The top submissions (max 20, first round) will be used to create image ads that will be shared as ads across different social media platforms.
Facebook, Instagram, X/Twitter, Snapchat, Reddit, Pinterest,
Most likely, all ads will be posted on Meta then the best performers will be shared on other platforms, because Meta facilitates fair bulk a/b testing in a very efficient way
I will resolve 100% to the best performing ad, by results (lowest CAC, CPC, etc.).
I might refuse to create/run an ad if the copy is obviously bad or offensive
Example ads from okcupid, Hinge, thursday, tinder, bumble
The best submissions will share in a M5,000 bounty!
If there is one that outperforms the others by a significant margin, it will receive the entire bounty.
If your messaging requires it, you can suggest an image/theme too. Default will be basic text on solid colour background + logo & "Manifold.Love". Finished image submissions that get used will all receive M500, regardless of results
If the lines get used in Google search ad copy, you will also get a bounty!
1,000🏅 Top traders
| # | Trader | Total profit |
|---|---|---|
| 1 | Ṁ2,914 | |
| 2 | Ṁ559 | |
| 3 | Ṁ412 | |
| 4 | Ṁ409 | |
| 5 | Ṁ345 |
People are also trading
FYI, tests were a bit scuffed
(A) - Kinda scammed the win, bc it didn't spend as much. Less data, but did better per-cost which is what I intended to resolve on.
(I) & (D) - Tied 2nd best. Spent more, probably the real winners, but were tied on performance and it would have been impossible to pick one winner from these two.
I'll send M2.5k bounty to the creators of (A) and (D)
Delays, approval issues, etc., meant that this didn't go as smoothly as planned. Thank you everyone for your input either way, it was fun!
@Joshua Yeah I probably would have multi resolved if I realised I was going to be the "winner". I consulted briefly with tumbles who agreed that this resolution was valid - he was not told which option was "scamming the win" just that it was kinda scuffed and imperfect.
I would have liked to A/B test them against each other battle royal style until a real winner, but bc Manifold Love is on ice for a bit there wasn't much point spending lots of time on it -- especially because of some really weird approval issues
Here's a snippet of what our DMs looked like. Genzy had said there was a clear top three options, though he didn't mention which options were which
If anyone feels like they could make a good ad for any of these concepts or thinks something should be included that isn't, make a 1080x1080 image/video and I will run it. Will give mana grants if they are good!
Update for bettors: some of the terms have been used in copy to varying success, and the terms that are being made into the fb ads are as follows, and I have tagged their answers with (A), (B), (C), (D), etc. to make clear which answer they correspond to. Bet accordingly.
ads will run in 1080x1080 and 1080x1920, on meta platforms, image and video versions.
"Find your statistically significant other" - will resolve (A)
"Where probability becomes possibility" - will resolve (B)
"Browse all existing profiles, filter your preferences, find your match", "All Profiles unlocked: filter for your match", (and any other variants) - will resolve (C)
"You're too special for an algorithm" - will resolve (D)
"Improve your chances with San Francisco matchmakers"/"Let Expert Matchmakers find your match" will resolve (E)
"Bringing back the best in online dating"/of online dating - will resolve (F)
"Somewhere on the end of the bell curve is a partner for you" - (G) - I really need a unique illustration/animation for this for it to work properly I think. Will probably work on that, but it won't hold up the rest of the set, it might run late, or not at all, data will be skewed.
"when surveyed.... " - (H) -this one is unique bc it's so long, it starts full page text and transitions to the CTA. Will only run it as a video, not sure how well it will do but after the first test batch I will probably move it to retargeting only (or cut it). not getting fair/equal treatment so keep that in mind
Finalizing the versions to start running this weekend (Monday at the latest) so the test batch will run through valentines day. Tests will start on the 4th in the US and run until at least the 16th/17th.
Ad copy may contain other submissions, (e.g. themes from (C) and (F) will probably feature in post text with the other images) but that isn't considered, just primary text on the ad image/video itself
"Love's a gamble, why not bet on it" - this and some others like it are not being tried, due to the platform changes and removing betting.
The Manifold Marketing team submissions (me and chatGPT):
"Set sail on the sea of love, be shipped by our matchmakers" (I)
"Beyond first impressions: Browse in depth profiles" (J)
"From compatibility to chemistry: score your perfect connection" (K)
"Unlock love's potential: The score that leads to more" (L)
"A compatibility score that leads to more" (M)
/Gen/will-chatgpt-beat-manifold-at-selec
ChatGPT selected "Find your Statistically significant other" as the best of these. It suggested that they should be bet on as follows:
(A) 20%, (D) 18%, (B) 15%, (L) 10%, (K) 10%, (E) 7%, (F) 5%, (M) 5%, (G) 3%, (I) 2%, (J) 2%, (H) 2%, (C) 1%
read criteria of the gpt market before betting there.
You can add new options still, but they won't be given fair treatment
edit: also some of these have already been used/tested, just not on meta platforms.
just a reminder that this isn't the vote if you're hoping to increase the pool of (heterosexual) women 🙃
There is mana to be won! Submit new answers before it's too late!
FYI these ads will be made soon, not at the close date (templates for all placements in progress now). I expect the tests will be concluded before the close date, giving us time to make valentines day ads and push those when the test sets are finished.
Find your Statistically Significant Other
is clever, but no way is it 15% likely to perform best.
Also, I will not resolve to Other, it will be the closest thematical match if I have changed it a bit, I have decided I will not add my variants to this list later, I will count them as the original prompt that sparked the variation if nobody submits variations themselves. This means, if the best performing ad is
Online Dating
used to be good
is good again
it would resolve 100% to
Online Dating used to be good. We're bringing that back.
and Joshua would get at least M1,000. Maybe all M5k if nothing else is close after testing
@PlasmaBallin Outside chance it performs well because people find it so strange that they have to click on it to see what it is
@PlasmaBallin Somehow I misread this as saying BTE instead of BET. So I guess it's actually not an inside joke. Though I still think "bettor" is a much better pun than "BETter".