Crowdfunding rewards: re-thinking your prices
Should you set your rewards according to what other creators are charging? Or are there other factors you need to consider?
You may recall a previous post on this topic where I lamented the analysis of the paper as it didn’t differentiate between reward types, which I felt was an important element for backer decision-making.
I went back to the literature and did a deeper search for more information on the analysis of rewards, and was delighted to find this paper Analysis of Rewards on Reward-Based Crowdfunding Platforms, which dug into these issues in much more detail.
This paper explores a small dataset of 3104 Kickstarter projects, collected between December 2013 and March 2014. While this is not a very large sample size compared to some of the papers I’ve previously discussed, their analysis and findings are interesting, and I would hope that this team - or others - are working on a similar paper with a much larger sample size.
As with most crowdfunding papers, it’s important to define the variables of interest for any potential analysis. The relevant variables in this paper are:
Number of rewards
This is simply the number of rewards offered for a specific project, and as discussed before, has a complicated association with the success of a given campaign. The findings in the paper showed a strong association (0.964) between number of rewards and the success of a campaign.
Pricing
The authors hypothesize that backers must not only be interested in the project, but also willing to pay the price of the reward on offer. My own experience concurs with this hypothesis, as there have been many times when a project looked exciting to me, but the reward tier was priced too high to make it worthwhile (this is particularly true for physical goods where the shipping cost is equal to the reward price itself).
However, given the analysis includes all types of Kickstarter projects, reward prices varied between $1 and $10,000, with 75% of all reward prices were between $1 and $200. Given this heterogeneity, the price itself is not a useful variable to consider. Instead, the authors identified two sub-variables of interest: price range and price/goal (p/g) ratio.
Price range: Using their own data, they grouped prices into three tiers: low, middle, and high using quantiles. Thus, rewards between $1-$10 were low price; $10-$200 were middle-priced, and $200+ were high-priced.
p/g ratio: This is an unusual (but smart) measure, as it is simply the price of an individual reward to the project goal. So, for example, if your goal was $1000, and you were offering a digital reward for $5, the p/g ratio would be 0.005. Using the quantiles of the dataset, the ratios are also grouped into low (<0.00143), medium (0.00143-0.03636), and high (>0.03636). In our example above, the digital reward would be in the medium p/g ratio tier.
The data showed p/g ratio was statistically significant. Successful projects had more medium and high p/g ratio rewards than unsuccessful ones, while unsuccessful projects tended to have projects had more low p/g ratio rewards. This finding agrees with intuitive experience. If you have a high goal for your project but are only offering cheap rewards, then you need more backers to hit your target. In our example above, you need 200 backers pledging at $5 in order to be successful, so you have to ask your self as a creator: is that a feasible number of backers for me to find?
Limited rewards
Limited rewards are clearly an incentivizing factor for backers, as discussed in the last post. This was also true in the smaller dataset analysis, where both successful and unsuccessful projects raised more funds if they had limited rewards than if they did not.
Reward taxonomy
The authors also tried to determine the impact of reward type on success rates, and identified 15 different categories in their dataset. So these included things like digital and physical copies of the product, but also “Thank you” tiers, “Early bird,” tiers “Spin-off,” tiers “Collaboration” tiers, etc.
Detailed analysis
The authors performed a multi-factor analysis to identify the impact of the above variables on the success of a project. Their analysis identified the following top five factors:
% of rewards with high p/g ratio
number of projects backed by the creator
has a late-added reward
number of physical copy rewards
total number of rewards
The most important factor was the share of high p/g ratio rewards (although interestingly, the share of medium p/g rewards didn’t even make the top ten). Analysis suggests that the higher the share of high p/g rewards, the more likely you are to be funded.
Recall our earlier example where the goal was $1000. In order to have a high p/g reward, it needs to be at least $37, so let’s assume we have a $50 reward. However, if the only rewards available are the $5 and $50 ones, the share of high p/g rewards is only 50%. While his may be sufficient to be successful, having another reward tier priced over $37 would make success more likely.
This is an incredibly useful finding for creators (with the usual caveat about sample size), as it allows them to consider how to price their reward tiers more thoughtfully. There are basically two ways to increase the share of high p/g rewards. The first is to consider a smaller goal, as this automatically increases the p/g ratio for all your rewards. However, this may not be possible if you need to hit a minimum target in order to produce a product. In that case, it’s preferable to work out the minimum reward price that meets the definition of a high p/g reward, and ensure most of your reward tiers are priced at this level or above.
The other variables in this top five are also important to consider, but I won’t go into all of them in great detail.
The number of projects backed by the creator is ranked second, and more important than other factors associated with the campaign page (for example). There are many reasons for why this is important, some of which I discussed in previous posts, but primarily this is a signal to other backers that the creator is actively engaged in the crowdfunding community, and understands the business model. This in turn may alleviate perceived risk in the minds of potential backers.
Late-added rewards are those that are added to the campaign after it has begun, typically such rewards appear during the project dead-zone in an attempt to excite the potential backer base. The authors suggest late-added rewards can positively impact the overall funding amount of succesful projects, but there are insufficient details to understand how - or why - this is the case.
The final section of their analysis focused on reward taxonomy, where they examined relationships between the 15 different reward classifications number of backers, and success rates. I reproduce their figure below to illustrate their findings.
The figure on the right shows the association with success rate. Projects that included a reward tier categorized as “Including all the previous rewards” had an average success rate of 94%. Projects including a “thank you” reward had a success rate of 60% even though backers don’t receive an actual reward!
What is interesting, is that “spin-off” rewards (such as mugs, t-shirts etc) have a higher success rate than digital, physical, and early-bird rewards. The authors suggest that “spin-off” tiers may be cheaper than other rewards, which is why they are more represented in the data, but I don’t agree. I’m not sure how likely it is that a “spin-off” reward is cheaper than a digital version of the project, at least for comics (which is the category where they are most prevalent, as you can see from the heat-map below).
The heat-map identifies which reward taxonomies are associated with success across different project categories. The darker the colour, the more successful the reward.
The most interesting feature of the heat-map, for me, is that although physical copies are identified for every category - they are not strongly associated with high rates of success! With that said, the dataset examined in this paper is a decade old, so some of these associations are likely different now, but it’s still interesting to explore how different rewards behave in different categories.
What does it mean for creators?
This paper is relatively old and only explored a small data-set of roughly 3000 projects, however their findings are very interesting for creators.
I found the taxonomy classification to be very interesting, although I may quibble about their overall breakdown. It’s useful for creators, because you can use this list as a framework for your own projects, and map your proposed rewards to each of the 15 categories to identify any potential gaps and opportunities. While some rewards may not make sense for certain projects, it’s worth thinking about how you could integrate them into your own campaign.
While there is no guarantee that having more reward types will increase your chance of success, it will likely create new reward tiers that have higher p/g ratio. For example, you could include a higher-priced reward which offers an experience (that you would need to tailor to the specifics of your campaign). I found this variable to be very interesting and plan to think more about how to use it myself in future.
For the moment, I went back to my own Kickstarter project data and broke down the rewards into the different p/g categories using the same cut-offs identified in the paper. I then calculated the share of high p/g rewards for each campaign, and then did a simple linear regression to understand the impact. My dependent variable was total funding raised (in dollars) with the independent variable being the share of high p/g rewards.
My regression found the following: R=0.85; F(1,2) = 5.11; p>.05; b=2154.
Although the power of this model is weak (due to the small sample size) the findings indicate a very strong correlation (but not statistically significant) between share of high p/g rewards and total funding raised, and that an increase in the share of high p/g rewards by 1 implied funding would increase by $2154.
So my data does agree with their findings, that there is a strong association between share of high p/g rewards and total funds raised.
The lesson here is that you need to consider your reward tier pricing carefully in relation to your total goal, and not simply look at what other (peer) creators are charging for similar products.
This may sound obvious, but it’s not that simple. There is a herding effect amongst creators, whereby they coalesce onto the same price point for similar items - irrespective of other factors. For example, you can often see comic campaigns offering physical copies of the book for almost the same price, even though the individual campaign goals are very different.
Just because one creator is offering a fixed price point for a product, it doesn’t mean that’s the market price for all similar products! As a creator, you have no understanding of the business models employed by other creators, who may - for example - be offsetting print costs through other mechanisms.
Social media does not help this situation, as backers (and creators) publicly argue about how much a specific product should cost on crowdfunding platforms, and frequently mock creators whose prices are out of step with “what things should cost.”
But I think this mistakes cost for value. Crowdfunding platforms aren’t online retailers (and we should encourage them not to be). The audiences are different and have different value propositions. This means - for example - comics don’t need to be priced at $3.99 like Marvel and DC books, but can be priced higher because they offer somethign of differing value.
I think it makes more sense for creators to understand their funding goal requirements, and then base their reward prices on that specific goal - and not on what other creators are charging. Using this approach, the likelihood of having a larger share of high p/g rewards increases, and therefore the probability of success also increases.
This is the ultimate take-away for me, and again serves to highlight the fact that reward-based crowdfunding is not merely retail, which comes with its own market-based theories and models, but something quite different - and this means we, as creators, also need to think and behave differently.
What do you think? Does this make sense to you? What other questions arise for you as you read this? Let me know in the comments.