WEBVTT - The Facebook Algorithm Problem 0:00:04.400 --> 0:00:07.800 Welcome to tex Stuff, a production from I Heart Radio. 0:00:12.200 --> 0:00:15.040 Hey there, and welcome to tech Stuff. I'm your host, 0:00:15.280 --> 0:00:18.840 Jonathan Strickland. I'm an executive producer with I Heart Radio, 0:00:18.880 --> 0:00:22.360 and I love all things tech and I don't get 0:00:22.360 --> 0:00:25.720 to do this very often, so indulge me. But in 0:00:25.840 --> 0:00:33.080 Shakespeare's classic tragedy Hamlet, our protagonist and perpetually moody Dane 0:00:33.520 --> 0:00:37.760 says at one point there is nothing either good or bad, 0:00:38.040 --> 0:00:42.720 but thinking makes it so. What he means there is that, 0:00:42.880 --> 0:00:45.680 at least in his eyes, all things in the universe 0:00:46.120 --> 0:00:49.960 are just as they are. They are neither good nor bad. 0:00:50.440 --> 0:00:53.640 It's only how we think about them that makes them 0:00:53.840 --> 0:00:56.560 good or bad. This is similar to a philosophy and 0:00:56.560 --> 0:00:59.720 I've talked about on this show when it comes to technology. 0:01:00.280 --> 0:01:03.920 Technology and of itself has the capacity to help or 0:01:03.960 --> 0:01:08.480 to harm. But the underlying tech isn't necessarily good or evil. 0:01:08.840 --> 0:01:12.959 It's how it's put to use that makes all the difference. Now, granted, 0:01:13.319 --> 0:01:16.640 there are some technologies that are harder to use in 0:01:16.680 --> 0:01:19.360 a way that is a net positive, you know, like 0:01:19.640 --> 0:01:24.000 nuclear warheads for example. Not that people haven't tried, they have, 0:01:24.600 --> 0:01:28.360 but you get my meaning. Well, one technology that has 0:01:28.440 --> 0:01:32.000 far reaching effects and consequences that can end up being 0:01:32.040 --> 0:01:35.920 positive or negative. Would be the algorithms that are underlying 0:01:36.000 --> 0:01:40.280 stuff like search and social networks. Specifically today, I'm going 0:01:40.360 --> 0:01:44.760 to be talking about Facebook in this episode, but algorithms 0:01:44.840 --> 0:01:49.000 underlie the various computer systems and online systems that we 0:01:49.040 --> 0:01:51.840 rely upon, and they have the potential to be harmful 0:01:51.960 --> 0:01:56.360 depending on their implementation. But before I start talking about 0:01:56.400 --> 0:02:00.320 the effects of algorithms, let's back up a little bit 0:02:00.440 --> 0:02:06.000 and define that what exactly is an algorithm. Basically, an 0:02:06.040 --> 0:02:10.200 algorithm is a set of directions or rules, and that's 0:02:10.240 --> 0:02:15.560 really it. There's nothing special about this definition. These rules 0:02:15.639 --> 0:02:20.679 guide how computers calculate operations and solve problems. It's a 0:02:20.800 --> 0:02:25.400 finite sequence of instructions for computers to follow when they're 0:02:25.440 --> 0:02:30.520 executing a specific task. These instructions are unambiguous, which means 0:02:30.840 --> 0:02:34.000 the computer system doesn't have to make judgment calls or 0:02:34.000 --> 0:02:38.519 anything like that while executing the instructions. There is a 0:02:38.520 --> 0:02:42.480 clear path to follow based on whatever the situation is 0:02:42.560 --> 0:02:47.079 at hand. A different situation might lead to a different outcome, 0:02:47.240 --> 0:02:50.560 but the instruction path would still remain clear. If you 0:02:50.600 --> 0:02:53.320 were to dig down deeply enough to see what was 0:02:53.400 --> 0:02:57.600 going on behind the computer's actions. The number of potential 0:02:57.639 --> 0:03:01.400 paths to various outcomes could be an warmous and so 0:03:01.480 --> 0:03:04.360 to us it would look like a really complicated system 0:03:04.360 --> 0:03:08.320 and very difficult to grock. But if you really got granular, 0:03:08.800 --> 0:03:11.639 you would see the logic, as it were, that connects 0:03:11.680 --> 0:03:16.040 the beginning state to the end state of any given operation. 0:03:16.800 --> 0:03:21.200 At Google, search algorithms include elements that determine what position 0:03:21.240 --> 0:03:25.320 of particular query result will take on a results page. 0:03:25.960 --> 0:03:29.480 So when we search for something using an engine like Google, 0:03:29.960 --> 0:03:33.160 there are several things going on in that split second 0:03:33.240 --> 0:03:36.760 before our results pop up. For one thing, Google has 0:03:36.800 --> 0:03:40.080 an index of all the web pages it has crawled, 0:03:40.480 --> 0:03:43.880 and so Google will reference that index to find all 0:03:43.920 --> 0:03:47.160 the entries and that index that relate to the terms 0:03:47.160 --> 0:03:49.800 that you put in your search query. And then those 0:03:49.840 --> 0:03:53.560 results are ranked by a few different criteria. So way 0:03:53.680 --> 0:03:58.480 up there is relevancy, like how relevant is each search 0:03:58.560 --> 0:04:01.960 result to the original search query. If I were to 0:04:02.000 --> 0:04:05.600 search for Captain America, Google is going to put the 0:04:05.600 --> 0:04:08.800 pages about the fictional superhero way up towards the top 0:04:09.360 --> 0:04:12.320 search results about someone who happens to hold the rank 0:04:12.440 --> 0:04:16.120 of Captain and there's somewhere a mention of the region 0:04:16.160 --> 0:04:20.160 of America would likely fall much lower on the search 0:04:20.200 --> 0:04:24.520 results list, because even though those pages will have relevant 0:04:24.640 --> 0:04:27.880 terms in them, the chances are it's not what I'm 0:04:27.920 --> 0:04:31.919 actually looking for. A context is important, but context is 0:04:32.000 --> 0:04:36.120 just part of it. There are thousands of pages that 0:04:36.200 --> 0:04:40.719 mentioned Captain America. Some are from really good resources, meaning 0:04:40.920 --> 0:04:44.560 they are resources that are dependable, they're informative, they might 0:04:44.560 --> 0:04:48.720 be entertaining, you know whatever. Some are from garbage sites 0:04:48.760 --> 0:04:52.800 that don't provide any real valuable experience. So then you've 0:04:52.839 --> 0:04:56.039 got the ranking system within Google that determines how the 0:04:56.120 --> 0:05:01.040 most relevant search engine results will appear on the results age. Now, 0:05:01.040 --> 0:05:04.000 if there are thousands or even millions of hits, how 0:05:04.000 --> 0:05:06.880 does Google determine which ones show up on the first 0:05:06.880 --> 0:05:10.520 page of results. Well, there's a complicated system that determines 0:05:10.680 --> 0:05:13.359 that ranking, and it's not really in the realm of 0:05:13.400 --> 0:05:16.200 this episode to go into it, but suffice it to 0:05:16.240 --> 0:05:19.120 say that Google takes a lot of things into consideration 0:05:19.320 --> 0:05:22.920 when ordering search results. Ostensibly, this is all in an 0:05:22.960 --> 0:05:25.840 effort to give you the most relevant and helpful results 0:05:26.080 --> 0:05:29.200 based on your search query. For the people companies and 0:05:29.320 --> 0:05:32.680 organizations that are making pages. It's always a goal to 0:05:32.880 --> 0:05:37.000 rank really high on that list, preferably on that first page, 0:05:37.000 --> 0:05:39.719 and if you can, the top half of that first page, 0:05:39.760 --> 0:05:42.960 because most people will only look at the first few 0:05:43.040 --> 0:05:46.320 search results and then they'll ignore everything that comes afterward. 0:05:46.920 --> 0:05:49.839 Over the years, Google has made it really tricky to 0:05:49.960 --> 0:05:53.480 game this engine, which means it's hard to come up 0:05:53.520 --> 0:05:57.440 with effective strategies to show up higher in search results. 0:05:57.920 --> 0:06:01.000 Your best bet is really to create just really good 0:06:01.040 --> 0:06:05.240 resources around the particular topic of interest in question. And also, 0:06:05.400 --> 0:06:08.440 you know, happened to be a really respected resource, which 0:06:08.520 --> 0:06:12.560 is not easy to do, especially if you're just starting out. Obviously, Also, 0:06:12.600 --> 0:06:17.120 whenever Google makes changes to this algorithm, it can drastically 0:06:17.120 --> 0:06:20.719 affect how your page appears in search results, even with 0:06:20.760 --> 0:06:24.240 a perfect query. And when tech stuff was still part 0:06:24.240 --> 0:06:27.200 of how stuff works, that was something we were always 0:06:27.240 --> 0:06:31.040 concerned about. A change in the Google algorithm would mean 0:06:31.080 --> 0:06:35.000 that our web traffic could either skyrocket or it could plummet. 0:06:35.560 --> 0:06:37.760 There was no change in the quality of our work. 0:06:38.000 --> 0:06:40.320 Our work was just as good as it had been before. 0:06:41.000 --> 0:06:43.480 It would just be a change in how our work 0:06:43.480 --> 0:06:47.760 would show up in search results. However, that's the Google algorithm. 0:06:47.800 --> 0:06:51.600 With other algorithms like the ones on Facebook, those can 0:06:51.640 --> 0:06:54.480 be easier to game if you know what you're doing. 0:06:54.839 --> 0:06:58.000 And what's more, there are algorithms on Facebook that allow 0:06:58.080 --> 0:07:03.680 for advertisers to engage in some discriminatory and potentially criminal behavior. 0:07:03.800 --> 0:07:07.280 So we'll look at how the various algorithms on Facebook 0:07:07.560 --> 0:07:10.560 can have a negative impact because of the way they 0:07:10.560 --> 0:07:13.800 are implemented and because of the way that people are 0:07:13.840 --> 0:07:17.520 taking advantage of them. Now, in the beginning, Facebook didn't 0:07:17.560 --> 0:07:21.040 have an algorithm at all. Heck, it didn't even have 0:07:21.080 --> 0:07:24.320 a news feed when it first started. Instead, you could 0:07:24.360 --> 0:07:27.360 post stuff to your profile, but then you could visit 0:07:27.400 --> 0:07:30.200 the profiles of friends to see what they had posted. 0:07:30.680 --> 0:07:34.320 You could leave comments on stuff, but it still required 0:07:34.360 --> 0:07:38.560 you to actually go over to your friends profile. There 0:07:38.640 --> 0:07:43.520 was no aggregator for content just yet that changed in 0:07:43.600 --> 0:07:48.320 two thousand seven when Facebook introduced the news feed. Now, 0:07:48.480 --> 0:07:52.320 your homepage on Facebook had a record of your friends activities, 0:07:52.360 --> 0:07:55.720 just listed in reverse chronological order. You know, the most 0:07:55.800 --> 0:07:58.920 recent things appearing at the top of the list. You 0:07:58.920 --> 0:08:01.280 could log in and see what your friends had been 0:08:01.360 --> 0:08:03.440 up to since the last time you were on and 0:08:03.480 --> 0:08:06.800 you could catch up. There's still no algorithm to speak 0:08:06.840 --> 0:08:10.080 of yet, and as would become a hallmark of any 0:08:10.280 --> 0:08:14.560 change made on Facebook. Ever, a lot of users objected 0:08:14.760 --> 0:08:17.040 to the news feed, and the main objection was that 0:08:17.440 --> 0:08:20.640 people felt it was a violation of privacy. They didn't 0:08:20.680 --> 0:08:23.800 like the idea that the activities they were doing on 0:08:23.800 --> 0:08:28.400 Facebook were being broadcast to all of their friends, which okay, 0:08:28.400 --> 0:08:30.880 now this one stumps me a little bit, because the 0:08:30.920 --> 0:08:35.840 profiles on Facebook were public unless you set privacy features 0:08:35.880 --> 0:08:39.520 to make it otherwise. So if you change your relationship 0:08:39.600 --> 0:08:43.360 status on your profile, for example, why would you object 0:08:43.360 --> 0:08:47.640 to it being broadcast to your friends. If anyone visited 0:08:47.640 --> 0:08:50.440 your profile, they would see the difference. So either you 0:08:50.600 --> 0:08:53.360 don't change it, or you change the privacy settings on 0:08:53.400 --> 0:08:56.280 your profile, But otherwise why would you be upset? I 0:08:56.320 --> 0:08:59.800 think the main takeaway is just that people hate change, 0:09:00.160 --> 0:09:02.920 and if you don't want something broadcast to everybody, maybe 0:09:02.960 --> 0:09:07.120 you shouldn't post it. A couple of years later, Facebook 0:09:07.160 --> 0:09:11.240 introduced the first iteration of their algorithm, which would determine 0:09:11.360 --> 0:09:14.320 what users would see in their news feed and in 0:09:14.440 --> 0:09:17.560 what order they would appear, so rather than getting a 0:09:17.640 --> 0:09:21.200 full account, of all of your friends activities in reverse 0:09:21.280 --> 0:09:25.000 chronological order. Users would see a selection of stuff that 0:09:25.040 --> 0:09:29.120 the algorithm had determined was the best for that user. 0:09:29.640 --> 0:09:35.360 So how did that work? Well, generally speaking, the algorithm 0:09:35.360 --> 0:09:39.560 took three main criteria into account to select the stuff 0:09:39.760 --> 0:09:43.760 that would show up on news feeds. One was user affinity, 0:09:44.080 --> 0:09:46.960 which was an indicator of how close the relationship is 0:09:47.520 --> 0:09:51.920 between one Facebook user and another, or a Facebook user 0:09:52.040 --> 0:09:55.960 and a brand. And this was measured in one way perspectives, 0:09:56.559 --> 0:09:59.400 and that can be a little confusing. So let's make 0:09:59.520 --> 0:10:03.360 a an example. We'll take three fictional people to explain 0:10:03.480 --> 0:10:08.240 how this affinity worked. So in this example, we've got Sam, Dean, 0:10:08.760 --> 0:10:14.080 and Castiel. Okay, So Sam and Dean are super close, right, 0:10:14.640 --> 0:10:18.160 and they often comment on each other's posts, and they'll 0:10:18.240 --> 0:10:21.000 hit like on each other's statuses and all that kind 0:10:21.040 --> 0:10:24.440 of stuff. So they tend to see one another's posts 0:10:24.520 --> 0:10:26.960 a lot in their own news feeds when they log 0:10:27.000 --> 0:10:31.400 onto Facebook. They each have a high affinity for each other. 0:10:32.040 --> 0:10:35.400 Then we've got Castil. So Castiel is friends with both 0:10:35.559 --> 0:10:39.720 Sam and Dean, but really he's much closer to Dean 0:10:39.880 --> 0:10:43.320 than he is to Sam. And Castiole responds to a 0:10:43.320 --> 0:10:46.720 lot of Dean's posts, and as a result, Castiel C's 0:10:46.840 --> 0:10:51.400 Dean's stuff pop up in his that is, Castil's news 0:10:51.400 --> 0:10:55.160 feed pretty regularly. Castil is not as close with Sam 0:10:55.200 --> 0:10:58.360 and he doesn't interact with Sam as much. So sam 0:10:58.440 --> 0:11:01.760 stuff can pop up in Castile's news feed, but it's 0:11:01.800 --> 0:11:04.280 not as frequent or as sure a thing as when 0:11:04.360 --> 0:11:08.800 Dean posts something. Now, on top of that, Dean likes Castiel, 0:11:09.080 --> 0:11:12.000 but you get the feeling that this relationship is not 0:11:12.080 --> 0:11:16.840 a perfect two way street. So Dean only interacts with 0:11:17.120 --> 0:11:21.240 some of Castile's stuff that he posts online. So Dean 0:11:21.559 --> 0:11:26.040 sees Castiel's posts on occasion in his news feed, but 0:11:26.360 --> 0:11:29.680 not nearly at the same frequency at which he sees 0:11:30.160 --> 0:11:34.160 Sam's post in his that is Dean's news feed. And yes, 0:11:34.920 --> 0:11:38.199 I just summarized the show Supernatural in an example about 0:11:38.320 --> 0:11:42.480 Facebook algorithms and affinity. Now, as we see a high 0:11:42.520 --> 0:11:46.679 affinity one person has for another helps determine how much 0:11:47.040 --> 0:11:50.040 of that other person's stuff the first person is going 0:11:50.120 --> 0:11:52.160 to see. But it only works in that one way. 0:11:52.480 --> 0:11:56.200 If I like your posts a lot, then I'm going 0:11:56.280 --> 0:11:58.600 to see more of your posts in my feed. But 0:11:58.640 --> 0:12:01.360 if you never interact with me, You're not going to 0:12:01.440 --> 0:12:03.800 see a lot of my stuff in your feed. That's 0:12:03.800 --> 0:12:06.280 what I mean by that one way perspective. Of course, 0:12:06.679 --> 0:12:10.240 there are other ways to wait this, more like you 0:12:10.280 --> 0:12:13.439 can tag someone in a post, which will definitely give 0:12:13.480 --> 0:12:16.839 their attention much more effectively. But let's move on. So 0:12:16.880 --> 0:12:20.360 the second component to the edge rank factor we had 0:12:20.600 --> 0:12:24.280 user affinity as a component one. Component two is the 0:12:24.320 --> 0:12:28.319 weight of the post W E, I, G, H T. 0:12:29.200 --> 0:12:33.040 So different kinds of posts have different weights, and the 0:12:33.080 --> 0:12:36.240 greater the weight, the more likely it will pop up 0:12:36.600 --> 0:12:41.360 on a given friends feed. So, generally speaking, with edge rank, 0:12:41.559 --> 0:12:45.520 stuff that had photos or videos had the most weight, 0:12:46.040 --> 0:12:52.000 followed by posts that contained links, followed by plain text posts. Now, 0:12:52.040 --> 0:12:55.640 the reason for this is that Facebook analysts had figured 0:12:55.679 --> 0:12:59.240 out that posts that had stuff like photos and videos 0:12:59.640 --> 0:13:03.120 with tend to get more engagement. I'll get back to 0:13:03.240 --> 0:13:06.520 engagement a bit later in this episode. Just know that 0:13:06.600 --> 0:13:11.440 it is a critical concept for Facebook. The third component 0:13:11.760 --> 0:13:16.200 factored into edge rink is called time decay, which essentially 0:13:16.200 --> 0:13:20.720 translates to how old is this post? And the idea 0:13:20.800 --> 0:13:24.400 here is that the more recent posts are likely more 0:13:24.480 --> 0:13:28.720 relevant than older posts are it doesn't do you any 0:13:28.760 --> 0:13:31.960 good to see the announcement of an upcoming get together 0:13:32.600 --> 0:13:36.640 if in fact that get together actually happened last week, 0:13:36.800 --> 0:13:41.000 so you're not likely to see the initial post pop 0:13:41.080 --> 0:13:42.679 up at the top of your news feed if you've 0:13:42.720 --> 0:13:47.200 already missed everything. As the post ages, the value of 0:13:47.240 --> 0:13:51.560 the post decreases. So when you log into Facebook or 0:13:51.760 --> 0:13:55.520 open the Facebook app or whatever, this algorithm would run 0:13:55.679 --> 0:13:58.680 with your particular situation in mind. It would scan through 0:13:58.760 --> 0:14:02.160 all the stuff that had been posted by the friends 0:14:02.200 --> 0:14:05.960 you have and the brands you follow since your last session, 0:14:06.280 --> 0:14:10.640 and then it would select and rank or order those 0:14:10.679 --> 0:14:13.880 status updates that you would see in your news feed 0:14:14.120 --> 0:14:18.440 based off of those criteria. But this would also mean 0:14:18.880 --> 0:14:22.360 you would miss out on stuff. No longer would you 0:14:22.360 --> 0:14:25.680 be able to see everything that all of your friends 0:14:25.720 --> 0:14:30.280 were doing, listed in reverse chronological order. In fact, at times, 0:14:30.360 --> 0:14:33.520 chronology didn't have anything to do with the order in 0:14:33.520 --> 0:14:36.960 which status updates would appear on your news feed. A 0:14:37.040 --> 0:14:41.160 post several hours old might appear above one that was 0:14:41.200 --> 0:14:44.600 posted just five minutes before you logged in, all because 0:14:44.720 --> 0:14:49.080 of those other factors. For some Facebook users, including me 0:14:49.360 --> 0:14:53.040 at the time, this was really frustrating. It made us 0:14:53.040 --> 0:14:55.880 feel as though we were missing out on stuff our 0:14:55.920 --> 0:14:59.680 friends were saying because we were. I found it particularly 0:14:59.720 --> 0:15:04.240 irre tating because I wasn't one to frequently engage with updates. 0:15:04.240 --> 0:15:07.440 I didn't like a lot of them, and I wouldn't comment. 0:15:07.480 --> 0:15:10.280 A ton of times. I would read everything, but I 0:15:10.320 --> 0:15:13.240 wasn't interacting with it, So the algorithm wasn't getting a 0:15:13.280 --> 0:15:17.880 really good idea of who my good friends were versus 0:15:17.960 --> 0:15:21.880 like people who are my acquaintances. So it meant that 0:15:21.960 --> 0:15:26.520 some people who were my closest friends in reality were 0:15:26.560 --> 0:15:29.640 not the ones that I was seeing, you know, being 0:15:29.680 --> 0:15:32.240 active on Facebook. Even if they were, I just wasn't 0:15:32.280 --> 0:15:35.840 seeing those posts. And my case was not unusual. A 0:15:35.840 --> 0:15:38.040 lot of people went through this and and then I 0:15:38.120 --> 0:15:42.520 missed stuff, sometimes big stuff like an engagement announcement or something. 0:15:43.520 --> 0:15:47.640 Edge rank was just the beginning, and actually that particular 0:15:47.680 --> 0:15:51.240 algorithm didn't stick around that long in the grand scheme 0:15:51.240 --> 0:15:55.760 of things. By late within Facebook, no one was really 0:15:55.840 --> 0:15:59.160 using the term edge rank anymore because they had moved 0:15:59.160 --> 0:16:04.080 towards a more sophisticated machine learning based algorithm. This one 0:16:04.160 --> 0:16:06.600 was far more sophisticated than edge rank. It was not 0:16:06.720 --> 0:16:09.760 limited to just those three main factors, although they were 0:16:09.760 --> 0:16:14.720 still components within the algorithm. Instead, according to Lar's Backstrom, 0:16:14.720 --> 0:16:17.920 who was the engineering manager for the news feed ranking 0:16:17.960 --> 0:16:20.960 at the time, this new one used somewhere in the 0:16:20.960 --> 0:16:25.800 neighborhood of one hundred thousand weights to determine what users 0:16:25.800 --> 0:16:28.600 would see and in which order. Now you would think 0:16:28.760 --> 0:16:30.680 that with that many weights it would be hard to 0:16:30.720 --> 0:16:34.160 gain the system, but you'd be wrong. Now, a couple 0:16:34.240 --> 0:16:36.320 other things I want to mention before we go to break. 0:16:36.920 --> 0:16:39.600 One of those is that, yes, there was an option 0:16:39.720 --> 0:16:44.080 still is to try and list things in most recent 0:16:44.240 --> 0:16:48.040 as opposed to what is most relevant or whatever Facebook 0:16:48.040 --> 0:16:51.480 thinks is most relevant, except that you're not really seeing 0:16:51.600 --> 0:16:54.880 everything that way. Either. You're not getting the most recent 0:16:55.440 --> 0:16:59.520 reverse chronological list of all of your friends activities. It's 0:16:59.560 --> 0:17:03.120 still a selection of what your friends are posting, and 0:17:03.120 --> 0:17:05.720 it's a selection that the algorithm has determined is the 0:17:05.760 --> 0:17:08.959 most relevant for you. It's just that they'll be listed 0:17:09.040 --> 0:17:12.440 in reverse order instead of kind of that haphazard all 0:17:12.440 --> 0:17:15.200 over the place method. But the fact is you're still 0:17:15.240 --> 0:17:18.160 not seeing everything, which still frustrates a lot of people. 0:17:18.840 --> 0:17:22.720 And when we come back we'll talk about the algorithm, 0:17:23.080 --> 0:17:27.840 how people have used the creative approach to game the system, 0:17:27.880 --> 0:17:32.439 and some unintended consequences that algorithm has led to. But 0:17:32.560 --> 0:17:42.600 first let's take a quick break. So we're at the 0:17:42.680 --> 0:17:47.000 point where an algorithm starts to serve up content catered 0:17:47.119 --> 0:17:51.920 to each individual Facebook user, and the goal was always 0:17:51.920 --> 0:17:56.280 the same from the very beginning keep people engaged, which 0:17:56.320 --> 0:18:00.800 really just means keep people on Facebook for longer. And 0:18:00.840 --> 0:18:05.719 this gets to Facebook's method of generating revenue. Facebook's primary 0:18:05.760 --> 0:18:10.120 source of revenue is through advertising. And when I say primary, 0:18:10.160 --> 0:18:16.960 according to Investopedia, advertising accounts for ent of Facebook's revenue, 0:18:17.359 --> 0:18:25.399 which for Q three was twenty one point five billion dollars. 0:18:25.840 --> 0:18:31.280 Wholly cal princely some doesn't even begin to cover it. 0:18:31.920 --> 0:18:35.320 Facebook sales ad space to various companies to serve up 0:18:35.320 --> 0:18:39.080 advertising to Facebook's users, so they're essentially think of it 0:18:39.200 --> 0:18:42.159 like a billboard. Facebook is like a giant billboard, but 0:18:42.160 --> 0:18:46.080 it's a very special billboard. It's incredibly valuable. There's another 0:18:46.119 --> 0:18:48.760 component to this I'll touch on later, which revolves around 0:18:48.760 --> 0:18:53.399 targeted advertising, but let's just stick with the basic value 0:18:53.400 --> 0:18:57.959 proposition here. So Facebook has space on its apps and 0:18:58.119 --> 0:19:00.879 its pages, and in that space ace it has the 0:19:01.040 --> 0:19:04.639 ability to show ads that they could be on the side, 0:19:04.680 --> 0:19:06.880 they can be fed into the news feed as well, 0:19:06.920 --> 0:19:10.120 so they're kind of interspersed along with the status updates 0:19:10.119 --> 0:19:14.280 from friends and stuff. Uh. The longer that people use 0:19:14.359 --> 0:19:18.480 Facebook per any given session, the more ads they will 0:19:18.600 --> 0:19:21.879 encounter as they interact with Facebook, which also means that 0:19:21.920 --> 0:19:25.520 they could potentially interact with the ads. So part of 0:19:25.520 --> 0:19:28.840 what Facebook really needs to do is to find ways 0:19:28.880 --> 0:19:32.800 to convince people to spend as much time as possible 0:19:33.359 --> 0:19:39.080 on Facebook. In this use case, time really is money. 0:19:39.320 --> 0:19:42.639 Greater engagement doesn't just mean more opportunities to see ads, 0:19:42.920 --> 0:19:45.639 it also means that Facebook can demonstrate its own value 0:19:45.680 --> 0:19:50.159 to advertisers and thus command higher prices. To put it 0:19:50.240 --> 0:19:54.320 another way, let's say you own a few billboards around town. 0:19:54.359 --> 0:19:57.080 Since I've already mentioned them, so let's say you have 0:19:57.119 --> 0:19:59.879 a billboard that is on a back road and that 0:20:00.359 --> 0:20:03.240 gets just a little bit of traffic every day, and 0:20:03.320 --> 0:20:06.520 you happen to own another billboard space which is on 0:20:06.560 --> 0:20:09.120 a much more heavily used road that goes right into 0:20:09.160 --> 0:20:11.600 the heart of town. The rates that you would be 0:20:11.640 --> 0:20:15.560 able to charge for the back road billboard are probably 0:20:15.560 --> 0:20:18.200 gonna be a lot lower than what you can command 0:20:18.240 --> 0:20:21.840 with the heavily traveled thoroughfare. More folks are going to 0:20:21.920 --> 0:20:25.919 see that second billboard, which makes it more valuable. On 0:20:25.960 --> 0:20:29.600 a similar front, if Facebook can show numbers that indicate 0:20:30.040 --> 0:20:33.399 that more than two billion people are spending more and 0:20:33.440 --> 0:20:38.000 more time on Facebook, it can command higher prices for advertising. 0:20:38.320 --> 0:20:41.199 Add companies and their clients want to be where the 0:20:41.240 --> 0:20:45.040 eyeballs are, and Facebook fits the bill nicely. So with 0:20:45.200 --> 0:20:48.440 that in mind, the real purpose of the algorithm ultimately 0:20:48.680 --> 0:20:51.240 is to serve up stuff to users that will keep 0:20:51.320 --> 0:20:54.600 them on Facebook longer and convince them to come back 0:20:54.640 --> 0:20:58.640 to Facebook frequently. Now, some of the things that will 0:20:58.680 --> 0:21:01.040 bring people back fall online with a few of the 0:21:01.119 --> 0:21:05.520 ideas that Drank demonstrated. If you logged onto Facebook but 0:21:05.680 --> 0:21:08.240 you only ever saw posts from people you only kind 0:21:08.280 --> 0:21:11.000 of sort of know, and you always missed out on 0:21:11.000 --> 0:21:14.000 the posts from your close friends, you probably not find 0:21:14.080 --> 0:21:17.280 it a very satisfying experience. So at least some of 0:21:17.320 --> 0:21:19.479 the stuff you'd be seeing would fall in line with 0:21:19.600 --> 0:21:22.200 the stuff you'd really want to see in the first place, 0:21:22.720 --> 0:21:25.359 But maybe there would be some posts you would really 0:21:25.400 --> 0:21:28.800 like to see, but the algorithm, for whatever reason, doesn't 0:21:28.880 --> 0:21:32.760 judge those as being particularly engaging. Well, the chances are 0:21:32.920 --> 0:21:36.240 more slim that you would see those posts. If you've 0:21:36.280 --> 0:21:39.640 ever posted something to Facebook that you thought was particularly 0:21:39.680 --> 0:21:44.399 important or impactful, but you didn't see very much reaction 0:21:44.520 --> 0:21:47.360 to it, there's a chance that not many people actually 0:21:47.440 --> 0:21:51.840 saw what you posted, So that's a real possibility. Of course, 0:21:52.040 --> 0:21:54.240 there's also a chance that they did see it, they 0:21:54.280 --> 0:21:56.840 just didn't care. It's kind of hard to say for 0:21:56.880 --> 0:22:00.639 any given post, but it's entirely possible for you to 0:22:00.680 --> 0:22:04.280 post something that people would have reacted to. They just 0:22:04.320 --> 0:22:06.240 never saw it, so they never got the chance to. 0:22:06.800 --> 0:22:10.120 But let's say someone in your circle posts something and 0:22:10.240 --> 0:22:13.359 it starts to catch on. Maybe it's someone in your 0:22:13.400 --> 0:22:16.159 circle that you don't interact with much, so usually you 0:22:16.200 --> 0:22:19.879 don't see very much from them on Facebook. But for 0:22:19.920 --> 0:22:23.560 whatever reason, this particular post that they made is getting 0:22:23.560 --> 0:22:26.520 a lot of interaction from other friends, which means the 0:22:26.520 --> 0:22:29.920 algorithm might say, hey, this post is really taking off 0:22:30.359 --> 0:22:32.960 a lot of people are interacting with it, so I'm 0:22:33.000 --> 0:22:36.440 going to serve it up to even more people, including you, 0:22:36.920 --> 0:22:41.600 who usually wouldn't see such things. From this particular distant friend. So, 0:22:41.600 --> 0:22:45.040 in other words, posts that get a lot of engagement, 0:22:45.240 --> 0:22:49.080 that being reactions and comments and shares and that kind 0:22:49.119 --> 0:22:53.919 of stuff, they tend to rise above other posts. Engagement 0:22:54.040 --> 0:22:57.879 doesn't have to be positive either. It's not really important 0:22:57.920 --> 0:23:01.560 to Facebook if the stuff you see makes you happy 0:23:01.680 --> 0:23:04.679 or unhappy, or if the stuff you see is true 0:23:04.880 --> 0:23:08.600 or if it's false. What matters to Facebook is that 0:23:08.680 --> 0:23:12.199 you spend as much time on the platform as they 0:23:12.200 --> 0:23:14.920 can coax from you, and that you feel the need 0:23:14.960 --> 0:23:19.520 to return to Facebook frequently. Maybe it turns out you 0:23:19.640 --> 0:23:22.720 rage posts on Facebook a lot. Maybe you see stuff 0:23:22.760 --> 0:23:27.919 that you vehemently disagree with, so you post lengthy comments, 0:23:28.560 --> 0:23:30.840 or you share a post you don't agree with, but 0:23:30.960 --> 0:23:32.840 you do so in an effort to put your own 0:23:32.880 --> 0:23:36.080 spin on the thesis of the original post, or to 0:23:36.160 --> 0:23:40.680 criticize it in some way. Your motivations or goals are 0:23:40.800 --> 0:23:44.840 not important to Facebook. The company could not care less. 0:23:45.280 --> 0:23:49.320 The platform just once your time. Well, that's being a 0:23:49.320 --> 0:23:53.320 bit too reductive. Facebook also wants your posts to inspire 0:23:53.400 --> 0:23:57.119 other people to engage using their own comments, likes, and shares, 0:23:57.440 --> 0:24:00.080 so that they get their time too. That's really it. 0:24:00.359 --> 0:24:04.320 They want everyone's time, not just yours. I'm sure most 0:24:04.359 --> 0:24:07.360 of you already see where this becomes an issue. As 0:24:07.359 --> 0:24:10.439 we've heard for a few years now, Facebook has become 0:24:10.480 --> 0:24:15.879 a hotbed for misinformation and disinformation, something that Facebook founder 0:24:15.920 --> 0:24:20.600 Mark Zuckerberg has tried to avoid addressing for years. There's 0:24:20.600 --> 0:24:25.080 a saying that's frequently misattributed to Mark Twain that says 0:24:25.440 --> 0:24:28.679 a lie can travel halfway around the world while the 0:24:28.680 --> 0:24:31.800 truth is still putting on its shoes. I think it's 0:24:31.880 --> 0:24:34.560 kind of fitting that Twain is often cited as the 0:24:34.560 --> 0:24:37.040 originator of the saying, but the truth of the matter 0:24:37.119 --> 0:24:39.480 is that it predates him. It might have been the 0:24:39.480 --> 0:24:42.960 product of Jonathan Swift, but we don't really know. But 0:24:43.359 --> 0:24:47.520 the saying is apt. Misinformation is not saddled with the 0:24:47.640 --> 0:24:53.480 heavy burden of being, you know. True misinformation can tap 0:24:53.520 --> 0:24:59.560 into biases and preconceptions. It can reinforce the beliefs we hold, 0:25:00.160 --> 0:25:03.199 whether they're positive or negative, even if those beliefs have 0:25:03.440 --> 0:25:08.080 no real basis in reality. And if we feel that 0:25:08.119 --> 0:25:12.720 our beliefs are validated, we're more likely to engage and 0:25:12.840 --> 0:25:16.119 share those posts and to feel reinforced in it to 0:25:16.160 --> 0:25:19.800 do more, to continue sharing that kind of stuff. Now. 0:25:19.840 --> 0:25:24.719 I've seen examples of this across different ideologies. The narrative 0:25:24.800 --> 0:25:28.000