0

Deploy an already built prediction website

$0raised
$12,500valuation

Project description

Several years ago, me and a couple of friends built a forecasting website. It is 100% complete, but currently parked in a github repo.

This would be a second attempt to ship the website. The first time around we failed to get users due to inexperience (and the 100$ marketing budget).

The concept of this project is to ship the website, but:

  • focus on the smaller, Bulgarian market

  • focus on Bulgarian politics, economics, international relations, and policy

  • hire actual marketing people to get initial traction

  • have (small) leaderboard prizes to get initial traction

The website concept is very similar to Reddit/hackernews, with the following differences:

  • there are no subreddits, just the one main page

  • some news articles have various forecasting questions attached to them

  • each user has different upvote strength, based on accumulated prediction scores

  • reputation decays over time

  • there is also a leaderboard with scores

Here is what it looks like - https://imgur.com/a/smvl0n3

The idea is to tie status in the community to accurate forecasting, but still leave room for corrections by the crowd (downvotes). The fact that reputation is based on actual predictions of reality means that it will not devolve into an echo chamber.

Distributing prizes based on the leaderboard, as opposed to bets on individual questions, means it's not classified as gambling in the local jurisdiction.

What are the benefits?

  • You will know whether this alternative approach is viable and/or better than existing platforms.

  • You will know whether a smaller linguistic noosphere can support a prediction community.

  • If we are successful, we will export the model to other European countries. It is valuable to have policymakers *aware* that prediction markets are a thing, before actually using them to guide policy.

  • Most importantly, it is valuable for more effective policy here! I.e. Bulgaria is looking to adopt the euro, most likely before 2025, and there is a lot of chatter whether this will cause inflation. So a question on the inflation, conditional on Euro adoption, could help.

What is your track record on similar projects?

Well, the website IS built, so you know we can do it.

I am currently a data engineer, so I can handle at least the engineering side.

In terms of community building, and growth, I have no experience, but I have some friends and contacts who might be a good fit for the job.

How will you spend your funding?

For the full amount, I will do the following things:

  • pay for various website hosting, firm registering, domain buying costs

  • run and maintain the website for 6 months

  • translate the website

  • have a (small) marketing campaign

  • have a (small) SEO campaign

  • do some (minor) UX tweaks

  • pay a community manager to moderate the content

  • have some (small) leaderboard prizes - on the order of 5000$/10k BGN total.

  • reach out to various political science departments and attempt to recruit their students to be participants

  • (maybe) send a press releases at the start and end of competition period

  • (maybe) send a press release comparing predictions of elections to polling agencies

  • apply for other sources of non-profit financing and/or sponsorships, in order to try to make the project long-term sustainable

For smaller amounts I will only afford to do some of these things.

🍄

Petar Buyukliev

over 1 year ago

@Austin

Yeah, building the community is definitely harder than building the engineering side. It is hugely impressive that you managed to bootstrap Manifold to have so much activity in so little time.

I agree, it IS a niche within a niche. But -- isn't it sort of the point to get to such a depth? My ideal version of the world includes using prediction markets for more than just US elections and crypto prices.

I want a decent number of people predicting whether a certain policy will increase birth rate, or whether a nuclear power station project will be built in time and within the budget. Those are indeed very niche, nerdy things, but also things on which it is important to get things right. So I am trying.

But also, even the case of failure is not in vain. If it turns out that it is impossible (or very hard) to get a community of 100-200 people in Bulgaria to be regular users, that is also a data point. The next guy will know that prediction markets for small countries are very tricky/hard. We will have some practical evidence against any futarchy ZEDE/charter city ideas.

>feedback from the first set of users

The feedback was generally polite interest - I didn't manage to get the ball rolling to get anything more.

>What do you plan on doing differently this time around to try and get more usage?

Specifically:

  • focus on community building, rather than on features

  • have leaderboard prizes - which is not trivial, even for small amounts, in poor countries

  • have a marketing budget

  • consult actual marketing/SEO experts

  • do things that don't scale that well - press releases, cajoling university students, facebook posts, get a core of users to participate by badgering them IRL, send personal invitations to the political analyst class, give out nice certificates that people can show off and put on linkedin, hire an intern or two, hold an essay contest

  • apply for more sources of financing

🍄

Petar Buyukliev

over 1 year ago

@Austin

Yeah, building the community is definitely harder than building the engineering side. It is hugely impressive that you managed to bootstrap Manifold to have so much activity in so little time.

I agree, it IS a niche within a niche. But -- isn't it sort of the point to get to such a depth? My ideal version of the world includes using prediction markets for more than just US elections and crypto prices.

I want a decent number of people predicting whether a certain policy will increase birth rate, or whether a nuclear power station project will be built in time and within the budget. Those are indeed very niche, nerdy things, but also things on which it is important to get things right. So I am trying.

But also, even the case of failure is not in vain. If it turns out that it is impossible (or very hard) to get a community of 100-200 people in Bulgaria to be regular users, that is also a data point. The next guy will know that prediction markets for small countries are very tricky/hard. We will have some practical evidence against any futarchy ZEDE/charter city ideas.

>feedback from the first set of users

The feedback was generally polite interest - I didn't manage to get the ball rolling to get anything more.

>What do you plan on doing differently this time around to try and get more usage?

Specifically:

  • focus on community building, rather than on features

  • have leaderboard prizes - which is not trivial, even for small amounts, in poor countries

  • have a marketing budget

  • consult actual marketing/SEO experts

  • do things that don't scale that well - press releases, cajoling university students, facebook posts, get a core of users to participate by badgering them IRL, send personal invitations to the political analyst class, give out nice certificates that people can show off and put on linkedin, hire an intern or two, hold an essay contest

  • apply for more sources of financing

Austin avatar

Austin Chen

over 1 year ago

Hey! I think it's cool that you've already built and shipped this once already -- I'd love to see more prediction sites flourishing! I appreciate that you provided an image of the site too; it looks pretty polished, and the image really helps us understand how the site would function.

Given that the site is already mostly built, it seems like your hardest challenge will be finding users who are excited to participate -- especially if you're targeting the Bulgarian audience, as forecasting is already something of a niche, so Bulgarian forecasting would seem to be a niche within a niche. To that end, I'd definitely recommend conducting user interviews with people who you think might be a good fit (I found the books "The Mom Test" and "Talking to Humans" to really help me get comfortable with user interviews).

A couple questions:

  • What kind of feedback did your first set of users have on the site.

  • What do you plan on doing differently this time around to try and get more usage?

🍄

Petar Buyukliev

over 1 year ago

@Ernest

>Where does the news come from?

The links are submitted by users, like reddit/hackernews. The markets are created and added by the team operating the site.

>What advantages does it have that aren't met by MM?

It's hard for me to really answer the question. What advantages did reddit have that weren't met by Facebook? None, it was just different.

And this site operates completely differently from manifold.

  • Different processing of information

  • Different topics

  • The structure encourages users to participate in every forecast.

🍄

Petar Buyukliev

over 1 year ago

@Ernest

>Where does the news come from?

The links are submitted by users, like reddit/hackernews. The markets are created and added by the team operating the site.

>What advantages does it have that aren't met by MM?

It's hard for me to really answer the question. What advantages did reddit have that weren't met by Facebook? None, it was just different.

And this site operates completely differently from manifold.

  • Different processing of information

  • Different topics

  • The structure encourages users to participate in every forecast.

ErnestFrench avatar

Ernest French

over 1 year ago

Where does the news come from? What advantages does it have that aren't met by MM?