Crowdfunding and the paradox of choice
Are too many reward tiers good or bad for your campaign?
I’ve been thinking a lot recently about Kickstarter rewards, primarily because I just wrapped another campaign, but also because I came across this abstract.
Sadly, the paper itself is behind a paywall, and I haven’t yet been able to access it, but I’ll cite the abstract here.
In this study, we investigate the effects of reward options and their prices on crowdfunding success. Rational economics predicts that the more choice potential contributors have, the more likely it is that they find a reward option that stimulates participation. However, experiments in behavioral economics and marketing show that providing someone with excessive choice (overchoice) might adversely affect participation. Using data collected from Kickstarter, a well-known crowdfunding website, we demonstrate the existence of the overchoice phenomenon in the context of crowdfunding, i.e., an inverted U-shaped relationship between reward options and crowdfunding performance.
I’m curious about this paper because it relates to the concept of over-choice, which is when there are too many potential options for an individual to confront. Some of you may also be familiar with the Paradox of Choice, which offers a similar thesis.
In most of my recent articles, I’ve examined how different elements of the campaign page can be associated with the success (or not) of a crowdfunding campaign, but one crucial element that I’ve not explored yet is the reward structure.
As a creator, this is often a tricky issue, as you want to offer a wide range of rewards that entice backers at different levels. But there are also recent trends in comics (for example) where creators offer large numbers of variant covers, which also increases the total number of available rewards to choose from, but can also impact someone’s ability to easily peruse the website, let alone make a decision about which reward tier they wish to choose.
Is there anything we can learn from the literature on this? I hoped the above paper would be helpful, and perhaps it still will be, but the reference to the inverted U-shaped curve makes me a little nervous.
There are a few papers in the literature that have explored the impact of the reward structure, but they have often identified contradictory findings - namely that too many (too few) reward options are associated with decreased success of campaigns.
The abstract of the article appears to suggest both these results are in concordance, by postulating the presence of a ‘centrality bias’ effect, a term that appears frequently in behavioural economics and basically means that things want to fall on a distribution where the maximum occurs in the middle of some range.
The reason I’m cautious about this result, is that there was another paper on reward structures that looked to examine centrality bias. This paper makes two reasonable-sounding hypotheses about project-based crowdfunding campaigns that they explore:
Backers tend to choose the donation option in the middle, independent of the price range.
Backers tend to choose the donation option in the middle, independent of the number of options.
To test the first hypothesis, the authors designed an experiment where participants were given a finite number of coins, which they had to use to back test projects. Participants were not given any information about the project funding goal or other observational learning information, so their decisions were not informed by herding behaviour.
The study featured three reward tiers for each of three products (e.g., a book), and participants were randomized to one of three price scales (cheap, moderate and expensive) for each product. For example, for the book, the reward tiers for each scale were:
Cheap scale: 5 coins; 10 coins; 15 coins
Moderate scale: 10 coins; 15 coins; 20 coins
Expensive scale: 15 coins; 20 coins; 25 coins
For each scale and product, most participants tended to back the middle option, however the authors also used logistic regression, controlling for various factors such as gender and age, to search for inferences.
Their model estimated that across all projects and scales, 21% of backers would select the cheapest reward, 44% would select the middle reward, and 35% would select the most expensive reward. Thus, they confirm their first hypothesis.
To test the second hypothesis, the authors designed a second experiment where there were more reward tiers available per project rather than price scales - but each reward level needed to have the same central reward cost. Thus, for the book example, there were options with three, five, and seven rewards:
Three rewards: 15; 20; 25
Five rewards: 10; 15; 20; 25; 30
Seven rewards: 5; 10; 15; 20; 25; 30; 35; 40
Again, logistic regression was used to model the data, and they determined that the probabilities were highest for the middle reward option.
For example, in the three reward tier case: 34% backers would get the 15 coin reward, 49% would get the 20 coin reward, and 16% would get the 25 coin reward. Again, this confirms their second hypothesis.
For creators, this finding suggests that centrality-bias is real, and that creators could benefit by dropping lower price rewards and thus ensure higher-priced rewards appear in the middle of the range.
But, looking at my own campaign data (and also some other campaigns), I don’t see this effect at all. This is where theoretical modelling breaks down because I think it’s too simplistic.
The above paper, although nicely written and certainly interesting, makes the assumption that all rewards are identical, differing only in price. This is not the case in real life, where value proposition come into play. For example, with shipping costs continuing to spiral, I will only back comic/publishing projects that offer digital tiers - which are typically much cheaper than physical tiers. Some backers prefer digital rewards only. Some prefer physical rewards only. Some prefer physical products if the project ships from the same country, but digital otherwise.
My own data bears this out, with digital tiers being the most popular (and cheapest) of the rewards for my campaigns. Even on campaigns where I offer several different digital tiers, the cheapest one is the most popular choice. For physical books the data is more complicated, but there is no clear centrality-bias.
This is because there are many other variables that impact a reward decision, not just its location relative to other rewards. The type (digital vs physical) and the complexity (number of different rewards options), and shipping costs, all play a role in decision-making.
This is where the theoretical assumptions fail, as they are not modelling a real-world scenario, but one that can more easily be analyzed. This fact is what makes me curious about the original abstract, and their argument for an inverted U-shaped distribution. What are they modelling? What methods are they using? What variables are they considering? What variables are they omitting?
While I’m still a big proponent of using the literature to help us better understand crowdfunding and crowdfunding behaviour, I also recognize that there is often a gap between theory and reality. The literature can help guide creators with some decisions, but when it comes to others, there is no better teacher than experience.
So does the paradox of choice have an impact on backers? I think it does, but I don’t have the data to back that up. Not yet, anyway.
What do you think? Do you see centrality-bias in your own crowdfunding data? Let me know.
From what I've heard from others, and my own anecdotal experience, too many options *can* lead to confusion and reduced interaction with your campaign. But there are ways to alleviate this in terms of Add-Ons vs additional core tiers, making certain the basics are covered in tiers with customization and options now being set up as Add-On additions. My own recent campaign, which failed to fund, I do believe I kept the tiers too simple/few. Not that this turned anyone off from pledging, but there simply weren't enough options to spend more money in ways that backers saw the value in that greater amount. So for my upcoming relaunch, there will be a few more tiers, more items and options. Not because "more is better", but just so the math actually maths. So that ~100 backers can fund me, rather than requiring a ridiculous number of backers just to make it to the funding line, because backers only had two or three moderately priced items to even spend on.