everything in our power to create a responsible and honest foundation, to know that we did not take short cuts here. 

someone asked me about this section so logging my response to them here publicly. i'm pretty sure no one reads this blog, so if you do happen to read this please say hi! would love to hear from you. my DMs are open twitter.com/cindywu

> I'm curious if you have any more thoughts or insights on this section.

briefly, i think four years later we stayed true to the "create a responsible and honest foundation" 

i haven't processed this entirely, but i do think that my action (or inaction) may have been the cause of the company being inadequately financed. i do think my decision to finance the company on a shoestring budget led to the failure because we operate in a very competitive market (google docs, notion, etc.) where table stakes are high. 

basically i only took money from founders and ceos, humans who i knew would support our team in charting an alternative structure where we bake in profits serve purpose instead of profits serving profits. though i had access (vcs funded my first company, a16z/index/lux/sv angel/two sigma etc.), i never entertained vcs. never even met with them. 

now, that is for jelly pbc. we never publicly announced this, but we stopped working on jelly pbc last october. i now work full-time at vercel. i'm not entirely sure what will happen to jelly pbc, but for now we've chosen to do nothing.

i think (and i hope) we will see experiment come full circle in the next 3-4 years. my current plan is to take that delaware c-corp that is funded by the typical sv vcs you would expect and transition it into a steward-owned corporation. in many ways openai is a steward-owned corporation. other examples of steward-owned corporations are zeiss and possibly patagonia. i'm also waiting on kickstarter to make a move. 

long story short, i did not take any short cuts but i also didn't get anywhere that feels meaningful lol

We made a conscious decision to stop working on Experiment in February of 2019. As we embark on building something new we felt it worthwhile to reflect on why we stopped building the first company, what we feel we want to repeat, what we feel we want to improve, and what we feel is just different. 

This post was not easy to write, as we are still surrounded by more questions than answers. This is me, Cindy, giving it a try. 




We have many hypotheses as to why we stopped working on Experiment, but there is no method anyone can use to scientifically uncover the 

 reason. What we can offer is a story about how we felt then and how we feel now. One thing we've learned in the past eight years is you can't argue with how someone feels. 

We officially stopped working on Experiment in February of 2019. One of the hardest things we did to close this chapter was sit in a closet in San Francisco and personally call every single investor that invested in Experiment to share three main things: 




It took us two full years to arrive at this outcome. By June of 2017, we had let go of all our staff in New York City, cut the company costs to a bare minimum, and stopped collecting a paycheck. The generosity of family and friends for free housing afforded us the time, space, and freedom to ask ourselves: 




We experienced the lifestyle of counting every last penny when shopping for groceries. We experienced the fears of not having health insurance when I got severely ill and couldn't afford to go to the emergency room. We revisited our values and independently reconstructed what matters to each of us. Through this meandering we questioned if the way we were building and the foundation we were building upon matched where we intended to take the community, or where the community intended to take us. 

We found a mismatch. 




We stopped working on Experiment because after trying everything we could think of to grow the crowdfunding business, we felt burnt out and unmotivated to continue. It felt like any incremental change we made wouldn't get us to where we wanted this community to go. 

If you are to judge Experiment as a traditional business, we successfully did the business thing of making more money than you spend. We provide something of value to customers and they pay us a fair price in exchange for that value. We built a sustainable business like building a profitable grocery store, but that was never our desired destination. The definition of a startup is a business that is growing. So, as much of an accomplishment being profitable was, it made us feel bad. 

Just because we felt bad didn't mean that the entire community had to suffer the whims of Denny and Cindy. We put together a plan to get Experiment on track to live without us. Today we spend around 4 hours a month as volunteers stewarding Experiment. The platform 

 who review proposals, onboard new scientists, and organize partnerships. 

We needed a way for us to get untired. For me, the best way to get motivated is to start building something I want. 



The foundation that we built on didn't motivate us

 

In efforts to regain founder motivation, we tried something different. We tried to build something new under Experiment. We quickly found that the agreed upon outcome of all stakeholders of the business didn't match what we as founders wanted to guarantee or deliver. Oops. 

Knowing how mentally and physically taxing it is to try to build a company, we felt if we were going to try again we wanted the assurance that, if we put our blood sweat and tears into making this thing, if this thing has a slight chance of working and becoming huge, we want to be able to look back and know we did everything in our power to create a responsible and honest foundation, to know that we did not take short cuts here. 

We wanted a foundation for us to optimize for our purpose, which is not always accurately measured by revenue or served by extracting for financial returns. 

After seven years of running a for-profit Delaware C-Corp, we learned that the traditional incentives didn't feel right for us. The structure didn‘t motivate us as founders to work hard or incentivize the stakeholders to build this organization to last. If we were to keep going down the prescribed traditional Silicon Valley startup path, the community most at risk would be the scientists that use our platform. We don't like that. 

Our foundation felt bad. When something feels bad, it is difficult to get yourself to work on it. If you can't get yourself to work on it, it will never work. 

We stopped working on Experiment because we didn't feel good delivering on the outcomes outlined in the legal structure, which is optimizing for a financial return for shareholders. While this structure seems to be very good at motivating a lot of founders who build incredible companies, the structure didn't motivate us. We needed something that would motivate us. 

We knew parts of the traditional Silicon Valley startup foundation didn't match our intentions, but we were still super motivated to build for science. We knew there must be a better way, so we went searching for something that feels right for us. To find what felt true, we looked at what we felt we did well and what we felt deserved improvement. 



 

We've always valued having real conversations with users. Through this we became lifelong friends with some of the scientists we met through our work on Experiment. One of the reasons why meeting with and learning from users was such a big part of our process at Experiment is because hanging out with scientists is one of our favorite things to do. 

We love hanging out with scientists so much that our vacations usually involve us going somewhere to meet with scientists. Yeah, we are that boring and take 

 very seriously. 

When we feel disconnected, pessimistic, or burnt out the best way to remedy that is to have real conversations with scientists. Meeting scientists where they are and hearing about their challenges and dreams reminds us of the opportunity ahead. Scientists are the humans that inspire us to keep making, the humans that remind us that we can be better. We must be better. 

Our assumptions are almost always grounded in real conversations with real humans. If we believe something to be true, we go out and ask the scientists we serve to see if their opinions and behaviors match our assumptions. We believe this is one of our strengths and one of the reasons why Experiment was able to reach profitability, while many other scientific crowdfunding websites either pivoted or closed their doors. 

We will continue the practice of collaborating with users on what we build. 



Spending entirely on staff and supporting staff

 

During the seven years of operating Experiment, all of our funds went to supporting our staff, recruiting good people and keeping them happy. We plan to spend money similarly. Our operating costs are entirely spent on staff and resources that help our staff do their best work. For the first two years of Jelly we intend to use all funds for staff and supporting staff. 



Excessively long window to exercise options

 

Typically, companies have a 90 day window for employees that leave a company to exercise their options. At Experiment, we extended this window to the maximum amount of time possible because we felt it was the right thing to do for the employees. 

The problem with a short exercise window is employees often don't have enough cash lying around to exercise their options. Even if they did have enough cash lying around, it is unclear if owning options in an early stage company is a good financial decision. 

Our lawyer asked us not to do this again. We probably won't do it in the same exact way, but we do intend to allow early contributors that earn options to have a seemingly unlimited time to exercise those options. While this can be inconvenient for our legal team and cost us more as an organization, we have an excessively long exercise window because we believe this is right thing to do for the people who work with us. 



Strive for growth without compromising on principles

 

Optimizing for a single metric is a practice we will continue to do, especially during the early days. Picking that metric is something we will be more careful about. The metric we choose has to accurately represent us serving our purpose. Accurately quantifying our progress and growing that quantity is how we measure our impact. 

We didn't always do this well at Experiment. Part of the reason why we didn't do this well is because the metric we chose was too early in the user cycle. We don't know what that metric is yet, but we know that focusing on growing a single metric needs to exist in this organization. 

To stay true to our principles we will monitor the health of the organization with measurements such as a set ratio between executive and median worker compensation, quantifiable measurements for quality in the scientific discoveries published on the platform, a measurement of our carbon footprint, etc. These principles will be defined by us as founders to start and we expect stewards to re-evaluate these principles as the company evolves. 

We will focus on growing a single metric, but not at the expense of our principles. 



 

As long as the organization is governed by stewards and not pure financial contributors, being publicly traded is a door we keep open and an option that we strive for if it allows us to serve our purpose. Financial supporters are encouraged to contribute, but financial support alone cannot buy investors voting rights. We intend for all financial investments to result in non-voting stock, which may attract a different kind of investor. 



Not for sale to entities with different values

 

We intend to keep the option of being acquirable open because sometimes the best option for serving your purpose is joining forces, but Jelly will not be acquirable just any organization. 

At Experiment, we can sell the company to anyone we want. Because we as founders still own more than 51% of the company, we can sell the company at anytime for any price. If we get greedy, we can sell the whole thing for private financial gain. That doesn't feel right. Once investors own more than 51% of the company, they can force the board (Denny and I are the board) to sell the company to anyone for any price. That feels really bad. 

Over the years, friends and advisors have asked if we would consider selling Experiment. Through those exercises, we realized that we would consider selling Experiment. The reason why we haven't is we haven't met a trusted buyer that we like. For us, selling has nothing to do with the money. For the right buyer, we would sell for one dollar. 

We intend to add a golden share, a single share that can veto the sale of the company or veto changing the corporate structure away from being steward owned. 

Can an entity like Elsevier acquire us? Elsevier would have to change their corporate governance to do so. If Elsevier transitioned to a steward-owned structure where financial investors are stripped of voting rights and the organization was governed by representation of stewards who are workers at Elsevier, journal editors who help edit and distribute new scientific papers, the scientists that provide the peer-review, the scientists that write the papers, then maybe! 

Do we demand that all peer-reviewed papers be open access? We feel we wouldn't need to demand this if Elsevier were governed by scientists. 



Broadening representation of stakeholders

 

In the traditional technology startup, the board has representation by founders first, then investors, and then maybe some independents if the founders fight for it. We want broader representation within our organization because we believe it results in a more equitable platform and a future that better serves the scientific community and the people that benefit from science. 

We don't know exactly what this looks like yet, but this is a strategic move for us. We believe that the organization will be stronger and more resilient when we make space to hear the needs of all stakeholders. The first step towards this is to separate voting rights from economic rights, where voting rights cannot be purchased. 



 

At Experiment, we never considered a day where we could shut down. It felt almost taboo to discuss failure. If we considered failure as an option, that felt like we were betraying someone. We're not sure who, but it felt like we were betraying someone. It kind of felt like a 

 situation. 

As a result, we never set aside a fund to be able to gracefully shut down. To this day, Experiment still does not have a fund to be able to gracefully shut down. This isn't right. 

Part of the responsibility of being an owner or creator of a thing is to take on the responsibility of delivering on your promise. It is hard to deliver on a promise if that promise is not explicit. 

With Jelly, we will set aside a fund for a graceful shutdown. This will ensure a minimum timeline for delivery of our service if we do need to shut down. This approach makes enterprise customers feel comfortable making a bet on the product because their financial forecasting typically covers longer multiple year time horizons. 

Because everything is built open source, people will be able to move off of our platform and self host. We have plans to make our content interoperable with other online platforms, so we plan to make it easy to export all your data and port over to another service of your choice at anytime. We believe scientists should own their own stuff and if you trust us to host for you, we promise you a set amount of lead time if you need to move to a different home. 



 

We will touch on being open source later, but one thing that we did very late into our time working on Experiment was documenting the intention behind product changes. Yes, anyone joining the team can see what code was changed and when in version control, but the 

 isn't logged. Only if you were there in person at the meeting would you know the 

. As people cycled in and out of Experiment, each member leaving took with them institutional knowledge that was nearly impossible to get back. 

Writing why we are building certain things is just as important as building the things. We think building open source and operating remote-first will help make this documentation more natural. 



Reward good behavior in users by making them owners

 

The only ways to become an owner of Experiment was to be a founder, invest in the company (you needed a minimum of $10,000 lying around and get Denny and Cindy to think favorably of you), work at the company for a minimum of 1 year, or provide enough value that Denny and Cindy wanted you as a trusted advisor. 

If this new company gets big, we want a mechanism for users to become owners too. You can become an owner that participates in the financial upside or you can become an owner that participates in the governance of the organization. 

We don't know yet how we are going to do this, but a first step is separating voting rights from economic rights. By making that separation we can grant governance rights to users without them having to invest to own shares. Broadening who has rights to purchase economic shares is something we plan to do later down the line. 



 

We understand that building Jelly the way we intend to build it may attract different kinds of funders and we are happy about that. We've budgeted in that it may take longer to secure the right resources from the right people and entities. 

If you've had a conversation with us about investing money in Experiment, you know that we test value alignment through a private conversation. No one in first round of financing is surprised that we had no intention of selling the company because we had stated explicitly in words and in writing that we intended to grow Experiment to be very impactful. If we succeeded, we hoped to make them a lot of money. There were no chances of making them a small bit of money. During the second round we made the mistake and didn't test this value alignment as strictly. I felt that funders looked at us and thought that because we were young and naive we might change our minds later (when our priorities in life may change). Our priorities in life haven‘t changed. A $1 million or $1 billion dollar personal payout still isn't appealing to us. What is appealing to us is accomplishing this mission and serving our purpose. 

Now doing this again, we are testing value alignment of funders using a legal structure. Investing without voting rights isn't going to work for everyone, but for the funders that do sign we know they are signing with the belief that this organization has the best chance of success in producing the impact by being self-governed. 



 

We are interested in building to last. It might take us seven years. It might take us seventy years. We hope that we can achieve our purpose as quickly as possible, but we will not compromise our people or the quality of the scientific work in efforts to produce a financial return. 

If this thing we make can exist and grow without us, that is how I will know we as founders did a good job. We think we can operate much faster than we did building Experiment because our cycle times are much shorter. Cycle time is defined as the time between deciding to do a thing and getting feedback from users who are using that thing in production. 

While our time horizons may be longer than a typical startup in a portfolio, you get an experienced founding team that has a track record of learning fast by producing more iterations per block of time. If someone is trying to diversify their portfolio, we are a lower risk, more reliable investment. 



 

For us, choosing an alternative corporate structure is as much of a business strategy as it is a personal decision. What we as founders and leaders are agreeing to deliver to our stakeholders matches with what we personally want to deliver. We sought to fix that mismatch we experienced at Experiment. 

Three things we found we care about: 



What we feel just needs to be different 

 

We don't think building a crowdfunding platform open source would have made much of a difference for Experiment. Deciding to build open source this time around has more to do with the nature of the product that we are building at Jelly. We have a better chance of successful adoption for Jelly building open source. 

We are betting that building out in the open will encourage others to contribute where they can. We feel building openly will result in a more reliable and robust platform, a tool that the community can trust.



 

When we built Experiment, we required that everyone that wanted to work for us move to be with us in our one bedroom apartment on 1st and Folsom in San Francisco. We had a blast 

. Looking back, this decision excluded a lot of talented people from working with us. As we've built up the community of scientists, engineers, designers, organizers from all over the world, we would like for all of them to contribute in a way that is meaningful to our mission. 

We want to give everyone the opportunity to contribute. If they choose to contribute, we want to make sure we are building an inclusive environment for everyone to do their best work. 

One step towards building an inclusive environment is practicing a remote-first culture. By normalizing being remote, we ensure that our process and practices are inclusive towards people who are not sitting in the same room next to us. 

Practicing a remote-first process takes longer to set up and more care and patience to do well, but we feel that building a remote-first team makes us more resilient in the long run. This does not mean we won't have an office. It just means that you don't need to come to the office to be a productive member of the organization. 



Balancing founder returns with worker compensation

 

When raising our first angel round for Jelly, a friend asked me if I hope that this company will make me rich. If the company is really financially successful, we don't mind being compensated for our work. If we are to pay ourselves a high salary, compensation across the organization needs to be balanced. One way to do this is for executives and founders to make a maximum of some multiple of the median pay for a worker at the organization. 

Take an example where executives can only be paid 50 times more than that of the median worker. If the median worker is making 15 dollars per hour or $31,200 annually the maximum an executive could make is $1.56 million annually. Do we each need to make more than $1.56 million a year? Probably not. The median salary at a tech company is probably closer to $80,000 which means the max an executive could make is $4 million annually. This is just an example. 



Separating voting rights from economic rights

 

For Jelly we intend to separate voting rights from economic rights. We intend to create a new class of shares with only voting rights and no economic rights, voting shares. Investing in the company should not automatically give an entity or individual voting rights. 

If you separate voting rights from economic rights, investors can participate in the upside of the business without interfering with the management of the company. There is an opportunity for financial investors to participate in the management of the company, not because they are buying their right to participate, but because they earn their right to participate the same way that any other community member, executive, or worker would. 

The voting shares will be held by the founders for now. If a founder leaves the company, the voting shares are returned to the company. We have yet to plan out how the voting shares will be distributed to workers, to users, to other stewards. The first step for us is to put our foot down and set in stone that voting cannot be directly purchased using cash. 



Building a company where profits serve purpose 


The best thing I feel we can do right now is make this thing work in a way that motivates us as founders, to attract the types of team members and trusted partners we want to attract. 

By saying no to some of the tried and true methods of building a startup it makes this organization that much harder to start. We accept the risk. We accept that risk because the cost is worth it. We may fail. But, if we succeed this organization will be ours, not Denny and Cindy's, but ours, belonging to the scientific community. 

Experiment was built by scientists for scientists. Jelly is built by scientists for scientists, and if we succeed will be governed and owned by scientists too. 