31 comments

[ 2.8 ms ] story [ 74.4 ms ] thread
Why not post the paper?
OP here. AFAICT the paper was internal to FB and leaked to Bloomberg:

> The paper, obtained by Bloomberg and discussed here for the first time...

The paper itself was not linked in the article; is it available?

It's not. It's a rhetorical Q for Bloomberg, junk news sources make money by interpreting primary sources for others instead of providing the data and adding value.
No, they'll probably milk the paper for a few stories before releasing it (I hope).
I hope too, but I'll bet it wont happen any time in the next ten years, if ever. The source relationship too valuable to them, and until the readers demand it, it will stay that way.
I don't believe the numbers. Clinton outspent Trump 2 to 1. She had the most sophisticated ops (including Groundwork (Eric Schmidt-backed startup)). How could Clinton campaign have spent half the other when it was their area of competence ?
I'm convinced 99% of T_D users who spent 12 months posting Pepe memes were actual people inspired by a paid 1%. You can probably save a lot of money if so many will "work" for free.
Your misunderstanding guarantees you will make bad predictions. There is even a place for US citizens to bet money on it; predictit. Re-stating you statement means a 1% of frog posts wins elections. Why? You really don't know. By your logic you should be able to use it though. All JB needs is the right frog...
I would agree with you if Trump weren't such a leftfield candidate. There were plenty of other much safer investments establishment Republicans that would give the 1% a return on their investment than Trump. No one believed he would win at every stage of the game.
From the article, "But Trump’s FB campaigns were more complex than Clinton’s and better leveraged Facebook’s ability to optimize for outcomes."

Also, "The data scientist says 84 percent of Trump’s budget asked people on Facebook to take an action, like donating, compared with 56 percent of Clinton’s."

She outspent Obama too early on.

Trump is a guy who does whatever he can to get attention. Given the pathetic state of the GOP slate and the less than inspirational Hillary campaign... people just needed a little push.

By pathetic, do you mean cleaning house? Last I checked, winning the presidency with Ohio or Pennsylvania to spare isn't pathetic. The old GOP is begging to lose the house and senate. Having control without a president playing ball exposed them way too much. Neither will happen, but it's instructive to watch Ryan's PAC fund DEM's.

Consider McCain. Voted how many times on a full repeal of the "Affordable Care Act" and then when the exact bill had a chance of passing?... Ryan... Graham... McConnel... same party.

Trump exists partially because the parade of candidates in the primaries was really... not good. For all of his faults, Trump isn't a plastic person.

There's a reason why Trump was seen as a breath of fresh air to many, and a reason why the Russians saw an opportunity in him.

Clinton outspent Trump over the entire campaign. Clinton spent $1.18 billion, where Trump spent $617 million.

https://www.bloomberg.com/politics/graphics/2016-presidentia...

So Clinton outspent Trump by nearly double in the entire election, but according to the article, it was the opposite when regarding Facebook spending.

Clinton may spent most money on traditional media channels and wasted her money as described in the article.
The most critical thing you need to consider when evaluating talent is value over replacement.[0] While Hillary Clinton certainly had fantastic credentials (Secretary of State, Senator) her track record for accomplishments[1] while in those positions is IMO indistinguishable from a fictitious replacement level politician. She was the presumptive nominee in 2008 and lost, but you could chalk that up to Barack Obama being highly competent. There isn't much proof that she was an extraordinary politician, in fact there is a large body of evidence to the contrary.

[0] https://en.wikipedia.org/wiki/Value_over_replacement_player

[1] https://www.thebalance.com/hillary-clinton-s-accomplishments...

> An internal Facebook white paper

that sounds like an unbiased source

I don't think there are sources that are simultaneously privy to details of an issue and unbiased.

(It's the same reason why regulatory capture and other crony clubs can happen entirely organically.)

Edit: The closest thing you can get to an unbiased source is a party that doesn't have a vested interest in the story. Given how Facebook sprawls a lot of interests - business, media, government/politics, culture/sociology, technology - there aren't many parties without bias.

could easily be less biased if it was a joint publication from facebook + whoever managed that campaign for trump and clinton. they both have access to those details.
Yes, of course they were. They also made better use of Twitter, and took advantage of the broken clickbait-based news market by generating cheap controversies and using strong language to get free publicity. There's no meme that spreads faster than anger, people love sharing things that make them angry, and it was heavily exploited by this campaign. No matter what you think about his policies, Trump is a master politician that knows how to play his opponents and the media like a fiddle. It's low brow, and it's brutal, but it works.
OR... it is possible that people researched the history and positions of the Bush/Clinton dynasties?
Possible that people researched something instead of falling victim of confirmation bias and internet memes? Unless you mean by researching something very different than me, you even thinking that is possible makes me guess your sample of "people" is heavily biased.
I don't assume people are so easily fooled. Especially by those two families. Take 5 minutes to research the latest creepy front runner (JB). There's not even a chance, but the central news pretends. I think the data (a solid win by an outsider) supports that.
(comment deleted)
(comment deleted)
tl;dr Trump's team tested roughly 90X the variations of ads that Clinton's team did (5.9MM vs 66K), and aggressively utilized Facebook's powerful lookalike audience tool, while Clinton used it for only 4% of her ads. Trump ads also allowed Facebook's ad algorithm to optimize for actions, where Clinton focused on simple exposure.

In other words, Trump's team let the Facebook ads system do its thing, utilizing the same tools that other marketers on the platform use regularly. Hillary's team used it more like a traditional media buy, leaving Facebook's most powerful advertising features largely unused. To the extent that one believes Facebook ads had any effect on the election, this seems to at least partially answer the question of how things ended up where they did.

I don't understand why are you downvoted. 5.9M of different ads vs 66K ads, that's ridiculous ratio.

It's more than tenfold difference, and it definitely influences the user engagement down to final conversion, err, sorry, vote.

(comment deleted)
> The paper, obtained by Bloomberg and discussed here for the first time, describes in granular detail the difference between Trump’s campaign, which was focused on finding new donors, and Clinton’s campaign, which concentrated on ensuring Clinton had broad appeal.

Here is the heart of the matter, sans the latest attempt at naming a scapegoat for Clinton's loss.

A populist candidate focused on leveraging the people who were becoming receptive to his message, while an establishment candidate focused on trying to nudge over people who weren't inherently receptive to her message. That's the whole campaign in a nutshell.