Manifold Team March Retro: What could have gone better?

100

Ṁ247

Ṁ515

resolved Apr 2

ALL

14%14%

Spikes in database usage routinely make for a bad user experience

12%10%

ND1 ~flat despite efforts

12%10%

Partner program hasn't brought in many new creators

10%9%

Low politics vertical usage

10%9%

Bet panel Yes/No outcome switcher is missed

10%9%

No major politics sponsorships/deals/promos

10%8%

Numeric markets taking a while and not obviously on track

7%8%

Pathos scam

6%4%

Buying the gold checkmark on X was not worth it

6%3%

progress on migration

2%1.7%

Vercel revalidation still erroring frequently

15%Other

Where did we drop the ball? The point of this isn't to point fingers or assign blame, but rather to identify where our processes and strategy are not serving us well.

This time, I will actually resolve the market to a rough approximation of what I think actually didn't go well. Place your bets!

Last time: https://manifold.markets/JamesGrugett/manifold-team-feb-retro-what-could

Get Ṁ600 play money

🏅 Top traders

#	Name	Total profit
1		Ṁ3
2		Ṁ3

9 Comments

5 Holders

13 Trades

Sort by:

Numeric markets taking a while and not obviously on track

[J] User feedback has not been good. Needs more iteration.
[Joshua] Could we switch the average number graph to the distribution chart
[Joshua] Not sure we should launch them when they are bad.
[i] Going to add limit orders and selling

Bet panel Yes/No outcome switcher is missed

But you don't hear from all the new users that they aren't confused about the new bet panel now

@ian Not being able to switch between Yes and No just makes me use the site less. I just close the market because I'm frustrated.

Partner program hasn't brought in many new creators

Seems it hasn't moved the user numbers yet and is expensive?

[Joshua] Anecdotally several users have stepped up the quality and quantity of markets they create. Succeeded in making people make good markets, which seems good in the long term.

[D] Strutheo said he created 500 binary markets that didn't go anywhere, and then one multiple choice market blew up, so he would do more of that.
[Joshua] Partner program has also incentivized people to do whatever to hack the algorithm, adding more tags, etc.

Spikes in database usage routinely make for a bad user experience

bought Ṁ10 Spikes in database u... YES

This to me is a slow killer. It feels like 1 of out 4 or 5 times I use the site it's annoying or unusable. This will make any user churn, especially new or casual users. It makes me churn and I'm a power user.

@ian Lots of CPU spikes. Will spike randomly, even at night. E.g. from 9pm last night.

[J] Marshall is skeptical about it being the API. He was was going to create a dashboard for non-cache hits for read api
[i] Could be people just hitting supabase directly