I really do think the number one feature that everyone really loves when they talk about the Facebook ads platform is the targeting parameters. That is the actual ways in which you can customize and tailor audiences to your specific business. Facebook is best in class for social media location targeting. That means that, really, you can find people exactly where you want them and send them an ad exactly when they're in a very specific place. This isn't limited to just targeting people in a zip code, like we will see with the Twitter ads platform. Instead, it's about creating fences in which advertisements are served. What I mean by that is, we can essentially put a little dot on the map, and that map can actually be the guiding parameter of whether we serve an ad or not. Let's go ahead and take a look here. If you can see here, we've placed a pin in downtown San Francisco. As a result, only people that are inside of this area that you see here in the shade of blue, actually see the advertisement. This is what we call hyperlocal in the digital ad world. That means that we can actually go beyond a specific area of the city, but instead, actually create an imaginary area in which we're serving our ads. This hyper precision really gives us the ability for us to advertise local businesses effectively. I want you to think about it. If you're a small business, what area of your town does your business own? What area does it want to own that it maybe doesn't already own? These are geofences that you can create in the Facebook ads manager. These geofences really give you the ability to only target the people inside of those areas, when they're in those areas, or people that regularly are in those areas. If you happen to have a business that's in a place where lots of people work in a common working area, you may not only want to target that area but you may want to target popular areas where people are commuting from. For instance, if I had a small business in Boulder that was on the hill, I might want to target common neighborhood communities that are associated with people that work at the University of Colorado, Boulder. So in my case, Lafayette or Louisville, places where lots of folks commute from to the University of Colorado Boulder. Again, all of these areas are easy to target with geofencing. Of course Facebook has the basic demographics down path as well. So it has really good data on your age, what your gender identity is, your relationship status, the languages you speak, the places you went to school, the major you had at school, they have the job that you work for, and past employers. This is a real key advantage of Facebook when compared to other platforms, compared for instance to Twitter. Twitter doesn't really give you the ability to put this stuff in your profile, Facebook does. So because most folks offer at least some of this information, and in fact I do believe that at least age and gender is required to sign up with the platform, you really do have the ability to not even have to guess at these things. One of the things that's really problematic about advertising on the Twitter platform is that age has to be inferenced. Even though the algorithms can be really good at guessing what someone's age is for an individual's Twitter profile, based on the stuff they post, the language they use, so on and so forth, it's still a guess and it's going to be wrong sometimes. Facebook really has the ability to have this standard data or this gold standard data which is people put it in, and because Facebook pages are generally based on people that are real, this data tends to be accurate because people don't generally lie about their age on the platform or if they do, they probably are somewhere in the ballpark of accurate. What really makes Facebook a powerhouse is the amount of data it collects on an individual. That's the most important thing to think about when you're thinking about the Facebook ad manager when compared to another ad platform. Facebook collects data in four different ways and each way, in and of itself, is really impressive. So the first way in which Facebook collects data is by just looking at your behavior on the platform. So the pages that you like when you're actually on the platform. Do you like certain bands? Do you like certain restaurants? Do you like certain other interests such as celebrities? Your behavior on the platform really speaks a lot about itself. So the stuff that you like is an indication that you actually have interest in it and that is one of the main inferences that Facebook uses. If you like a page, Facebook does generally think that you're interested in it. So beware. If you're just liking your friends band on Facebook, Facebook's going to take it that you actually not only like that band, but like the genre that band represents. Similarly, the actual post content that you make or the actual updates that you give Facebook that you want to share with your friends says a lot about you. It's easy to know whether someone's been married because it's easy to look through their photos and see that they have photos of weddings or a photo of themselves in a tux. So there's a lot of ways that Facebook actually learns about you and your life, not only through the textual content that you post, but through the videos that you post. One of the big initiatives of Facebook over the last few years has been building AI that can go through images and video and extract meaning about who you are and what you're doing. Other businesses also have the opportunity to upload their data to connect you on the platform to them. So essentially, businesses can upload lists of customers and say, "I'd like to reach these people on the Facebook platform." Along with that data, they can upload some stuff about your actual identity that the platform may have not known, such as an alternative e-mail address or a phone number. Facebook tracks the activity of its users after they leave the site. So web sites that they visit, games that they play, apps that they install can all be tracked as a part of this broader landscape. So Facebook knows the types of websites that you visit, it knows how often, it knows what actual product categories are associated with these websites. Facebook has a really good idea as a result, whether someone is looking for a mattress or not, whether someone is looking for a new computer or not, someone's looking for a cellphone or not, this all comes through browsing behavior. Especially for products that require that level of research before purchase, Facebook has a really good way of knowing whether you're interested in buying that product or not. It doesn't know exactly when you're going to pull the trigger, and we'll talk about that lack of intent at the end, but it definitely knows when people begin to start to research products that need to be researched before they're purchased. Let's imagine that I've moved to a new town and I'd like to figure out which coffee shop I went to frequent. Facebook isn't really going to know the initial search that begins this quest. So local coffee shops Boulder. That search really isn't a part of this advertising landscape data. However, Facebook will notice if a number of the websites that it tracks that are associated with coffee, begins to be visited by a particular user. So you could imagine if I visit the website bozocoffee.com and then I also visit a website such as Caribou Coffee, I am now all of a sudden looking up coffee to an extreme compared to my typical browsing behavior. This sends an indication to Facebook that I am interested in coffee now. While it may not mean that I actually want to use a coffee shop right at that very moment, it does mean that I'm interested in it. Facebook's really good at capturing these interests as they happen. Finally, Facebook actually has a really good way of approximating your location. So not just from the place that you put in a message or tag in a photo when you upload it to Facebook, there are other ways in which Facebook inferences location. The Wi-Fi connections that you use, the actual location of your phone are all really good indicators to Facebook exactly where you are. As a result, it's tracking, it's probably the most sophisticated of all social networks. So when you zoom out here and you look at all the different types of data that Facebook can collect on a consumer, it's kind of overwhelming. Everything from their demographics and geographics, to the friends they have and their interests, to the individuals interests, to specific types of data like mortgage data from banks, to all of or at least a majority of website traffic, where they actually go and visit online. When you take all of this stuff and you combine it together, you have a big pool of data on every individual user. This is data that you can action easily using the ads manager. There are even more inferred demos that we don't have a lot of time to talk about, but Facebook is always adding these specific types of lifestyle or life moment interests to their ad targeting. So if you want to reach people that have just recently become parents, that's possible using the Facebook ad manager. If you want to target people that are interested in a specific political viewpoint, you can do that with the Facebook ads manager. Of course, Facebook political ads are now protected as a result of the 2016 election. Hopefully, that means that these types of inferred demos aren't manipulated. Facebook also has ways of approximating ethnicity. Finally, Facebook has the ability to inference important life events, such as when you are married or your birthday and so on and so forth. So I think it's important to take a step back and just assess the quality of the targeting parameters that you'd like to use before you'd actually use them. So this isn't black and white, but from what I understand, interests are mainly based off of Facebook's internal data. So the stuff that you actually do on the platform. Therefore, we can probably be a little bit more sure that this stuff is accurate. Your interests or your Facebook account really is tailored to you whereas your computer or your web browser could be unrestricted and many people could use it. As a result, the interests in the behavior such as the websites you visit, could be off. So Facebook buys this secondary data in a number of ways as well such as the infamous mortgage example. So if you've just recently acquired a mortgage, Facebook could potentially have that data on you. These data sets really vary depending on the quality. For instance, Spotify is a data partner for Facebook and the data that it shares about your music interests is pretty high quality. As long as your account is attached to your Facebook account or your Spotify account is attached to your Facebook account, there really should be a one-to-one with little error. Other interests such as website behavior or behavior on specific apps that Facebook may acquire may not have that high quality standard of accuracy. So let's think about what Facebook is attempting to accomplish with interests. If we think of an industry affinity, this is really Facebook trying to find folks that are interested in specific types of products. Are you interested in retail shopping? Are you interested in gourmet dining? Facebook wants to know what types of content you're interested in and entertainment is a big type of content. So they want to try to match that up. What types of music are you interested in? What types of shows are you interested in? How can we show you more content that matches that? That's a part of their ad platform but also a part of their news feed algorithm in trying to show people content that is relevant as possible. Facebook is actually good at inferencing things like whether somebody is related to somebody else. It's actually not that hard to think about. If someone has the same last name, or if someone has a lot of friends in common and they talk regularly, that could suggest a family relationship or at least a deeply connected relationship. In general, Facebook wants to know what you're excited about, and those interests are really what it tries to leverage because it doesn't want to just show you ads that are irrelevant to you, it wants to show ads to people that are relevant. So this is a really big part of the innovation that's happening under the hood at Facebook. The Facebook behavior data is more myopic and that usually is indicating something specific happening in an individual. Again, imagine that I visited four.com, chevy.com, and honda.com and spent time on each page looking at different types of cars. If I did that, Facebook's going to use that in that combination of behaviors, as a trigger to say that I now am interested in buying a car. So I'm now what we say in market for buying a vehicle. Facebook's going to look at these types of behaviors across many different types of websites. So if I visit a charity website, for instance, and a combination of charity websites in a short period of time in particular, that really can suggest that I'm interested in making a charitable donation. Similarly, it's not hard for Facebook to understand if you work at a particular company or not, whether you put it on your Facebook page. If you regularly visit specific websites associated with one company, such as IBM, it shouldn't be hard for Facebook to make that inference that you actually work there. Facebook does have the ability to track different types of purchases and purchase behavior. Although there's not a lot of literature on this and it's really hard for me to understand what data they have and they don't, but at some level, Facebook does know the purchase behaviors of consumers. If you're interested in learning more about this, the best way is to actually go into the app platform itself and begin to play around with these behavioral parameters. You'll be able to see what you can target and what you can't. It won't always be intuitive as to what in market categories exist for what products or what don't, but in general, when there's more money to be made or there's a higher margin such as the sale of a car or the sale of a cellphone, these things tend to have more of these behaviors built from them because advertisers want them. Smaller products where the margins are smaller, there may not be a behavior that perfectly maps to the moment that someone is in market for your particular product or service.
