Dev insights - Creator Partner Program
Creators earn $10k, how the program was created, and what's next
Introduction
Platforms such as TikTok, YouTube, and Twitch offer individuals an enticing opportunity – monetising their passions. But what makes this dream so powerful is the uncapped potential it offers. Even if you reach your dream job at your favourite company, the scope of your impact will be limited to your position within the company as would your compensation. Content creation platforms come with no such constraints.
Making decisions that increase the surface area of your luck, in other words, embracing higher variance within your life is an extremely powerful tool. People are subconsciously drawn to this and just like the allure of content creation, it’s responsible for the success of casinos, the lottery, and trading stocks/crypto. It is why so many people are enamoured with owning a business when data suggests it’s usually more work for less pay.
As a prediction markets site, Manifold would love to tap into this and enable people to make lots of money by forecasting and asking the right questions. Three-months ago during a Hackathon, I wondered how we could offer this when real-money trading was off the table due to regulatory challenges. Well, if we couldn’t pay the traders, what about the creators? Thus, the creator partner program was born. We now approach the end of the first quarter and I want to provide some insight into my thoughts when designing it, what we’ve learned so far, and what we plan to change.
And yes, since its inception we’ve announced prize point markets which, if things go as we hope, could lead to real cash prizes for traders. But even with this, we are still excited to see how we can reward creators for their work!
What is the partner program?
In essence, Manifold pays market creators real money based on how much value they bring to our site. In its current state that means $0.10 per unique trader on each market (in addition to the 10 mana all creators get), $1.00 for each new signup, and $10 if that signup comes back to trade on 5 days. A market must also have at least 20 traders to count with hopes that this disincentivises creators from spamming lots of low-quality markets to get just a few traders on each.
To become a partner, one must surpass the minimum requirements of 1250 total traders and at least 10 markets in the past month with more than 20 traders. We made exceptions for people we thought could bring a lot of referrals to the site.
The primary, longer-term goal is to lay the initial building blocks for Manifold becoming a place where people can one day make a living creating questions.
A more immediate goal is to leverage our existing creators by incentivising more thoughtfulness when creating markets and rewarding those who can bring new users to our site. We want our creators to be excited to create markets and feel able to put a lot of effort into managing complex resolution criteria. This includes those striving to earn their place in the program in addition to those appointed partners.
The second goal is to target individuals with high social reach who would be great to have on Manifold but don’t care enough to use it when they could instead be engaging with their followers through other mediums that offer financial upside. Someone who spends 5 hours building their political Substack could sacrifice 1 of those hours to build an additional stream of relatively passive income by creating markets on Manifold.
Designing the partner program
Scrapping the first version
The initial proposal was an entirely different program from what it is now. Originally I wanted to increase the cost to create a market by five times for partners and not pay them any of the usual mana bonuses. Then, only pay them USD if they hit 30 traders on their markets (for reference even our top creators only have around 30% of markets they create get more than 30 traders).
My thoughts behind this were as follows:
Our power users are mana-rich, so this increased cost won’t be difficult to swallow. In return, we get higher liquidity markets.
The increased cost paired with the 30 trader threshold will ensure partners aim to create very high-quality market. Other sites such as YouTube will punish their creators if they upload unpopular videos by showing future videos to fewer people. We don’t have a complex algorithm that punishes creators for uploading bad content, so this could be an alternative way to achieve the same outcome.
This original model would have increased mana’s value significantly as it becomes a more scarce resource to creators with a way to for it to facilitate real money earnings as you create more markets. This would have had knock-on effects on the whole site, not just creators, and could have led to more mana sales.
However, as expected with a 1-day old idea with minimal outside feedback, this design had some major oversights.
The first oversight was assuming a significant overlap between being mana-rich, being a power user, and being a big market creator. In reality, some of our biggest market creators are the most mana-poor users on our site as their mana is constantly tied up in liquidity in all the markets they create. Up to that point, I had subconsciously organised our users into 3 categories: Power users, casual users and new users. While this is a valid form of categorisation, it isn’t a helpful one for designing the creator partner program. The 3 distinct categories of users on our site I should have been considering were: creators, traders and viewers. While there is some overlap between these, oftentimes big creators are really bad traders. And while the power user traders are very rich, the same is not true of power user creators.
Thus, when I announced the initial partner program and asked for applicants, it became a game of calculating whether it would be “worth it” to become a partner, rather than all the people I hoped to be partners excitedly registering their interest. Many felt they would run out of mana too quickly and so it wouldn’t make sense. Thus the appeal of the partner program was limited to the few who were both great traders and prolific creators who had lots of mana that they were happy to funnel into their markets in exchange for real money. This is probably less than 10 people.
With this feedback, I decided to simplify the program to give out rewards to qualifying users who made good markets and removed the risk-cost aspect to leave us with what we have today.
Incentives
Another key challenge when designing the program was ensuring the rewards were aligned with what creates value for Manifold. This was challenging as our unique trader mana bonuses did a pretty bad job at this. But I wanted to keep as much parity between the mana and real money bonus structures as possible to avoid over-complication.
While it allowed us to get the program started, mirroring unique trader mana bonuses was not sustainable. Not because of the cost, but because it failed to reward what mattered. Here’s an example of a pattern we saw between two different creators and how they are rewarded:
Creator A: Creates an interesting long-term question which is constantly engaging users and bringing people back to the site (eg. the LK-99 market). Let’s say 100 traders return each day to trade. This market requires constant updating, responding to questions about how new events impact the criteria, and possibly a lot of work to resolve (especially true of independent multiple-choice questions). After 3 months the market ends with around 1000 total traders providing the creator with $100 in partner rewards.
Creator B: Creates the same question each day which resolves daily. The question always has the same criteria so it takes them a few seconds to duplicate the market and change the date. The same 20 traders come back each day to trade. After 3 months of this the cumulative trader count across markets is 1800 providing $180, even though the total unique traders is probably somewhere around 20-50 people.
Both are creating value for Manifold and it’s hard to quantify the value each creator is generating, but in this example A is definitely generating more. Yet, with our current bonuses, B would be getting paid a lot more after a few months of constantly getting daily unique bonuses which A would not receive. Ideally, we want to find a way to bring B’s earnings more in line with A’s and probably slightly lower.
In our first quarter, Manifold is spending around $10,000 split between 30 partners. As with most things, this has a top-heavy distribution with our top partner earning 1/5th of this.
Did we achieve our goals?
The partner program did have a big impact on how some users engage with the site. Strutheo came out guns blazing making prop bets markets left and right funnelling every morsel of spare mana into a new market.
We saw other users such as Lion, Mint, and Ziddletwix, who were initially less known than other partners, rise to the occasion by creating more thoughtful markets and engaging in Discord.
While the creator A vs B scenario described earlier did happen to an extent, we saw improvements (it had already been an ongoing problem with mana bonuses). The 20 trader threshold per market before getting partner rewards was a high enough bar that some creators were forced to change their creation habits to include less mass-produced content. Creators were more willing to talk with me understanding that this was a partnership of sorts and collaborate on new ideas to try out. Overall I would say the program had a positive impact towards our first goal - enabling existing creators to invest more energy in creating meaningful markets.
I suspect this could have been marginally improved if we did a better job communicating what we were looking for in our partner's markets. There was initially confusion from partners who asked, “What do you want from me? Why are you paying me for just doing more of what I was already doing?”. But we didn’t want anything new from you, except to do more of what you were already doing and maintaining a high standard. Perhaps I was too lenient in my initial selection of partners and invited some users who didn’t meet this high standard. This then led to more confusion about what we were trying to reward and were looking for.
The second goal, incentivising high-reach individuals, has yet to be realised. While disappointing, in hindsight this should have been anticipated. The quality of people we are hoping to attract gain little upside from being a pioneer in this program. We first need to provide proof of concept and point to how much our top partners earn before they bite. If this is you and you’re reading this and getting interested, contact me!
Upcoming changes
With our old mana trading bonus (10 mana per unique trader) being replaced by a share of trading fees in our pivot, it's time for the partner program to adapt.
Now that the share of trading fees is paid out in prize points, there is a pathway for every creator on Manifold to build a real source of income (assuming the sweepstakes model goes to plan). But if this fulfils one of the primary goals of the partner program, then what purpose does it leave the partner program to serve?
In a world where Manifold thrives and trading volume increases, creators of the largest markets should expect to receive a share of fees equivalent to tens or hundreds of thousands of dollars. Polymarket, which has fewer users than Manifold, has $133 million in trading volume on their presidential election winner market. With our current fee structure, this would result in over a million dollars in trading fees, half of which the creator would receive. In such a scenario, owning large prize point markets where the majority of trading volume occurs becomes crucial to be a profitable creator. Being a partner will be a strong signal of quality and trustworthiness that will attract traders and help position you to capture more of this trading volume and prize point markets. Once we are at this point, any money earned directly as partner bonuses will be negligible compared to the share of trading fees, and we will re-evaluate whether they are even needed. However, we aren’t there yet, and in the interim partner bonuses will be critical to ensuring partners can on average expect a positive financial return when fronting liquidity for their markets.
The old mana bonuses were substantial. The top contenders to win the master league often relied more on their creator bonuses than their profit. With those bonuses gone and the cost to create a market increased relative to users' current balances, it is challenging to expect a positive mana return when creating a market. Two years ago, we might have thought this was acceptable; the creator is asking for information, which is being provided for them, so it is reasonable for them to pay for it. However, creating a question with robust resolution criteria and managing it takes considerable effort and time. Most of our creators don’t directly benefit from knowing the probability to their question, rather they are providing a public service by figuring out which questions to ask that are impactful for society as a whole. This is hard to do and so a creator’s value in the ecosystem is clearly more than just being a liquidity provider.
That's why, while trading volume is still too low for fees to cover the cost of markets, we intend to use the partner program to help creators cover the cost of creating a market and provide additional compensation. We want to continue paying out a similar amount to what we previously did—around $3500 per month split among our partners. This will be paid on a monthly cadence, rather than quarterly, which will give faster feedback to partners and allow them to make markets more quickly with what they’ve earned. But, with unique trader bonuses gone, we need to find something new to tie it to, this time something that actually rewards what Manifold values most.
When we tell partners that to scale the program we need to ensure what they create provides value to Manifold, to which they respond, “Well what type of market does provide Manifold value?”. Rather than using one individual metric, we shall be using a formula that takes into account all of the following things.
Number of markets we can make prize point eligible:
Short-term timeline (right now this means resolving in < 1 month but we will soon be expanding this to markets < 1 year).
Objective criteria
Appeal to a wide audience
Usable by Manifold in marketing and outreach
Quality of market:
The question’s probability reflects something of real-world importance and impact
Unique and thought-provoking
Has either unambiguous resolution criteria or detailed explanations of how various situations might influence the eventual resolution
Having the potential to scale to high liquidity and traders, the structure of the market can be pretty important here
Markets with too short a timeline don’t have enough time to build up liquidity and traders
Independent multiple-choice markets with too many options are valuable because they are fun, but struggle to have enough liquidity and traders for each contract.
Has good generic stats such as trading volume, comments, number of traders, and views.
Markets that are well-managed:
Respond to comments quickly
Resolve and close markets timely
Act as a good mediator and communicate well during disputes about market interpretation
And, to a lesser extent, the number of markets created and the amount spent on liquidity
FAQ about the new structure
Why a fixed amount of partner bonuses?
So partners and Manifold know what to expect during a time when market metrics are hard to predict. We don’t want to come up with a formula tied to creation metrics which ends up paying creators too little because they find themselves unable to create as many questions as we expected due to the change in pricing. On the other side, as a business, we also don’t want to find ourselves owing an unsustainable amount during an uncertain period.
Why $3,500 a month?
For us to spend more we need to see the site grow and scale. But really, as the site scales the trading fees are where most of the compensation for creators will lie. We chose $3,500 as it is comparable to how much we were previously paying partners and it felt like a healthy amount.
Doesn’t the partners sharing this pool cause bad direct competition?
Yes, it does cause more competition. Being a market creator is always competitive when there are a limited number of traders with limited balances. That said, directly competing over a fixed partner bonus is admittedly not the best system, so after your feedback, I am considering a structure that includes “stretch goals” that when collectively met by partners would increase the fixed amount. This will hopefully emphasize collaboration to unlock a higher amount to be shared among partners rather than competition.
Will you be adding more partners? Will this affect the fixed amount?
Yes, we will add some more partners. No, it will not affect the fixed amount. I already have a good idea of how many new partners I will add next month and it should be comparable to the number of people who stop being partners/creating markets. 1-month is a short enough time frame that I plan to add new partners at the start of each month to keep things simple and predictable.
Will there be any other benefits added?
Maybe? The biggest perk we see to being a partner long-term is the status and increased trading volume it provides. We might want to introduce other benefits which help increase market exposure and trading volume.
Thanks for reading! When the partner program’s first quarter ends in a few days, the new changes will go into effect immediately with 1 month-period starting. I am still listening to feedback and may make some last-minute changes if I hear good ideas. To follow along with these updates make sure you check the partner channel in Discord.
David Chee