FB - Facebook

[Liberty]

https://www.washingtonpost.com/technology/2019/01/18/us-regulators-have-met-discuss-imposing-record-setting-fine-against-facebook-some-its-privacy-violations/

[johnny]

Also wanted to revisit this Pew survey from Summer '18:

http://www.pewresearch.org/fact-tank/2018/09/05/americans-are-changing-their-relationship-with-facebook/

 

26% of surveyed adults said that they had deleted the app from their phone in the last twelve months.

 

I think it's safe to say that, looking at the completely lack of any movement in DAU/MAUs for core FB during (and after) the relevant period), that to the extent this happened at all, it was substantially short-lived for almost all of those people. More cynically, I suspect a lot of people in that Pew survey are just giving the answer that they believe is mandatory in the cultural moment (sounds nicer than lying, yeah?).

 

I was actually reminded of this because of how dangerous the numbers seemed for younger users: 44% of 18-29 year olds in the sample said they had uninstalled. This was in the back of my head as I was combing through the study data for college-students in the above post. Doesn't really add up, does it?

[JayGatsby]

Their app is super annoying and doesn't do anything you can't do from the mobile website. People probably got tired of constant push notifications about irrelevant stuff and deleted it.

 

DAU is a silly metric. Are people active for one minute a day or 2 hours a day? Given there's almost no original content left on my feed I can't see how that number hasn't declined.

 

Our younger college/below employees use Instagram instead. Nobody below the age of about 25 wants to be on a social network with their parents.

 

Their ad platform is incredible, and still cheap. The best thing about it is it's real time, so you know if an ad is working or not. So the investment in ads that don't work is close to $0. I put an ad up last night that bombed, so I killed it, and now have one that seems to be working.

 

 

[johnny]

DAU is a silly metric. Are people active for one minute a day or 2 hours a day? Given there's almost no original content left on my feed I can't see how that number hasn't declined.

 

Over what time period? Their ARPU increase in the US market is so strong--hard for me to square that with a headwind like what you propose.

[JayGatsby]

DAU is a silly metric. Are people active for one minute a day or 2 hours a day? Given there's almost no original content left on my feed I can't see how that number hasn't declined.

 

Over what time period? Their ARPU increase in the US market is so strong--hard for me to square that with a headwind like what you propose.

Since 2005.  The decline of original content has been fairly widely reported: https://mic.com/articles/176599/the-facebook-wall-is-dead-social-network-struggles-to-get-personal-again#.wpMRLu0CI

 

The ARPU increases have surprised me as well. They've increased ad load and pricing. Ad load is probably maxed out now, but they may be able to continue to increase price. It's all an auction system, so in theory if inventory declines because people spend less time on facebook prices will increase.

[johnny]

It's definitely true that the 2019 core FB product is totally different than the 2009 product. Sort of what you'd expect with ARPUs going from $1 to $100.

 

It's also true that, subjectively, I find nothing appealing about the core 2019 product v the 2009 product. This feels as much a Johnnychange as a Facebookchange, though. And again, even if "everybody agrees" core Facebook sucks now, that only makes the situation more curious--why does everybody continue to use something that they universally agree is bad and terrible? Do we put a higher or lower multiple on a business which is able to keep its users fully onboard, over the long-term, even when half of them at any given time are adamant, in a survey, that they wish to quit?

 

It's possible that the average "DAU" is actually much lower quality today than even 4 or 5 years ago, but it's hard to imagine that this wouldn't suggest itself in some of the data. If the core product is so miserable now that the average DAU is spending much less time, shouldn't there be some detectable DAU drop-offs on the margin? Instead you see some of the most distressingly stable figures you could hope to see.

[walkie518]

It's definitely true that the 2019 core FB product is totally different than the 2009 product. Sort of what you'd expect with ARPUs going from $1 to $100.

 

It's also true that, subjectively, I find nothing appealing about the core 2019 product v the 2009 product. This feels as much a Johnnychange as a Facebookchange, though. And again, even if "everybody agrees" core Facebook sucks now, that only makes the situation more curious--why does everybody continue to use something that they universally agree is bad and terrible? Do we put a higher or lower multiple on a business which is able to keep its users fully onboard, over the long-term, even when half of them at any given time are adamant, in a survey, that they wish to quit?

 

It's possible that the average "DAU" is actually much lower quality today than even 4 or 5 years ago, but it's hard to imagine that this wouldn't suggest itself in some of the data. If the core product is so miserable now that the average DAU is spending much less time, shouldn't there be some detectable DAU drop-offs on the margin? Instead you see some of the most distressingly stable figures you could hope to see.

there is likely more churn under the hood, there could be possible growth at Instagram and decline in Facebook, and it's also likely that those numbers have been somewhat adulterated

 

we learned that 1.3B users were created by third parties to later be deleted by Facebook as those accounts were deemed fraudulent, but by whom?  By Facebook or the no-name third-party Facebook hires to clean its dbs?

What would happen if more than 1.3B were reported as fraudulent?  Facebook also claims that this happened simultaneously, they were simultaneously created and deleted.  I have a very hard time believing this...

 

Facebook also claims that if you know how to use their tools effectively, your ad placements are effective...that is, if you don't know how to use their tools, it's likely that you are subject to click-fraud

 

I'm curious how many DAUs Facebook's IT security personnel and/or db admins believe they have and what trends they see, trusting the public statements poses many problems.  If push came to shove, I'm sure your intuition is correct that the number of real DAUs is far less than reported

[JayGatsby]

Click fraud really isn't an issue with FB. You choose how you pay (reach, clicks, views, etc) and they provide all the tools to measure if your campaign is working. It really all comes down to conversions, and their pixel lets you track that.

 

Click fraud is really only an issue on third party websites. If CoBF had Google adsense (or whatever they call it now), we could all click the ads and money would be generated for CoBF. On facebook that isn't a thing, unless facebook was actually clicking the ads themselves which would just be stupid.

 

I believe they have a lot of DAUs, I just think it's a silly metric. Same reason netflix focuses on screen time.

[johnny]

there is likely more churn under the hood, there could be possible growth at Instagram and decline in Facebook, and it's also likely that those numbers have been somewhat adulterated

 

My first thought too, but then I remembered they're actually specifically surfacing "core FB" numbers; they rather notably refuse to provide the same level of visibility into the use/user trends for the other products (except intermittent, non-systematic bragging about arbitrary thresholds being crossed).

 

That said, churn is still a possibility.

 

we learned that 1.3B users were created by third parties to later be deleted by Facebook as those accounts were deemed fraudulent, but by whom?  By Facebook or the no-name third-party Facebook hires to clean its dbs?

What would happen if more than 1.3B were reported as fraudulent?  Facebook also claims that this happened simultaneously, they were simultaneously created and deleted.  I have a very hard time believing this...

 

If there are a lot of "fake" accounts, that's not so much a problem. The problem would be substantial changes in the ratio of fake-to-real accounts. In other words, whatever their current fake-account-load is, that is baked into the cake financially--we are comfortable with those numbers, whatever they are. With the fake hunting/purging they've done, their reported numbers have remained impressively stable, to the extent you are inclined to trust their reporting.

 

I think I lean towards giving them the benefit of the doubt here: speaking from both personal experience, and anecdotal/network stories, they are in fact almost totalitarian in their imposition of certain "rules" about accounts representing actual real humans (to the point of, quite creepily, demanding that I and other people I know upload government-issued ID to prove legal names, etc.) Don't get me wrong, this doesn't prevent fake accounts from being created and maintained, but it suggests to me they take the issue seriously and would not be likely to mischaracterize it knowingly. But I agree with Jay that the way they present use statistics isn't maximally compelling. Very crude and very incomplete, and not much of a typical corporate-strategy rationale for being especially secretive about it.

[walkie518]

there is likely more churn under the hood, there could be possible growth at Instagram and decline in Facebook, and it's also likely that those numbers have been somewhat adulterated

 

My first thought too, but then I remembered they're actually specifically surfacing "core FB" numbers; they rather notably refuse to provide the same level of visibility into the use/user trends for the other products (except intermittent, non-systematic bragging about arbitrary thresholds being crossed).

 

That said, churn is still a possibility.

 

we learned that 1.3B users were created by third parties to later be deleted by Facebook as those accounts were deemed fraudulent, but by whom?  By Facebook or the no-name third-party Facebook hires to clean its dbs?

What would happen if more than 1.3B were reported as fraudulent?  Facebook also claims that this happened simultaneously, they were simultaneously created and deleted.  I have a very hard time believing this...

 

If there are a lot of "fake" accounts, that's not so much a problem. The problem would be substantial changes in the ratio of fake-to-real accounts. In other words, whatever their current fake-account-load is, that is baked into the cake financially--we are comfortable with those numbers, whatever they are. With the fake hunting/purging they've done, their reported numbers have remained impressively stable, to the extent you are inclined to trust their reporting.

 

I think I lean towards giving them the benefit of the doubt here: speaking from both personal experience, and anecdotal/network stories, they are in fact almost totalitarian in their imposition of certain "rules" about accounts representing actual real humans (to the point of, quite creepily, demanding that I and other people I know upload government-issued ID to prove legal names, etc.) Don't get me wrong, this doesn't prevent fake accounts from being created and maintained, but it suggests to me they take the issue seriously and would not be likely to mischaracterize it knowingly. But I agree with Jay that the way they present use statistics isn't maximally compelling. Very crude and very incomplete, and not much of a typical corporate-strategy rationale for being especially secretive about it.

Your argument holds and is likely why Facebook trades at rich multiples to cash flow.

 

Putting aside Cambridge Analytica, the second half of your argument is that Facebook appears to "take the issue seriously and would not be likely to mischaracterize it knowingly."  I disagree on this point for a few reasons.

 

Before the scandal broke, FB's chief security officer published a paper (attached) on information amplification, fake news, and fake accounts on Facebook and the mounting problem.  I believe not long after, Zuck said that Facebook had no fake accounts etc.  Then the scandal broke and Facebook failed to communicate. 

 

That Zuck was not aware of the paper to be published would mean that he doesn't talk to his chief of security with any regularity.  I have a very hard time believing that Zuck did not know and a harder time that he was providing information to the best of his knowledge in that interview. 

 

Additionally, Sandberg is no better though seemingly less disingenuous.  She can't be less informed than Zuck? Sandberg also had a habit of speaking very publicly.  In some ways, she is a trailblazer, in other ways, she's indistinguishable from others in similar roles despite difference in gender.  Her outspokenness prior to the scandal, I think, speaks volumes of her character.  She is either oblivious to the soft spots of the company she runs or she believes everything she says publicly wholeheartedly.  In either case, this behavior in light of the scandal either highlights a level of moral corruption or sheer incompetence.  I have my vote as to which one is more likely.

 

at the highest level of the corp ladder--who is to say that this kind of moral corruption does not exist further down the totem pole? 

 

Is it because they hire the brightest minds who are less likely to be dishonest?  Or is it that the ones who win are dishonest and therefore those who are both bright and dishonest outperform those who are bright and honest? 

 

This sets a very, very poor example not only for corporate stewardship but for the everyone in the business.  It's also likely that this kind of behavior echoes and in its own way amplifies...a corporate culture like this might be aligned with Zuck's every want and desire, but it certainly is not in alignment with shareholders. 

 

How this reflects on the share price is much more theoretical, however.  I would think that poor governance and a rotten corporate culture would eventually yield a poor stock price, but at 22x P/E and 14x cash flow from ops, and numbers indicating path to higher prices after baking in a little growth, pulling out cash on the books, and a little compounding of the existing businesses, it's hard to say when the next scandal breaks or if anyone ever takes any real responsibility?  Those who advertise on the platform are also complicit: they are telling Facebook with their dollars that what has happened is ok. 

 

The unilever executive made a stand (though not sure if they stopped selling ads on the platform), but this all gives me the willies.

facebook-and-information-operations-v1.pdf

[Spekulatius]

^ I see this different. Social media was like the Wild West in the 1900 Century You Grab the land first, no matter what and figure out the rules later, because there are none yet. Now the sheriff comes into town and the place gets cleaned up.

Sure, you could say that this should have happened sooner, but the fact is that everyone tried to grab as much land than they could, nothing else really mattered, that’s why they made it to the top.

[Liberty]

Zuckerberg WSJ op-ed:

 

https://www.wsj.com/articles/the-facts-about-facebook-11548374613

 

Also:

 

https://www.nytimes.com/2019/01/25/technology/facebook-instagram-whatsapp-messenger.html

[walkie518]

https://www.bloomberg.com/news/articles/2019-01-28/facebook-watch-struggles-to-deliver-hits-or-advertisers?cmpid=BBD012819_CUS&utm_medium=email&utm_source=newsletter&utm_term=190128&utm_campaign=closeamericas

[Shooter MacGavin]

https://www.plainsite.org/realitycheck/fb.pdf

 

Did anyone see this?  really interesting read.  Seems very biased but worth understanding the negative side of the story.

[cameronfen]

https://www.plainsite.org/realitycheck/fb.pdf

 

Did anyone see this?  really interesting read.  Seems very biased but worth understanding the negative side of the story.

 

I'll definitely read.  I dont understand though, whether facebook has 1 million fake accounts or 1 billion fake accounts, ultimately advertisers are paying based on irr.  Unless the irr calculations are not based on purchases and clicks data but on less obviously related (fake) account impressions, I dont think the number of fake accounts matters to advertisers.  Maybe he addresses this, but curious...

[Spekulatius]

https://www.plainsite.org/realitycheck/fb.pdf

 

Did anyone see this?  really interesting read.  Seems very biased but worth understanding the negative side of the story.

 

I'll definitely read.  I dont understand though, whether facebook has 1 million fake accounts or 1 billion fake accounts, ultimately advertisers are paying based on irr.  Unless the irr calculations are not based on purchases and clicks data but on less obviously related (fake) account impressions, I dont think the number of fake accounts matters to advertisers.  Maybe he addresses this, but curious...

 

That’s correct. If FB indeed has that many fake account, it doesn’t matter because the FB ad $ are driven by ROE ( people who click and buy) not the amount of accounts. In a way, more fake accounts are even bullish, since it means that the addressable market may be higher than thought. Right now the thesis is that everyone has already a FB / Instagram account in NA.

 

My bigger concern is the total addressable market.  FB and GOOG have already a significant part of the ad market, which tends to have a more or less fixed percentage of the GNP. Facebook needs to expand the TAM, in ads to Europe and Asia and ultimately beyond ads, if they want to continue growing.

[Jurgis]

I know people (mostly younger generation) who have more than one FB account. Mostly it's "I graduated from HS/college, good bye FB account and its history, getting a new one". They mostly don't care about the history (or actively want to get rid of it) and don't care about FB "real name" protections.

 

 

 

I only skimmed the plainsite doc, but it seems like "I knew Zuckerberg in 2004 and OMG WHAT HE WROTE IN INTERNET CHAT" collection. More than half of the doc is pre 2004 history... I liked the doc about sex orgies at Google better.  8)

 

[walkie518]

https://www.plainsite.org/realitycheck/fb.pdf

 

Did anyone see this?  really interesting read.  Seems very biased but worth understanding the negative side of the story.

 

I'll definitely read.  I dont understand though, whether facebook has 1 million fake accounts or 1 billion fake accounts, ultimately advertisers are paying based on irr.  Unless the irr calculations are not based on purchases and clicks data but on less obviously related (fake) account impressions, I dont think the number of fake accounts matters to advertisers.  Maybe he addresses this, but curious...

yes, unfortunately, IRR is more important, but one has to keep in mind that at the very top of the organization, moral corruption is present; this way of thinking will likely result in many future scandals regardless of the fires put out today

 

 

[cameronfen]

https://www.plainsite.org/realitycheck/fb.pdf

 

Did anyone see this?  really interesting read.  Seems very biased but worth understanding the negative side of the story.

 

I'll definitely read.  I dont understand though, whether facebook has 1 million fake accounts or 1 billion fake accounts, ultimately advertisers are paying based on irr.  Unless the irr calculations are not based on purchases and clicks data but on less obviously related (fake) account impressions, I dont think the number of fake accounts matters to advertisers.  Maybe he addresses this, but curious...

yes, unfortunately, IRR is more important, but one has to keep in mind that at the very top of the organization, moral corruption is present; this way of thinking will likely result in many future scandals regardless of the fires put out today

 

So maybe one can say Zuck is laser focused on returns and the good of the company at the expense of morals, but I dont think that's true.  Was facebook not paying enough attention to the problem, yes.  Did they try to cover stuff up, yes.  Will further data breaches happen, of course.  When your business model is predicated on matching the data of every single (non-chinese) person in the world, with basically every single advertiser, there is no chance no matter how much you spend that a breach somewhere wont crop up, exposing millions of peoples data.  With software that complex people dont understand that you can never write code that will take into account everything and the average person is becoming fine (I think) with risking the victim of data breaches, just like we are now desensitized to the fact that windows has like 1000s of security flaws to the point where most windows users dont patch their system even when Microsoft is begging them to. 

[Shooter MacGavin]

https://www.plainsite.org/realitycheck/fb.pdf

 

Did anyone see this?  really interesting read.  Seems very biased but worth understanding the negative side of the story.

 

I'll definitely read.  I dont understand though, whether facebook has 1 million fake accounts or 1 billion fake accounts, ultimately advertisers are paying based on irr.  Unless the irr calculations are not based on purchases and clicks data but on less obviously related (fake) account impressions, I dont think the number of fake accounts matters to advertisers.  Maybe he addresses this, but curious...

 

That’s correct. If FB indeed has that many fake account, it doesn’t matter because the FB ad $ are driven by ROE ( people who click and buy) not the amount of accounts. In a way, more fake accounts are even bullish, since it means that the addressable market may be higher than thought. Right now the thesis is that everyone has already a FB / Instagram account in NA.

 

My bigger concern is the total addressable market.  FB and GOOG have already a significant part of the ad market, which tends to have a more or less fixed percentage of the GNP. Facebook needs to expand the TAM, in ads to Europe and Asia and ultimately beyond ads, if they want to continue growing.

 

Well a few issues with fake accounts. 

 

1) How much of an ad impression that is viewed is delivered to an audience member that isn't intended?  Not all of the ads on Facebook are performance marketing relying on click-throughs.  Some portion of the ads are about creating awareness or reinforcing a brand (CPMs or likes) in which case the advertiser thinking they are paying to deliver an ad impression to a particular user is not getting that benefit.  I'm not sure if these fake accounts take away from that.  In fact, I'm not even sure I understand the business model of fake accounts besides creating ads to influence elections etc.

 

2) Fake account traffic MAY be contributing to DAU/MAUs that aren't genuine and MAY be masking an underlying trend of engagement erosion.  It's possible but not sure. But my confidence in FB's self reported DAU/MAUs is going down actually.

 

3)  Are "fake accounts" or what i would call "nefarious advertisers" paying for politically motivated ads that may not be a source of recurring revenue?  If Facebook somewhat successful at stamping them out, how much of those ad dollars in prior years go away?

 

It's just food for thought.  I don't know the answers. I'm long the stock, but those are my concerns.  Incidentally click-fraud and impression fraud have been around the internet for a long time.  Advertisers aren't unaware of it, but they still use Google, FB others due to the high direct conversion rates and/or awareness that they deliver anyway.   

[cameronfen]

https://www.plainsite.org/realitycheck/fb.pdf

 

Did anyone see this?  really interesting read.  Seems very biased but worth understanding the negative side of the story.

 

I'll definitely read.  I dont understand though, whether facebook has 1 million fake accounts or 1 billion fake accounts, ultimately advertisers are paying based on irr.  Unless the irr calculations are not based on purchases and clicks data but on less obviously related (fake) account impressions, I dont think the number of fake accounts matters to advertisers.  Maybe he addresses this, but curious...

 

That’s correct. If FB indeed has that many fake account, it doesn’t matter because the FB ad $ are driven by ROE ( people who click and buy) not the amount of accounts. In a way, more fake accounts are even bullish, since it means that the addressable market may be higher than thought. Right now the thesis is that everyone has already a FB / Instagram account in NA.

 

My bigger concern is the total addressable market.  FB and GOOG have already a significant part of the ad market, which tends to have a more or less fixed percentage of the GNP. Facebook needs to expand the TAM, in ads to Europe and Asia and ultimately beyond ads, if they want to continue growing.

 

Well a few issues with fake accounts. 

 

1) How much of an ad impression that is viewed is delivered to an audience member that isn't intended?  Not all of the ads on Facebook are performance marketing relying on click-throughs.  Some portion of the ads are about creating awareness or reinforcing a brand (CPMs or likes) in which case the advertiser thinking they are paying to deliver an ad impression to a particular user is not getting that benefit.  I'm not sure if these fake accounts take away from that.  In fact, I'm not even sure I understand the business model of fake accounts besides creating ads to influence elections etc.

 

2) Fake account traffic MAY be contributing to DAU/MAUs that aren't genuine and MAY be masking an underlying trend of engagement erosion.  It's possible but not sure. But my confidence in FB's self reported DAU/MAUs is going down actually.

 

3)  Are "fake accounts" or what i would call "nefarious advertisers" paying for politically motivated ads that may not be a source of recurring revenue?  If Facebook somewhat successful at stamping them out, how much of those ad dollars in prior years go away?

 

It's just food for thought.  I don't know the answers. I'm long the stock, but those are my concerns.  Incidentally click-fraud and impression fraud have been around the internet for a long time.  Advertisers aren't unaware of it, but they still use Google, FB others due to the high direct conversion rates and/or awareness that they deliver anyway. 

 

My (relatively unresearched) thoughts:

 

1.) First as an advertiser, even if you are doing a CPM type add, you can see engagement statistics on Facebooks website.  If you aren't getting the kind of engagement that you are expecting, I think you would pull the ad.  Furthermore, people that advertise online are less likely to want to place these kinds of awareness campaigns (namely awareness campaigns where they aren't tracking statistics--some may advertise to increase likes on page but that is a very concrete objective), because, from a relative value standpoint, it makes more sense to place ads on TV, radio, and billboards because they don't offer the capability to track viewers.  Why pay up (even though you aren't actually really paying up--its really historical inertia), for the data aspect and the ability to quantify irr if you don't need it.  Even though I offer a counter-argument, I think this is an important point. 

 

2.) If the growth in accounts is mostly fake, I think it is unlikely that Facebook growth in accounts just happens to saturate when MAU is right about a level where everyone who uses the internet (outside of China) uses some facebook platform.  Maybe it's a happy coincidence, even so, it's about irr and facebook can't fake that to advertisers. 

 

3.) This is probably something, but just from my personal experience, I have never seen the suspicious ads for conspiracy theories or getting people to not vote etc.  Doesn't mean they don't exist.  I could be gullible or not the target audience, but considering every firm with a reasonably sized marketing budget is on FB, this number probably isn't going to put a serious dent in earnings. 

[johnny]

I continue to think the bigger risk to FB's ad revenue isn't the fake account load, but the very likely possibility that advertisers are in fact not evaluating their spend in typical IRR terms, but are largely in a "reach scale or die" venture-backed dynamic that is causing them to vastly overestimate the lifetime value of customers, and thus overbid on ads, regardless of what metric they're targeting.

 

This thing about getting teenagers to install dev certificates so FB can go through every bit on their devices is a pretty huge deal too. They know the PR risks here, so this story makes me lean slightly more towards the idea that there may be some fundamental business model concern that is driving FB to take these PR risks, rather than just engineer autism being unchecked.

 

 

 

 

 

[Spekulatius]

^ “ Salad oil scandal”, imo. You couldn’t tell all these negative headlines, if you just look at the business numbers. Customer/ users don’t care - user numbers are up and revenues are up 30% YoY. That’s what matters.

[johnny]

I think it merits concern even if it doesn't show up in the income statement. Their current network rents are resilient to the tide of public opinion, but public opinion will matter once they need to make a move (acquiring Instagram was, in retrospect, a massive company-saving maneuver. Could they do that today?).

 

And like I said, I'm more concerned about the -why- here. The salad oil scandal had a pretty obvious and uncontroversial incentive structure: some dudes did illegal fraud stuff because it was clearly in their financial interest to do fraud. My question is why should Facebook need to be so aggressive/creepy in its attempts to collect data on teens? The whole idea behind paying 15x earnings for this company is that you sort of assume they already are getting all of the data they need to maintain/extend their moat just from the normal use of their products.

 

It also matters that this very weird Apple/FB cold war that has been brewing for a few years seems to be getting much more violent. Those two companies can do a great deal of damage to one another if they decide to go full-blown K-Street on each other.

 

[JayGatsby]

Their ability to continue to increase ad revenue in US/Canada amidst flat users (and I'd argue decreasing screen time) is impressive. They've said in the past the can't really increase core ad load much more, so either they're monetizing new channels (stories, messenger, etc) or rates are going up.

 

Most venture companies I think track their spend/metrics fairly closely. That's the key to the raising the next round.