers that had way more resources and accomplishments to leverage. Small projects can fail to compete when the advertising budget of a large campaign is larger than the small projects minimum goal. It is easy to burn out a team producing video updates up to par with what some of the biggest campaigns could produce. As well, just because a campaign raises a lot of money doesn't mean its execution was well done. It is still useful to look at campaigns like
Shovel Knight, but consider it an outlier. It is the medium sized campaigns that can be a better source of examples for how to run a project.
I generally have been recommending the same process over and over when designing reward structures. Why? Because I find it to be more adaptable. Decisions can be made to fit needs more than just deciding based on gut feelings. Sometimes it creates conclusions that a reward should be priced much higher than it was originally going to be offered for.
First, brainstorming all the content that will go into reward tiers like the instruction manual PDF and soundtrack. It is good to prioritize content that isn't a physical reward.
Second, estimating costs for that content including time costs.
Third, arranging the content based on costs from low to high and also sometimes factoring in the prestige of a piece of content.
Fourth, combining that content into reward tiers that don't have a price yet. This can be situations like beta access and the soundtrack being introduced in the same tier.
Fifth, assigning prices to those tiers that provide a healthy margin for the development budget to receive while still flowing well from one tier to the next for up-selling. Content can still be shuffled around.
Finally, selecting from one of the reward structure models such as the most well known being "Includes all of the above". Even before getting into various types of strategies like early-bird rewards, this process can usually produce an effective rewards structure because it is being built from the bottom-up and focusses on costs.
An issue with the above process is that it can be very easy to give away too much content in the lowest priced tiers. That same content, like beta access, could go for a higher price in the medium priced tiers. That is a risk, but I also feel that it is better to have the problem of being less efficient at covering funding the distance per backer but accumulating a lot of backers than to have the problem of not getting people to want to pledge to a campaign because the rewards are priced too high.
I noticed, by examining other campaigns, that there's a huge difference between the 20 and 25 dollars gap (major drop in backers). So I think it would be wiser to remain under the 25$ while making the rewards as interesting as possible. If this turns out to be the most successful tier, I thought it would be better not to include any physical rewards there so that I don't end up with very high shipping fees.
There often is a big difference. I consider a 75% drop from one tier to the next to be acceptable around that range. If the rewards don't flow well, I can actually see an even worse drop from one tier to the next. If that drop can be reduced to around 50%, then a campaign could get funded much easier.
"Postcard with original sketch" sounds like each backer gets a unique drawing on the post card. That could be very time expensive even if a small doodle. It could also be interpreted as a postcard and a production sketch on another piece of paper both mailed together.
Add-on rewards examples are those for
Twin Souls and
Jenny LeClue. They allow for more flexibility, but new backers also need to be educated how they work. The risk of add-ons if what if only one backer adds onto his pledge for a specific item. There wouldn't be much economies of scale. A solution is to ask for a minimum number of backers to contact you saying they would pledge for that item before it is officially added to the add-on list. An example is saying that at least 20 people need to buy the shirt for it to be worth doing. When 20 messages are received, it gets added.
The pricing of the rewards and the minimum goal amount are often still changing all the way up to a few days before launching. The rewards can also have very significant impact on how large the goal is.
Distribution platforms want their cut, but some Kickstarter campaigns were actually able to get a batch of free keys that covered more than 50% of the copies for backers specifically because they games were crowdfunded. This negotiation should happen before the campaign even launches. Apple
gave 3,000 keys for free to the developer of Zombies! Run after negotiating. I could talk about what the experience was like with one of the big console companies, but I know I probably shouldn't go into that. The company would wave some costs for the game because it was being crowdfunded, but not others.
To really reduce the goal there is even the cheap route for a game to distribute DRM-free copies of the game through an online file locker instead of relying on someone else's store that would take a cut. That option can work, but it can also be less convenient for backers and would be too unprofessional for a medium sized project.
Yes. Steam does take a 30% cut of the revenue which is standard for the industry. You won't find better (other than itch.io which will generate you little to no income) and you will certainly find worse. To simplify things, Steam will make up approximately 90+ % of the revenue you generate. Steam earns their cut and "indies" should count their lucky stars that Steam has adjusted its business model to the more open store-front approach. You may find some luck with Humble & Gog, but the percentages of sales drip dramatically after Steam so you must consider the efforts involved of getting your game onto every platform and managing the expenses and taxes from each distribution channel.
Some good news is that if a developer requests a reasonable number of Steam keys for a reasonable reason, Steam has been known to provide the developer those keys for free. Why? Steam is not really known for nickel-and-diming its customers or developers. It also benefits from Steam gaining more users and good games. Humble Bundle had to implement a $1 minimum not because Steam was asking for more money from them, but because Steam disliked how Humble Bundles were being exploited to cheat more entries into
one of its raffles. In the FAQ is says "
If my game is accepted through Steam Greenlight, can I give my previous customers keys for the Steam version? Once your game is accepted for distribution on Steam, we will give you as many keys for your game as you want at no cost". I haven't heard of Steam not giving a developer more than enough free keys to cover the Kickstarter backers. The catch is that developers are not given the power to generate Steam keys on their own. When a developer asks for 5,000 keys, someone at Valve has to press the equivalent of an approve button.
I should begin by saying that I contacted Lobstersundew via PM after reading one of his message in the
Moonman thread. I was well impressed by in-depth, informative message about Kickstarter. I had been roaming about the forum and the internet to this type of analysis backed up with
data (as opposed to impressions given at a whim).
When I discuss a graph I post it on
Imgur. I was previously studying to be a biologist until the college I went to had to shutdown the program I was in. It is because of that I have a need to actually measure what is happening to find insights instead of mainly anecdotal evidence with survivorship-bias like many other project creators have to rely on. I still have descendants of the carnivorous
sundews (The inspiration for my username) I experimented on in my basement under a grow light. Without data, the human mind can sometimes see patterns where there is just chance variation. With data, associations can emerge and reveal patterns. It is also important to provide raw data so that even if a wrong conclusion is made, others can see the data and find the correct answer.
I began posting in
Steel Assault's devblog thread because I wanted to warn them about the risk of launching on that old date they had picked. I didn't want bad timing to ruin its early momentum. It is still not easy for some developers to know when it is a horrible time to run a campaign. Recently some 30 day length video game campaigns launched that are going to run into GDC, IGF,
MIX and PAX all within the window of March 2nd to 8th when these campaigns will be experiencing their final countdowns. Getting press and shout-outs from well known developers is going to be more difficult when many will be in San Fransisco. If a project creator isn't being featured at those events then it will be hard to compete against all the other indie games that will get featured.
While I've occasionally clicked on
Moonman's thread I haven't posted there yet.
I normally just lurk on TIGSource even before I made an account here. I could quickly become a hated
Simon Cowell like figure for crushing people's Kickstarter dreams telling them why they are very unlikely to succeed and backing it up with graphs and examples. To avoid/postpone that I prefer to wait until asked to do a critique.
I was reminded of the life of the
.
The CPU in the movie
Small Soldiers is like an
Arduino boards gone bad horror scenario.
For the dough a parody of
in
I Have No Mouth And I Must Scream could funny.
I watched
Cybersix back when it was on
Teletoon. That mud character Terra had two competing good and evil personalities. A possible character arc could have a fragment be separated from the main dough antagonist and develop its own independent personality that would then battle with the original when merged back into the dough. It could be like the baby metroid ending.
The Bread Pub Brawlers is another bread-themed game.