0. In-app Notifications

[Austin] Yup, in-app notifications are absolutely lacking in our site :P

1. Trade Log => Volume Chart

Untitled

[Austin] Agreed that the trade log is pretty noisy; could condense to a comments-only view, and have a volume chart. (Or maybe hide it altogether? I'm noticing that more "consumer" fintech sites like Robinhood/Coinbase don't include it at all)

[James] The most useful info from the trade log is the most recent activity for the market. You want to know

The normal graph is somewhat a substitute, but not really. The volume chart fills this need, I think, but might be hard to decipher for regular folks.

Maybe we could have one line with the 24 hour volume: e.g. “$50 YES and $10 NO bets in last 24 hours”. If that volume is 0, then instead show 24 hours from the most recent bet: “$20 YES and $5 NO bets 7 days ago”

[Akhil] Yes, a few different options to retire the full trade log. Could do with or without the volume chart, or even integrate it into the main chart (see screenshot).

Screen Shot 2022-04-05 at 6.47.32 AM.png

Though to take a step back, generally what I’d like to know is:

There are a few different design elements that can communicate most or all of this without adding to the noise or using too much text.

2. Profile + My Trades => Portfolio

Untitled

[Austin] I've wanted to do this for a while now haha; both "merging Your Trades as a tab in your profile" and "replacing Your Trades with a table/spreadsheet". Thanks for including the screenshot - curious, what site is that?

[Austin] Actually - I've wanted to make Your Trades public, at least for the trades that have resolved (and possibly all of them), as a form of accountability, and as a way to see people's activity. Curious if you have any gut reactions to this. (Possibly the easier sell would be showing all your comments, publicly on your profile)

[James] I like this. Each row is a market, but maybe if you click it, it can expand / popup to show your individual trades for that market.

One reason we don’t do this already is it’s hard to fit a table on mobile. (Especially with long market questions.)

[Akhil] Yeah, I’d really like to see this info personally (Austin, orig screenshot I pulled from image search based on what looked clean UI, not sure which app). If the market question is on its own row followed by the position data row, should be mostly doable on mobile too.

[Akhil, via email] Oh, and yes - Austins' point that the predictions default to public for public questions makes a lot of sense; I'm also more likely to engage in markets that my friends / people I follow have interacted with; and to see how they've bet - all of which make being able to follow people (or import social connections from other platforms) an important part of the ecosystem.

3. Prediction Position

[Austin] We did think about this actually, especially when Loans just came out. I'm in favor, but think it's lower ROI than the other suggestions (it's a bit technically complex, I believe).

[James] I think these are good ideas. Very trader friendly.

[Akhil] One of the big reasons for suggesting this, is in order to drive updates to predictions. An overall question that emerges is whether one imagines a buy and hold approach to most markets, or news/comments driven updates to existing positions ⇒ more engagement.

Also, adding a position, or increasing/decreasing a position should probably use the same UI element. Click UP/YES or DOWN/NO followed by Confirm. Long press or multiple clicks have some amplification behaviour. Reduces me having to enter numbers using the keyboard.

[Akhil] re: technical complexity, You’d lazy load the portfolio data on most pages to be able to add user/portfolio specific context, so shouldn’t be too technically complex?

4. Browse

[Austin] Yeah, my very rough sense is we've leaned too far in the direction of "free for all, anything goes" to the detriment of the typical user browsing our site, and should test a more moderated approach (e.g. having major platform-owned categories of "Sports", "Politics", "Finance", "Crypto" that we curate).

[Austin] I'm less sure about increasing info density (though, I did hear the exact same feedback from another product person I respect a lot!)

[James] Better browsing could have great ROI. I think we should invest in this as well as the feed.

Perhaps the default Explore view (formerly Markets) should be organized by category.

One reason to focus on the feed is for mobile. You can only only show 1-2 markets at time on a mobile screen. Average users don’t want to click to navigate to different categories, they just want to scroll.

[James] Re: density. A unique part of markets is you want check back to see how the prices/probabilities changed. A denser, dashboard-like view for all the markets you care about seems like a use case that we should fill.

[Akhil] More on browse in the twitter vs. reddit question below. re: information & engagement density - at the risk of stating the obvious, but relevant to many of the suggestions:

The Portfolio would probably cover the dashboard like use case; the important question is what info should be included by default everywhere in order to drive engagement and prediction updates. My guess is that Daily/Recent Changes and my existing position on the market would likely be some of those things.

5. Growth Incentives

[Austin] We've thought about a personal affiliate link, more similar to Coinbase/most financial sites that pay for direct referrals. Tracking signups from a market instead is super interesting - reminiscent of what StackOverflow does when you link to an answer.

[Austin] Maybe instead of signups, we should gamify views/traffic; we could just show the number to start, and think about explicitly incentivizing via payments later. Signups are super spiky/rare, traffic is much more granular; if you tweet a market you'd usually get 0, maybe 1 signup. Relatedly, I thought this new section on Reddit was super cool:

Screen Shot 2022-04-03 at 11.43.44 PM.png

[Austin] So far we've considered viral growth mechanisms less important than making the core loop good, but I think soon (in the next 1-2 months) we'll want to change our focus from retention to growth.

[James] Good idea for the near future. Agree with Austin

[Akhil] Yes fair enough. Nailing the core flows and product before opening the floodgates makes a lot of sense.

6. Reddit vs Twitter

A major difference between reddit and twitter is that the former centers around the content, and the latter around WHO created the content. They both have advantages/disadvantages of course; though I do consider reddit more conducive and welcoming to new/unknown users getting the community to engage with their content. Is this something you've explicitly thought about, and taken a call on one way or another?

[Austin] That's a great distinction! The axis I had been thinking about was "community/topic-driven vs algorithmically/follow-driven", but that's a mechanical split; your division captures the spirit of the two sites. Thus far, I've been referencing Twitter as a guide to product and visual design (as you may have noticed!). However, I'm now less sure that's the right model; I think having an engaged community eg Reddit is more important than speed of creation.

[Austin] Some models to consider, by how good they are for Manifold (in my rough estimation):

  1. Reddit / Hacker News / LessWrong - Subcommunities coalescing around the news of the day

  2. StackOverflow / Quora / Wikipedia - Wisdom of the crowds seeking out a single, collaborative truth

  3. Twitter / Substack / Twitch - Emphasis on individual creators putting out entertaining content

  4. Discord / Slack - Gated communities allowing for quick, intimate discussion

[Austin] As you might know, we already have communities in our app; they’re currently an optional “super-tag” that aggregates markets based on whether they fall into a tag. I’d propose one simple but major change: require that every market be grouped into one of eg 10 major categories:

So that as with Reddit, all markets are categorized under at least one community. There are further questions around who moderates these (I think us, at first, and then we can delegate or hire for this as good moderators step up), and when to create new categories.

[James] This is an important question and Austin’s suggestions sound pretty good.

I think that because mobile pushes usage into one feed, that also leads to a world where sub-communities are less important and discussion on individual markets is more important.

The creator of a market is still important because they decide the resolution, which is why I think we should support following individual users like Twitter does. But the feed ultimately should be algorithmically driven IMO because that surfaces the best content.

I think communities are something we can expand later. Austin’s categories does sound like something that could be useful today, so maybe we should try that as well.

[Akhil] Yeah I’d think one could follow both topics and people; as do both reddit and twitter. The question then is one of whether the product and engagement revolves around individuals or topics/markets, and the design decisions that follow from that choice. It sounds like you guys are generally in agreement that the individual, while important, isn’t the primary focus?

re: feed vs. communities - related to James’ point, the default user is unlikely to perfectly curate their communities; nor should that be a requirement to start using the product. So surfacing stuff algorithimically makes sense. And the relevant question is one of depth vs. breadth.

A product mechanism which I think can accomplish both is Swipe vs. Scroll. Swipe being the primary way of navigating, and scroll being a signal to see more of the same (comments, news, related questions etc.). Importantly though, they needn’t be mutually exclusive (eg. swipe may also bring up related markets) - they only serve as suggested signals to see adjacent/general/broad content (swipe) or more depth (scroll).

Combine this with news driven new markets, news driven updates to existing predictions, and minimal clicks to add/update a bet, and I’d probably start every day on Manifold!

Ok, thanks for engaging with all the feedback and my brain dump 🙂. Sounds like these are all things you’re already thinking deeply about, and I look forward to seeing where this goes.