So let's do a little exercise here and try to unpack why certain ads are served and why. In doing so we'll learn a little bit about some common targeting practices on the platform. So this first ad here is for Flonase Sensimist and it is a specific product that is designed to fix a specific ailment, this case allergies. How do you think this ad was served to me? By the way, all these ads were served to me. This ad was served to me because in my bio it says I am a doctor. Well, I think you know enough about me now to know that I'm not the kind of doctor that can prescribe Flonase. I really want you to be critical of the approach that they used here. Just because someone has an indication in their bio that they have a PhD or a doctorate doesn't make them a medical doctor. So be really careful when you consider using heuristics such as what someone identifies themself as, there can be false positives. Now, let's consider this post. Again, you don't know a ton about me, but I already have all of the degrees that I probably am going to get my lifetime because I am a professor and I don't really have a clear need for an additional degree. Yet this ad was served to me. Why do you think that is? Again, if we go back to my bio we know that I have an interest in marketing. In fact, the word marketing is listed in my profile. So if I had to bet, this ad was targeted around folks that are interested in marketing. Whether it's the marketing buzz word in my actual profile or the content of my tweets, which also talk about marketing or the people that I follow. There's a lot of indication that I love marketing but there's one piece of evidence here that makes this add irrelevant. And again, that's because I already have all the degrees I want. Professors generally aren't seeking advanced degrees. In fact, anyone that teaches at a university probably is satisfied with the number of degrees they have. So this is an example of a Twitter ad limitation that just wasn't thought of. This is something they should have thought of when they were building the targeting parameters. They should have excluded people that work at universities or have titles such as instructor, professor and so on. Unfortunately, this is a very common mistake and most of the ads that I see on Twitter are actually for graduate programs. Again, I'm not the ideal market. So remember that every parameter that you use can have false positives. So let's think about this exercise critically. What can I do to better target perspective graduate students or perspective masters of science students? One, I could exclude titles that I know won't be interested. Very senior folks in business are unlikely to go back and get a graduate degree. So titles such as vice president or president or CEO or so on, CMO, would be excluded from my targeting list. So just as I can add behaviors in that I want to target, I can add behaviors in that I want to avoid. I don't want senior folks in business. Adding this would really help hone in that. More specifically to the use case of why I saw the ad we could similarly exclude folks that are instructors, that are professors, or have an affiliation with like a university currently. You can't go and get a graduate degree if you're already in school. You may say, well, seniors and undergrad could be relevant and I would agree. So even more parameters need to be built around these specific use cases. So think about it in those ways. One of the common mistakes that I see is this idea of only adding parameters that add and not adding parameters that honed and subtract. Just as you want to build a broad audience to make sure that your ad is seen as by as many people as possible. You want to try to build subtractive parameters to hone and tighten and increase the relevance of ads. In fact, I have to say if you don't do this most consumers that find ads irrelevant will be annoyed. I will point out that there is a level of irony that MBA programs are serving ads in inaccurate ways, right? So now let's think about another ad that I saw that was irrelevant. Here's an example of an ad that was served by the Marine Corps. And again, it suffers from a lack of exclusion. Professors are unlikely to go into armed service. I'm not making judgment statements here, I'm just going off of the data. Twitter actually gives you the ability to look at why an add was served. And if you'd ever get an ad that was served to you that you think was off, you can always click the little why was this ad serve button at the bottom of an add to unpack then? Well, I encourage you to do that because it helps you learn more about targeting parameters, specifically ones that don't tend to work well. And you can see here that the targeting parameters were just simply too broad. Only targeting based off of age and location is likely to garner a lot of people but it's really not having the relevance that we would want. And as a result, the metrics are going to suffer. If I cared to venture this may be because the Marines are unable to target on specific parameters for discrimination reasons or other reasons, but as it stands this digital advertisement is too broad. Now, let's consider this ad for Project Fi. For those of you that don't know, Project Fi is a rising cell phone service provided by Google that rivals cheaper plans such as T-Mobile and Sprint. Why do you think this ad was served to me? If we click the expand option, we can see that it was targeted to a age range and a specific location. So this is Denver Colorado. And if I had to guess, I would assume that they chose Denver because one, Google Fi is available in that area. And two, because of the tech leaning focus that the broader metro area tends to have. So in some ways this ad is actually spot on, I'm actually very interested in Google Fi so interested that I'm already a customer. So what can I do to prevent current customers from seeing ads? If your company is collecting data on its customers, that is collecting email addresses when they check out at your point of service via Square or via some other type of data collection service or a list of email subscribers or even a list of customers as they are registered for your product or service people that have registered for your eStore so on and so forth. Those people can actually be entered into the Twitter ads ecosystem and excluded from the ads that you serve. So simply put Google ads folks could take lists of customers that already subscribe in the Denver area and exclude them from their ads making sure that existing customers in Denver don't see the ads. This is a more advanced tactic, but it's certainly one that's possible. One point that I'd like to make is that spam on social media platforms still exist today. And will happen that spam gets mixed in with other promoted content and I think it's important to understand that this is a real problem that platform still face today. It's kind of a whack-a-mole thing and that Twitter attempts to review ads, but it can't review every ad that goes across its platform and therefore bad ads will air for a period of time before they're caught and removed. And so you have to be willing to advertise in an ecosystem that lives with spam. It's just a limitation. I just want to show you a quick example. Even though Twitter moves to remove specific types of ads that consumers hate, these ads still exist on the platform. So after Twitter vowed to remove Bitcoin related ads, I still continue to see them on my feed and it's something that even when I report them it takes a long time for them to be removed. This tweet that I reported a year ago still exists.
