Facebook, The PPC Show, The PPC Show [Podcast]

Trusting the Facebook Algorithm To Deliver Results

Welcome to episode #84 of The PPC Show. This is the second half of our interview with Merry Morud, Associate Creative Director at Aimclear, to discuss the recent Facebook changes in light of Cambridge Analytica and the impact advertisers can expect and get ready for.

This is part two of our conversation. Stay tuned as we continue to talk about:

  • Trusting the Facebook algorithm
  • Kim Kardashian opening for Merry at a conference
  • Updates to Audience Insights and Custom Audiences
  • Facebook Ad creative best practices

(Tune in to part one here).

Listen to the Episode

You can also listen and subscribe on Apple Podcasts, Stitcher, PlayerFM, Overcast, & Tunein Radio.

Show Notes and Transcript

We're picking back up in the middle of conversation so be sure to check out part 1.

Let the Facebook algorithm do it's thing

Merry Morud:        I also think it will encourage some other advertisers to let go of targeting and kind of let the algorithm and based on conversion angles kind of do its thing, which, I mean, I've seen work really well. Facebook has been trying to get us to do this for years and because, quote, it worked really well for Netflix and they were talking to us about a client who does like B2B sales software that caused like hundreds of thousands of dollars and we're like, "Yeah, great for consumer brands. I think we need a little more targeting qualifying on B2B or really, really, niche products," so, certainly have a targeting layer that qualifies, but I think that kind of giving it up to the algorithm is worth a test too. I think this will also encourage Facebook engineers and Facebook data people to go back to the drawing board when it comes to targeting.

Facebook data is so cool and we tell Facebook so much about ourselves and because we are connected with people, Facebook knows that we're connected to those people and Facebook knows what kind of household you live in because you're all accessing Facebook off the same IP, and so they understand your household and probably more than we think that they do. So I think they'll come up with more segments to replace that third party data targeting, maybe not all of it, but certainly they'll come up with some cool things. That's a prediction. I also think they'll come up with segments we haven't thought about or segments we wished were possible.

I also think they'll come up with segments we haven't thought about or segments we wished were possible.

For example, they had graduate targeting, maybe like two or three years ago in the spring, so they had like people who were graduates in 2015, let's say, and then parents of graduates of 2015, which was great because graduation is such a turning point and like a life cycle reset. So your kid is leaving the house and whether that makes you an emptynester or you have other kids that you're still raising, you know, you have to help your kid get to this next stage in their life, typically.

So you're buying things, you're helping guide them through the process, it's also a very emotional time for parents as well, and then, for the graduates, who are graduating high school for example, they're going off to college, they may need to be on a different phone plan, depending on the service availability, they may be buying a car because they're not going to be around mom and dad anymore, they may need a new bank account that's connected to their university, there's so many big things that are really, really, hard to predict when someone will change a bank. But we know that if they go to college and TCF is connected to the University of Minnesota Duluth, like, that's a great opportunity for the right person and the right time.

I haven't seen graduation targeting for like two years in Facebook, and it's like, why? Like, you have this data, you know these high school seniors have been sending you signals that they're about to graduate and you know it and you know who their parents are by who they're connected to on Facebook. You probably even know their grandparents, and I think you would also break that out even deeper into high school graduates and college graduates, and also, like college graduates who are 45, right? So like someone graduating with a degree at 45 is in a different life stage or has different needs than someone who is graduating college at 22, like totally different.

So like those, like that kind of targeting is so possible in Facebook and I just wish they would surface that.

JD Prater:                     Yeah, I think every EDU wants that, yeah. I think, to your point there, just to kind of jump in, was I think you've been drawn to really couple things. Like, Facebook is optimizing for us, they're doing a lot of great stuff around bidding, around optimizing for conversions, right? It may not necessarily apply to you, you may not have 50 conversions per week per ad set to really take full advantage of it, but I do think it is extremely valuable to give some stuff up to the algorithm and to trust Facebook and to kind of let them take over some things. There's a lot of things that Merry just mentioned that they can't do, right?

They don't own your strategy. They can help you with targeting options but it's up to you to select those and to do the research, and that's the part that's not gonna go away, that's the part of your job that you're always going to have, which is a nice part because I don't want to be in there messing with bids all the time.

Merry Morud:              Yeah, I don't want to be just the channel jockey, just like pushing levels and yeah, that kind of stuff. But yeah. To your point, like, giving it up to the algorithm is a bold move and I think when we say, "Trust the algorithm," I almost think people are like booing and hissing in their car-

Trust the algorithm?

JD Prater:                     Yes. Of course they are.

Merry Morud:              They're like, "Trust Facebook? No. No way," but think about it, Facebook has so much data on us as users and much, much, much, much more data and nuance data that we could even conceive of targeting as an operator. So they're looking at way more data points per user than we can even think of or would even really think about or there may be signals that a user is sending to Facebook that we wouldn't thread that needle that like that's a signal that they would buy our product. For example, if you were a hair stylist or like a barbershop and you had like really awesome colorists and that's what you're advertising ...

Facebook has so much data on us as users and much, much, much, much more data and nuance data that we could even conceive of targeting as an operator.

For example, my hair right now is many shades of pink for the most part and so, quite honestly, there's nothing on my profile, aside from photos, that would lead an advertiser, like myself, to know that I would be a proponent for this service. I don't, quite honestly, I don't know what that signal would look like or would be, like what does a person who would color their hair pink or green or blue or something that is just like not natural, like, what would those signals be?

I'm not sure, but Facebook is looking at very, very, nuanced signals that they could probably put together, like, maybe this person who likes Kylie Jenner and reads Vogue and has like a Stitch Fix and a Bark Box, those are the people who ... and that's not ... I don't have a Bark Box or Stitch Fix, but I'm throwing out ideas-

JD Prater:                     Kylie Jenner all the way.

Merry Morud:              Like, the Kardashians, keeping up with them, for sure. But funny story, Kim Kardashian opened for me at a conference-

JD Prater:                     What?

Merry Morud:              In So Cal.

JD Prater:                     For you?

Merry Morud:              Oh yeah, totally. She was just the session before me. But anyway.

JD Prater:                     She opened. Nice.

Merry Morud:              Well, yeah. She was before me, so she's an opener.

JD Prater:                     She's an opener. Yeah. Love it.

The gabillions of data points Facebook looks at

Merry Morud:              Anyway. Anyway. So Facebook's data algorithm, creative serving algorithm, is looking at all these like nuance data points and putting it together based on what you tell it to convert towards and again, it's something that we as people, we can only think of so many things. They're looking at hundreds of thousands and gabillions, well, however many that number is.

JD Prater:                     It's a ton of data points. It's a good point too. I remember, even anecdotally, so, working with ... in the agency side, working on a popular weather app, right? So we're doing a lot of weather targeting, right? So, Merry, I'll ask you this, and this was a fun realization for us, do you have a weather app?

Merry Morud:              On my phone?

JD Prater:                     Yeah, on your phone. Or do you just use the native one?

Merry Morud:              I just use the native one.

JD Prater:                     Okay.

Merry Morud:              Though I do really enjoy the screen caps that I get from friends who have like the really snarky one that it's like-

JD Prater:                     Yeah, that one is a good one.

Merry Morud:              I like that one.

JD Prater:                     What we found was a lot of people, let's just say that they had a weather app, okay, the next question is would you consider yourself a weather geek or would you say that you were interested in the weather, right? So, at what points on your profile are you saying, "Yes. I love the weather," but all we're targeting were like weather related stuff. And so you end up just kind of like tapping out that audience and you end up-

Merry Morud:              Yeah, totally.

JD Prater:                     And it's like oh, there's a whole nother segment over here that aren't interested in the weather but yet, they want a weather app because maybe they're taking their kids to school.

Merry Morud:              Right.

JD Prater:                     Right?

Merry Morud:              Exactly. So parents-

JD Prater:                     Parents.

Merry Morud:              They have to know if it's going to be like-

JD Prater:                     How to dress, right?

Merry Morud:              Way too cold for them to like ... their little ones to stand outside waiting for the bus or if a storm is coming and they may have to pick their kids up early from school or I mean, maybe it's people who work in construction outside a lot, people who are, you know, I wouldn't say like servants to the elements, but like what they do is greatly impacted by the elements.

JD Prater:                     Yeah. And exactly it, right? So this is after you do some research, but I think what I'm trying to get to is we started off with such low hanging fruit but then we realized that's not scalable and like you have to dig deeper, and this is the part that you're really getting at with the old school ways, not so much relying so much more on these partner categories or the go to audiences that we've always had. You have to dig deeper, find those different audiences, and now, the next part is really fun, is crafting the creative. We're gonna get into that in just a second because if Merry is here we have to get into creative, but one more thing with audience insights, we can't leave that one hanging. We've gotta talk about audience insights and custom audiences as well.

Changes to Audience Insights and Custom Audiences

Merry Morud:              Yeah, so this one really hurt because audience insights was just such a fantastic tool and again, this is like a knee jerk reaction by Facebook because someone found a loophole with their bug bounty and so they shut down audience insights and they shut down the audience reach when you use a custom audience in targeting, even if you use it just to exclude. That makes it really hard for us to plan budgets because we don't know how big an audience will be if we use a custom audience and psychographic targeting. We still get the audience preview in the audience tab, but yeah, in terms of audience insights, and that was visual, where you could plug in a custom audience, whether it was emails or people who bought something on your website or app or whatever, and then Facebook would give you this really colorful data about who those people were, like what kind of other pages did they like, what was their demographic makeup, what kind of households did they come from, like what was their income, and so that went away and that really hurt.

It seems like a simple enough solution for Facebook would be to close that loophole would be to just randomize the audience rounding and prediction, you know, so maybe randomly round it up three percent and the next time it's rounded down five percent and the next time it's rounded up one percent or whatever it is, just like kind of randomize it. That seems simple, but I'm also not an engineer. For now, I think it's gone but I bet it will come back. It's certainly going to be re imagined, I would think, because ... that Minnesota accent came out there, re imagined.

JD Prater:                     Don't you know?

Merry Morud:              Oh yeah.

JD Prater:                     Oh yeah.

Merry Morud:              Oh, it'll have to be re imagined. Yeah. Might've been Canadian.

JD Prater:                     A boot. Yeah.

Merry Morud:              Yeah, so some of that insight was tied back to third party data, you know, when it came to like household income and what persona profile they belonged to, whether it was like a young, up and coming, city dweller or a silver senior, whatever those mean.

JD Prater:                     Yeah. I love those.

Merry Morud:              I know, I'm just like, these both sound like the same person. Well, some of them do. But I would put money that it would come back. But again, re imagined with Facebook data. They're figuring it out. How soon it will come back remains to be seen. They'll surprise us with it one day.

JD Prater:                     Of course. Of course.

Merry Morud:              One day it'll just be like it's back and then we'll all rejoice.

JD Prater:                     Someone on Twitter will have a screen shot and they'll like, "Hey, look, everyone. It's back," everyone is like, "Wait, really?"

Merry Morud:              And quite honestly, it'll be like back for a week before anyone notices.

JD Prater:                     Always. Always.

Merry Morud:              Unless Facebook makes a big announcement and we catch on to it. They won't.

JD Prater:                     All right. So, well, we've got you here. Let's call this part three. Part three of this podcast or chapter three, however you want to say it, but I really ... while you're here we've gotta talk about ad creative. You've given a ton of conference presentations on this and I think, I'll let you speak to it, but it seems like your research is what really makes you so good at this, right? I mean, you do extensive research, you understand these audiences, and so you really kind of get into this empathy aspect of it. So I'm gonna turn it over to you and everyone is listening, so we're talking about data, we've got things going away, things are gonna be maybe coming back, but what can we take away from all of this and how can we make our ads the best possible ads?

How can we make our Facebook ads the best possible ads?

Merry Morud:              Yeah, so, most recently, I've been most passionate about creative and it's probably why I'm associate creative director, it's why my talks at conferences have steered more towards creative because, quite honestly, when you work with big budgets, like we do at Uber, we can run a lot of tests and we can figure out what works in terms of campaign optimization, targeting, et cetera, very fast. And so, there are just only so many optimization levers and audiences, buttons to pull and push and whatever, and so at the end of the day, once you've kind of optimized your ad type, your campaign bidding, and all, like, placement and that kind of stuff, day of week, where do you go from there? Where do you improve? And the answer is always creative because you can always make your message, you can always make your creative better.

Where do you improve? And the answer is always creative because you can always make your message, you can always make your creative better.

There's always improvement and also, even though you're creative right now, maybe amazing and you've worked really hard and done a lot of tests to get it to that point, great. It will probably not be amazing in six months or it will probably need something in the cultural zeitgeist may change to make that irrelevant or make that motivational message not as motivating as it once was, so creative, often times, needs to be reworked and refined and that's just for one audience. When you layer in acquisition and remarketing and re engagement and retargeting and all of these different steps throughout the funnel, there are endless ways that you, as a creative, can refine that message to move those people further through the funnel and that's an important point I want to make about Facebook ads is we, as advertisers, get so bogged down, thinking that ad has to make the sale in that Facebook ad or that search ad or whatever or display ad. At the end of the day, that ad's only job is to get the audience to take the next step. That's it.

If you can make a sale on that Facebook ad or display ad and that was enough, then more power to you, but that's not that ad's role and it rarely ever is. This has always plagued me as a copywriter because I want to say everything in one ad and I want to say to everybody all at once, but you don't and you can't. A, people aren't gonna give you the time to say it all and so you really have to boil it down to what's the motivator to get people to take the next step. And you may figure out what that motivator is, but then you'll have to refine how you say it or address it and so there's always, always, always, creative iteration and improvement that can be made, even on ever green. We, as marketers, and we have so much data and so much optimization, algorithms, and mechanisms and whatever, we often forget that just having really amazing creative can have as much business impact or as much ROI impact as just the next latest, greatest, optimization setting or algorithm.

So creative can move that ROI needle for you and your clients or you and your business, if not as much as those optimizations, I mean, like, if not more or you can do it every week or every month, like, improve on that creative and whittle away those cost pers.

JD Prater:                     Yeah, I think that's a really good one. I really like ... we think too much about maybe just the sale, but it's the action to the next step. What is the next step? And I like that differentiation of thinking through that within your own funnel, but really thinking through where is this audience in their customer journey, right?

Merry Morud:              Right.

Incite the click, incite action

JD Prater:                     And when you think through the life cycle of a customer, what do we want them to do next? And I think that's ... I really like that one. So that one is really good. So let's break it down and give the audience even some more really good stuff. When you talk about in incite the click, incite action, how do we do that?

Merry Morud:              When we think about an ad's job is to take the next step, the visual element of your ad, be it motion or static, that has to get people to stop scrolling. And Facebook says, you know, people scroll on mobile at a rate of one point seven seconds per post, so you have about a second to grab someone's attention with an image, and that's really, really, fast. So think about the image or the visual as the thumb stopper. This is why really just like weird images work, because our brain is confused and when our brain is confused we seek to rationalize or seek to appease that confusion by looking at the context around it, and that means reading the headline, reading the text copy. So, quite honestly, when it comes to an in sighting ad, it starts with an image that's going to get people to stop because if they're not going to stop, they're not reading your copy because they're not reading that copy in a second. They're already scrolling through it and they're not gonna read the copy unless that visual element stopped them first.

They're already scrolling through it and they're not gonna read the copy unless that visual element stopped them first.

So that's the visual element's job. The ad's job, as a whole then, is to take the next step. And how you insight the click is to ... it boils down to like being relevant, understanding your audience, and, like I mentioned, that can be like writing about a nod or a wink to an inside joke or a nod or a wink to how or what you targeted. We did Facebook ads for a food delivery company a few years ago, it wasn't Uber Eats, should have been.

JD Prater:                     Next year.

Merry Morud:              Yeah, next year. Next year. But we did targeting around all of these, what I call them, couch events, because it's food delivery, and so I wanted to hit people who A, were gonna be at home and wanted to be at home and wanted to be, you know, engrossed in something at home, so they're not, like, you know ... a podcast, you can kind of cook and listen to a podcast or whatever. But in terms of TV events that happen-

JD Prater:                     Super Bowl.

Merry Morud:              Yeah.

JD Prater:                     Something like that, right?

Merry Morud:              Super Bowl.

JD Prater:                     Yeah. Okay.

Merry Morud:              This was around the time of March Madness.

JD Prater:                     Oh, nice.

Merry Morud:              So we had to elude to the fact that people should order in for this event without saying March Madness, because the NCA would be down your throat faster than you can press publish on top editor. So we had to have a little wink and a nod to the audience that yeah, you're gonna be watching basketball for weeks straight, hours on end, and I think one of the lines was madness is a crowded bar or trying to watch the game at a crowded bar or something like that, or madness is like ordering crappy pizza for the big game or something like that.

JD Prater:                     Yeah.

Merry Morud:              We're like, "You have a lot of basketball to watch," you know, eluding to like this week's long event, not like a month. It was the bane of my existence for a long time. Too much basketball. But anyway, so March Madness or Game of Thrones had just come back or was about to start a new season, and so we would pump out ads that said put Joffrey's] 77 courses to shame, so that was around the Purple Wedding. We would make like fun winks and nods to the audience and those went gangbusters because they were relevant, they were kind of funny and they were smart and well targeted, and I mean, quite honestly, food is easy. It's beautiful, people love seeing that, those visuals on Facebook, so that was helpful that we had a really sexy product, which food really is.

JD Prater:                     But you also identified a motivator, right?

Merry Morud:              Mm-hmm (affirmative).

JD Prater:                     And so whenever you're thinking through, you know, we were kind of talking before, like, motivating versus like a barrier, right? So, how do you kind of play with those two things? Motivations versus like a barrier to entry or something like that. Do you think through those things as like a brand or a category whenever you're thinking through a copy and images?

Motivations versus barrier to entry

Merry Morud:              Yeah. Well, before you even kind of start brainstorming about your creative approaches, you want to identify brand motivators and barriers and then identify like those category motivators and barriers, so this is almost like a swat analysis, where you're thinking like okay, why does someone need a product like this as a category and what's the motivator for our product? How is our product different? Or what is going to make someone not take the next step in buying our product? Is it our brand? Is it our price point? Is it just they a lack of awareness or knowledge to what this is or does? Is it do we have bad reviews online? All that kind of stuff, and then identify what are the reasons that a user can overcome that barrier and purchase from you?

JD Prater:                     Cool. Cool. Well, I also saw that you had mentioned write soundbite style copy. What do you mean by soundbite style copy?

Merry Morud:              Yeah, so I've been writing in headlines or thinking about copy in terms of soundbites for, I mean, like, Marty taught me to do it years and years ago. Write like clips of copy in soundbites because, essentially, when we talk about people scrolling through at one point seven seconds, that's all you really get. They're not there to read a whole sentence sometimes, so think in like snappy headlines in terms of writing in soundbites. So what I'll often write in soundbites and then kind of refine and then identify either what needs more explanation in that soundbite or sometimes it's a question, has the audience ask itself a question, and like you really have to then thread that needle into the landing page or other supporting copy, whether that's the text copy or landing page or maybe it's like text on the image. So identifying those really snappy soundbites and then it's just like old advertising hat, right?

We always wrote in like headlines, snappy headlines or like three word statements. It's very like mad man, right? It's toasted.

JD Prater:                     I just remember, there's this one episode with Don, you know, and I think he's like, "Give me 20 headlines by lunch," or something, something like that, and I remember thinking through and always hearing this as well, like, the first ten are always the easiest. You can write ten headlines pretty quick. It's the last ten that you really have to think through, and a lot of times those are actually your best ones rather than the ones that come to you really quick, because those are probably, again, the easy ones, but it's the ones where you use the research and you really kind of put yourself in their shoes, getting the empathy for the user that's gonna be reading this. And sometimes those later ones will actually be the best performers.

Going way too far with ad copy and then walking it back

Merry Morud:              Yeah. I mean, really, it's about riffing on those initial ideas, and when you're either in a brainstorm with other people or just by yourself, write out, like in your head, sometimes you're like, "I want to say it's like this and that's really ineloquent, or maybe it's just totally wrong or really un PC," that's okay. Those ideas don't have to be set in stone, but write out that abstract or what you're trying to say, and then kind of boil it down. I also am an advocate of going way too far with ad copy and then walking it back, because if you write ad copy for marketing legal or like for a brand, your client, that you're writing it to get approved, you're always gonna have really safe, boring, ad copy.

And so, go a bit too far, have legal walk you back a bit or have the client walk you back a bit, but then they'll understand what you're trying to say and maybe just sanitize one word instead of a whole sanitized message that's not very compelling. Yeah. They'll say like, "You know, I really like it, but we can't say revolutionary," like, no brand has ever let me say revolutionary. I guess they think the public thinks they may actually start a revolution, which whatever, but-

JD Prater:                     That's a great example.

Merry Morud:              It's like software companies and we're like, "It's accounting software revolution," and they're like, "Not, not, revolution because no," and I was like, "There's not gonna be a riot of accountants. We're not starting an accounting revolution."

JD Prater:                     What if I start a PPC revolution here?

Merry Morud:              Yeah.

JD Prater:                     At AdStage. Yeah.

Merry Morud:              That's our next podcast. PPC Revolutions.

JD Prater:                     Well, cool. Let's really kind of take it home, bring it home. So we really talked quite a bit around Cambridge Analytica, Congress, we talked about ad copy, we talked about images, we talked about partner categories, third party data, audience insights. Where would you want to leave the audience with? Like, you know, what are some of those takeaways where you would say, "Hey, this is what I want to say."

Key takeaways from all Facebook changes

Merry Morud:              Regulations is coming more. More disclosure and regulation is coming. It's already happening and I think it will happen as a whole to the internet, so not just Facebook. And what happened when Zuckerberg went to Congress was he freaked them out, and not just by what Facebook knows, but he freaked them out as to what is possible out there online in terms of data and privacy, so I would expect more of that to come. Maybe not at the level of Europe's GDPR, but at some level. It's gonna come and it will probably be over corrected at first and over reached, but we'll likely walk back because certainly that there are big companies who have lots of lobbying money to write those, quote unquote, plain English legislative documents. So that's coming and just get ready and make sure that what you're doing is above board. In terms of what went away with Facebook and in terms of third party data targeting, not worried, really excited. Advertisers out there in podcast land, you know, start transitioning audiences so it's not a complete surprise when that targeting goes away and you have to start a new audience.

I would start them both now, so you get at least a bench mark of expectation, of like we expected this and this is what we were getting when we were using third party data, and this is our cost pers or what we're getting when we use Facebook based data or just letting the algorithm run wild. So start doing that now. Start mapping out what audiences need to be transitioned to the new data zeitgeist. When it comes to audience insights, like I said, I think that'll come back and then the audience reach, when you use custom audiences, I think it'll come back once they figure out how to close that loophole. It'll happen.

I'm very confident of that. It was too good of a tool to take away from us and it was so helpful and I know people are upset that it's gone and Facebook knows that. When it comes to creative, always, always, always, improve creative. I know you think that maybe just changing a couple headlines or body text isn't going to move the needle a whole lot but really, it can. It can if you do it in a very thoughtful way, in a strategic way. And really, it boils down to getting people to stop scrolling. So your job, as a Facebook advertiser, when it comes to the creative piece, is to get users to stop scrolling and you do that with an image that grabs their attention and then get users to take the next step with copy that insights that click, invites that click, and gets users to continue their journey or start a journey. Yeah. I think that's it.

JD Prater:                     Nice.

Merry Morud:              I mean, I feel like there's more I could say, but like-

JD Prater:                     Oh, we could go on and on.

Merry Morud:              I feel like we have a good solid takeaways there.

JD Prater:                     Awesome. Well, Merry, thanks again. It was such a pleasure having you on the podcast that was actually recorded, and it was also a pleasure having you on the one that wasn't recorded. But thanks again for your time, effort, and for all the things that you do for the industry.

Merry Morud:              Yeah. Thanks for having me, JD. We should do it again.

JD Prater:                     We should. Well, maybe like a year later, this is what we've learned from this podcast, it's gonna be a good one.

Merry Morud:              Next time Facebook gets in trouble. Hot takes.

JD Prater:                     Yeah. Call in Merry. That's right. Well, cool. Well, thanks again, everyone, for listening in to the podcast and we'll see you guys next week.

Merry Morud:              Bye.

JD Prater

JD is the Director of Growth Marketing at AdStage. He’s a savvy marketer, digital strategist, and avid cyclist. A stereotypical coffee snob and recovering Coloradan, he’s a creative thinker who sees the big picture but loves getting lost in the details.