Transcript of Putting Social Media Myths to the Test
Back to Podcast
John Jantsch: This episode of the Duct Tape Marketing Podcast is brought to you by Opteo. That’s O-P-T-E-O, dot com, slash ducttape. And if you go to that link, you’re going to find out how you can get a six week extended trial of this Google Ads optimization software.
John Jantsch: Hello and welcome to another episode of The Duct Tape Marketing Podcast. This is John Jantsch. My guest today is Scott Ayres. He is a content scientist at AgoraPulse, and he’s also a contributor to the Social Media Lab at AgoraPulse, and that’s what we’re going to talk about today. So Scott, thanks for joining me.
Scott Ayres: Hey, appreciate you having me on. I’ve listened and followed you for, well, over a decade now probably. So it’s great to get to talk to you for the first time ever. Yeah.
John Jantsch: Well, I’ve been a fan of what you guys are doing there at the lab. I think everybody really likes to see … Experts like me can pontificate about stuff all day long, but I think people like to see results, don’t they?
Scott Ayres: Yeah. I was part of the problem for a very long time. I was telling somebody other day, for eight, nine years I’ve blogged and written about social media, but it was usually that opinion type stuff. “Here’s what I think should work,” or the top five things to post to Instagram. But there’s no data behind it. It was just off the top of your head. And so I think we, as experts, have unfortunately fed that too much. So I love that we’re running these long form experiments and really just testing weird little things to see if they really work or not.
John Jantsch: Well, and I think the right answer to a lot of people’s questions is, “Who knows?” I mean, because, I get asked all the time, “How often should I blog? How length should a blog post be?” And really, my best consultant answer is it depends. And I think that that’s probably what the data is actually proving out to be. So explain an experiment. I mean, how do you pick it? What does it look like? What do you hope to gain? Just kind of go through the process of an experiment that you perform.
Scott Ayres: Yeah. So basically what we do is we actually do follow as close as we can the scientific method. You remember that ninth grade biology. So what we do is we kind of go out there and see what people are talking about or see what people have posed questions to us. “Hey, what is this? Can you do this?” Or, “Is this better than that?” And then we looked through what everybody said and see if we can test it first, because there’s some things we just simply can’t test. If someone comes to me and says, “What content works best for the 50 and over crowd?” Well, unless I have a bunch of pages and accounts targeted towards that demographic, I really can’t test it because the product may not fit them.
Scott Ayres: And then we form some hypothesis based on all of that research that we’ve done from experts, if we can find anything. A lot of times I’ll find stuff, it’s just opinions. There’s no data on it. And then we go through the process. We do the hypothesis. We start the test and run it. Typically, most of our tests are going to run from anywhere from 10 to 30 days, just depending on what it is that we’re testing. Paid ads tend to be about 10 to 14 days, just based on the ads typically. And then we pull the data and I spend an enormous amount of time in spreadsheets because I’m the type where even though I can go to tools, even like Agora Pulse or whatever else out there and pull data, I like to go straight to the source just to make sure that everything’s legit and it’s 100%, nobody can question what the data says.
Scott Ayres: So we take all that data and then try to draw a conclusion on it. And what we do is we use a a thing called a statistical significance calculator, which is a phrase I couldn’t say two and a half years ago because it’s like a tongue twister. But basically what that does is you put all the numbers into this formula and it puts out what’s called a P value. And if the P value’s at least 95% or higher, it means you have at least 95 to 100% certainty that if I run this test again or if you, John, ran this test, you should get the same results. And as far as the data science geeks are concerned, 95% is the minimum, where you and I probably would have said, “Hey, 50% is pretty darn good,” but we had to have that 95% before we can say, “Yes, this is statistically significant.”
Scott Ayres: And a lot of times we find out, like you were saying in the beginning, sometimes we found out it didn’t really matter. So if one of them was easier to do than the other, you might as well do that, or if you prefer it better, you might as well do that. And so sometimes we come up and our results just don’t show us anything. What it does show is that, hey, that effort that took you 30 hours a week is a lot. You shouldn’t do that when you can do 10 hours a week. So that’s kind of how we go through the process in a quick nutshell there.
John Jantsch: Yeah. I kind of live by the motto that 50% of the time, 90% of the statistics are made up to prove the results that you want, right?
Scott Ayres: Right, right.
John Jantsch: So how do you, though, account, and I know that the calculator and the P score is trying to say, “Yeah, plus or minus so much,” but how do you in social media account for … There’s so many variables that come into play in a lot of the experience. On a Tuesday in Texas, if the wind’s blowing, Facebook’s going to show something else.
Scott Ayres: Yeah. I mean, we try to do our best anyway, and sometimes we just simply can’t. And we’ve gotten better as we get older here, but we try to do it across a couple of different accounts depending on whatever it is that we’re testing.
John Jantsch: Well, I got to stop you. You said you’re getting better as you get older?
Scott Ayres: Well, as the lab gets older, and I’m getting older too. I’m in my 40s now, so I’m a little less stubborn. I’m getting more stubborn actually probably, but I’m actually learning, which is hard for you when you’re a guy in your 40s. But yeah. So I mean, what we’re trying to do now is we try to test across multiple accounts and multiple industries, because if you just test, the bad thing about a lot of us in social media marketing … You can probably attest to this … We say, “Hey, go do this.” Say back when Periscope started. I started a Periscope account and got 10,000 followers watching it. Well, you have 100,000 followers. Yeah, sure. You probably did get 10,000 people to watch it. But Bob’s Shoe Shack has 10 Twitter followers, no one’s going to watch his Periscope.
Scott Ayres: So I think in the social media marketing world, we’re guilty of that too much, just kind of pontificating what worked really well for us we assume will work well for everybody else and it doesn’t. So I like testing on small business accounts and local accounts. Right now on Instagram, I’ve got like eight or nine accounts I’ve been working on for months that are just, they’re entertainment pages, if you will. They’re about animals or about cars, motorcycles, fitness industry, that sort of stuff. But they’re all different niches that are very targeted to getting their followers and being engaged. That way, now I can test on them and kind of get an average across the board. That really helps, because what would happen a lot of times is you have two accounts in the same niche and you tested on it. Well, it just may work for that niche. It may not work for the other one.
John Jantsch: Yeah. And I think that’s the caveat with all of this. What you’re in some ways doing I think is providing people maybe a shortcut but they’ve still got to do their own testing, don’t they?
Scott Ayres: Well, yeah. I think everybody should always test no matter what it is. A lot of stuff we put out there just kind of gives you a guide. This did work. You go try it and see. If you’re someone out there who has a business and you’re just listening 100% of what somebody says and doing it and never testing, then you don’t know. You may be missing out on something that could have been working for you just because you read the blog on it. So I think you should always test stuff and change up what you’re posting constantly because like you said earlier, social media changes so quickly that on Tuesday, the wind blowing across here in Texas, the algorithm changes. And so you’ve got to constantly move around. But I do think our goal is just to kind of help you, if you’re starting out at square one, maybe we can help you get to square two or three a little bit faster. That’s really our main goal.
John Jantsch: So let’s talk about a couple of the experiments and you can expand on it, but I’ll start by kind of what you were trying to test. So one of the more recent ones and a lot of people in Facebook, they’ve given you lots of options. Now you can have a video carousel image ad, a single image ad. So you were trying to get trial signups and you were testing the carousel. The hypothesis was the carousel ad will outperform the single image ad and generate more free trial signups. And I guess it’s worth noting that you were hypothesizing a result. You weren’t just saying, “Let’s test these two things.” You were actually suggesting that you thought a carousel ad would outperform a single image ad, and I would be behind that. What’d you find out?
Scott Ayres: Well, we initially did an experiment we had done by a guy named Charlie Lawrence. Give Charlie a shout out. He’s the guy over in the UK. Love Charlie the death. Sometimes we have guest bloggers do stuff for us, which is kind of fun, because they have a different set of accounts than we have. And so it’s always nice. We’re doing a test right now with a couple of other companies. I won’t name them here just in case you have competing sponsors. But we love when we get other people to get on there.
Scott Ayres: So Charlie ran this experiment trying to see if you can get trials over to Agora Pulse. And in the end, he didn’t really find out much. It was almost a wash. The clicks on, let’s say the free trial [inaudible 00:09:19]. Let’s look at that number because that was the one we focused on. The carousel ad format got 51 free trials. The single ad image got 50 free trials, so basically the same. We spent the same exact amount of money on it. The reach was almost identical. The link clicks were almost identical. And so what I found on the end is neither one of them generated more signups. Neither one of them outperformed better than the other one.
Scott Ayres: So kind of what that tells me and tells us is, while it might look a little bit better to have the multiple carousels, you don’t necessarily need to do it and take the time to do it, because there’s a lot of people who don’t have that option. Maybe they don’t have enough images to do it, especially I’m thinking about a small business or something didn’t have it set up. So in our case, it doesn’t matter if we did it either way. Now, the caveat there, we did it on one account and we just did it to our free trial signup. So obviously, you’ve got to know that when you’re reading this and it may not apply to you if you’re trying to get foot traffic into your store or something like that.
John Jantsch: Well, and just to put out another shout out for testing, I mean, you theoretically could have changed out that single image and maybe it would have bombed or you could have changed so many variables.
Scott Ayres: I know, especially on the paid ads. I’m doing a lot of our paid ads right now. I just started doing them a lot for the lab now and the stuff you can do, I will say the cool thing in Facebook ads manager that I’m just in love with is the ability to AB test in the ads without having to … It used to be, remember, you had to run two separate ads your own and set them up? Now it’s just like, “Boom, I want to test two different images,” and now you can just test two different images. I’m in love with ads manager right now because of that.
John Jantsch: To the same audience. That’s what was always a killer. You had to create separate audiences and everything. So even then you created your own variables.
Scott Ayres: Yeah, and then the ads were running at different times. The worst thing you can do when you’re running an ad experiment is accidentally let it go to a different placement than the other one. I just set up one here recently testing an image that was our graphic from our blog to drive traffic versus a graphic with me. We’re on a podcast so you can’t see us, but normally I’ll wear this big orange wig and a lab coat when I’m in our live show, we’re doing presentations. So the other one is that image of me with my hands up in there or something like that. So I’m going to be kind of curious to see if people click based on an actual photo versus the typical featured animated image we use for our blog. But it took like two seconds to set it up in ads manager and we were done.
John Jantsch: One thing that I’ve learned over the years is there’s a lot of times when … I mean, it just makes me hooked on testing. There’s so many times when I’ll go, “Well, look at this image. It’s awesome. It’s going to kick butt.” Never does. I’m always wrong. And so there’s no accounting for taste.
Scott Ayres: Yeah, I think that’s true. And it’s times, it’s what’s going on in the world on that day that you tested. There’s so many variables that pop in and you may post something that … like our guys had tested for Agora Pulse, and this wasn’t part of the lab, but they were testing images of people versus this goofy animated bear jumping out of the bushes, and the stupid bear outperformed the regular image. It didn’t make any sense. We’re like, “This shouldn’t work,” but people were signing up for free trials like crazy with it. So they use it all over the place. Just, hey, why not? If people are going to sign up, might as well.
John Jantsch: Well, I see a lot of people actually advocate that idea of all you’re really trying to do with the image is get somebody to stop.
Scott Ayres: Yeah, stop the thumb. You got to stop the thumb.
John Jantsch: Yeah. So that could be what’s going on there as much as anything.
Scott Ayres: Right, true.
John Jantsch: This episode of The Duct Tape Marketing Podcast is brought to you by Opteo. It is a Google ads optimization software tool, and let’s face it, if you run Google ads, if you run multiple ad accounts for your clients, you probably know that it’s not the most user friendly interface, and that’s where Opteo comes in. You can actually sign up for alerts that you’ll get in an email or in Slack and they’re going to tell you, “Hey, something’s wrong. There’s a big cost spike or a spike in impressions or something’s not running. You’re over budget, under budget. Your conversions are going crazy.” You’re going to get that every day as an alert so that you can go there, you can fix it, you can tell a client about it. You don’t have to wonder every day, “Do I have to go and check out every single one of my accounts to make sure nothing is going wrong?” That’s up Opteo, O-P-T-E-O, dot com, slash ducttape. Go to that link and you get a six week extended trial. Check it out.
John Jantsch: So I’ll give you a chance to do a little shout out for AgoraPulse. How does AgoraPulse, the tool itself help you in some of this? If I’m a person saying, “Yeah, I need to do more testing,” how could AgoraPulse play a role in that?
Scott Ayres: Of course, we have a lot of different things our tool does, but for one, at the basic level, it schedules your content, which is nice when you’re doing testing, especially across multiple accounts. I’m running tests on 8 to 10 Instagram accounts. I could never do that inside Instagram. I’d have to be logging in and out all the time. So at the minimum, at least we’ll do that. But from there, it’s getting the reports, it’s getting all the engagement numbers that are right inside the app. It’s managing all your comments, especially ad comments. That’s real important. If you’re running ads, you can’t really manage those on Facebook. It’s impossible basically. But we handle that.
Scott Ayres: So right now, we’ve got an ad running that’s getting a lot of interaction. It’s about one of our lab posts, actually. It’s getting some disagreements on my data, which I’m all for the discussion, and so I’m having to make sure I keep up with that ad comments and I do it right inside of AgoraPulse where I don’t have to wait for a Facebook notification and be like a day late on that response. It’s all right inside there. But you can have team members, you can have all different kinds of things in the app. So our app really, for me, I only use it for myself and for our accounts, but if you have team members or you’re an agency, it’s one of the best tools you can have. And I used it before I worked for AgoraPulse. I was paying for it before I came onboard. So now I don’t have to, luckily.
John Jantsch: So there’s a special breed of animal that comments on Facebook ads, it appears.
Scott Ayres: At times, yeah, at times. You don’t get a lot of positive yeahs. You get a lot of this negative stuff on ads, it seems like.
John Jantsch: Well, I get stuff sometimes that I’m like, “I don’t even know what you’re asking or saying.” It’s like-
Scott Ayres: Well yeah, you get the weird troll type stuff, or, “Hey, come join the Illuminati,” or something like that.
John Jantsch: All right, let’s talk about another one. Posting content on social media, one of the hot topics, and you specifically went to LinkedIn, but one of the hot topics is, do you get more engagement with long stuff, short stuff? I think when people are just, “Hey, here’s my latest blog post,” you’re probably not getting much engagement but you tested thousands of words of basically a blog post on LinkedIn. So what did you find out? Long versus short for engagement?
Scott Ayres: Yeah, this was an interesting test that Melonie Dodaro, who’s kind of known for her LinkedIn marketing abilities, she was seeing an interesting trend on her text only updates. And so we did a test with her first and found out that our text only updates on the different accounts that we tested got like a thousand percent more views and engagement. It was ridiculous.
Scott Ayres: And so then we said, “Okay, let’s take it a step further. Let’s see if long versus short works better.” And what we do on most of these … I’ve done one on Facebook, I’ve done one on Twitter, I’m actually about the publish one in a few days that’ll be on Instagram character link, we kind of give an old hat tip to Twitter, the 140 characters, under 140 being short, over 140 being long, just to kind of make it easy to kind of look at. But we tested across, we did three different accounts. We did Agora Pulse’s account, did my own personal account on LinkedIn, so you have that mix of a personal account with about 9,000 connections when we did it. And then I used a small local business. Actually, I’m sitting in their office. I run an office from a recruiting firm in the construction industry, so very niche, and we tested on their LinkedIn account.
Scott Ayres: We did like 14 short posts, 14 long posts and did it over a two week period. And then what I always wanted to do is just average all those numbers together and kind of see which one did better on it. And so basically when we get done … I’m scrolling down and looking here on my other screen here … What we found the short text only updates on LinkedIn got about 13.85% higher views compared to posts that were over 140 characters long, so almost 14% more views. In marketing, that’s a pretty good amount. It’s not statistically significant in our little nerdy calculator, but it’s enough for me that made me go, “Okay, maybe long is not always better on LinkedIn. Maybe people want to have conversations.”
Scott Ayres: And if you think about LinkedIn, LinkedIn’s become cool again, last year, too. Everybody’s flocking back to it. I’ve been on LinkedIn longer than any other social site out there other than MySpace, which I guess nobody uses anymore. And LinkedIn, they tend to be in your industry. So whatever industry you’re in, they tend to be in that and they love to talk about subject matter around that. So if you’re doing short updates that kind of spur on conversation, it makes sense that people want to talk about it. So if you’re a realtor and you say, “Hey, what’s your best tip on an open house?” Well, all your realtor connections are going to want to add in their two cents on it. But if you write a long blog post on LinkedIn, while they might read it, they’re less likely to take the time to engage with it. You remember back 10 years ago, people comment on blogs like crazy. Now they don’t. So I think it’s the same on social. People don’t comment as much.
John Jantsch: So how do you feel about the LinkedIn algorithm? I mean, obviously all of the social networks, particularly Facebook, have kind of made it so that it doesn’t matter what kind of following you have. Nobody’s seeing your organic stuff. So do you feel like LinkedIn’s still a little bit wide open in that regard?
Scott Ayres: I thought it was. I mean, I think it still is, but I think so many people have flocked to it and people are getting too many connections. For me, for example, I was looking at mine today. I get about 20 or 30 connection requests every day and I’m one of those guys who just accepts all of them, because whatnot. But now my feed is miserable. I’m looking at, I’m going, “This is way too much. I can’t take it in.” And so I think there’s a balance you’ve got to figure out as a business or as a person. But as the business side of it, you’ve got to stand out, whether that’s now the hashtags you can use on there, which is something worth testing and looking into. I think the shorter updates of live video is coming around for LinkedIn. I think it’ll end up being just like Facebook and Instagram and Twitter where the majority of your followers aren’t going to see it unless they go to you or they’ve got you in a list or you’re active in a group or something like that.
John Jantsch: So I’ve invented a drinking game for LinkedIn, and so it goes like this, that when people give me a connection request, I look through my connection requests for the day and I bet on the ones that are going to reach out to me and try to sell me something within 24 hours. And you have to take a drink every time they do.
Scott Ayres: Oh, I would have taken about 30 shots today alone, because I logged in earlier. I’m looking at them right now. I’ve got like five or six just, “Hey, wanting to connect and see how are you doing,” and there’s some link in there. There’s a Bit.ly link. That’s the problem with LinkedIn, I think. I wish they figured that out where … I think there’s a setting. Judy Fox probably would know. You could go in and turn off the ability for people to send you these random messages. On LinkedIn, the one thing I hate the most is the anniversary, happy anniversary of your job or your birthday, because you get 9,000 connections, you’re going to get a lot of them. I can’t get to the stuff that I might actually want to get to.
John Jantsch: All right. So any huge surprises. Over the time you’ve been doing this, did you just get blown away by how definitive something was that you didn’t think was?
Scott Ayres: Probably the one that’s gotten us the most traction and is the one right now that’s getting us the most … I was talking about people commenting on ads. We’re running an ad right now to an experiment we did on Instagram that is it better to put the hashtags in the original post versus putting it in the first comment, because a lot of people for years have been teaching, “Put it in the first comment, put it in the first comment.” And up until recently, there was no tools, legitimate tools, anyway, that would do that. There are a few now they’ll do the first comment and Instagram’s allowing it for whatever reason.
Scott Ayres: And so we tested that because my theory was, my hypothesis was that you would have a higher reach if you put it in the original post because I’m not a fan of stuffing it into the comments because to me it just looks a little … I don’t know. I just don’t like the look of it. It looks a little spammy. It’s like you’re trying to get around everything. So we tested across three different accounts, our Agora Pulse account, my personal account, which is set up as a business account. At the time when I ran this test, I actually had a local bounce house business, you know, right now, bounce houses and water slides and stuff. A fun little business. I loved it to death. Did it for about five years. I sold it to my brother recently. I got tired of being out in the heat in Texas.
Scott Ayres: So what we found was pretty interesting, was when we put the hashtags in the original post, so whenever you schedule it or post it, you put it in the original post, it had 29% higher reach than if you stuffed it into the first comment. That’s a lot. It’s a big difference. And if I can get 5% more reach on Instagram organically, I’m going to do that. 29% says a lot there. Granted, it was three accounts. I’ll probably go back and test this across some more accounts just to kind of see if it changes anything. But what that tells me [inaudible] a lot of our friends in the industry have said, and I actually got a couple of not so nice messages on this one, because a lot of people are teaching this and they’ve taught their followers this for a long time, “Stuff it in the first comment. Put it in the first comment.”
Scott Ayres: But what it tells me is Instagram has gotten smart to that, for one. Any time marketers come up with something and it kind of gets around the algorithm, what happens? They always change the algorithm. And so the hash tags aren’t for your followers. They’re for people who are going to discover you and find you. So you want to get it out there as fast as you can. If you don’t get it out there right at the beginning or if you put it down in that comment, your likelihood of being in the explore feed is decreasing obviously because it’s timestamped. So that was probably one of the ones that’s got us the most engagement, the people disagreeing with us. But I challenge that. I’m like, “Okay, go test it and tell me what you see.”
John Jantsch: And I think you bring up one really good point. If you do a test and let’s say we go back to our original Facebook one and the carousel ads are just slamming the single images, it’s like, “Yeah,” well, everybody’s going to go that direction. And then guess what? A year later, nobody wants to do carousel ads because everybody’s doing them. I’m sure that some of your experiments, you could go back and retest a year later and because of whatever variables, you’ll get completely different results.
Scott Ayres: Yeah. And that’s just the name of the game in social media. Facebook and Instagram and Twitter, LinkedIn, I mean, they are businesses who are out to make money. And so, I mean, they’re going to make sure that for one, that they create a need for you to run ads. Let’s just be honest, and so organically when the marketers have figured out a way to get around the algorithm, they’re always going to adjust. It’s not like they don’t know.
Scott Ayres: If you remember back years ago on Facebook, there was that whole thing with, there was a lot of clickbaiting. You put an image of a baby up and then you had some funny description, but the link would be to something else, to a landing page or a quiz game or something like that. Facebook got smart to that, and what’d they do? They diminished the reach of those sort of posts that had a link with the photo. And so they are always going to figure this stuff out when you do that. Now, I guess a year later, they’ll come around it and might change that. So it’s a constant thing you’ve got to test.
Scott Ayres: But for me on this one, if I’m going to get at least 5 to 10% more reach by just putting it in the first comment, I’m going to put it in the first comment … I mean, in the original post, excuse me … even if that means you do the whole white space thing. You make it where it goes to the read more on Instagram and then you kind of put it below that where no one in the feed sees it, but it still shows up in the explore tab. Even if you do that, at least, at least get it out there when you schedule it and post it and don’t waste your time.
John Jantsch: It’s funny. I’ve seen a few people doing that in LinkedIn now where they’re just making their comment as big as possible and then putting a link in it so that it takes over the entire thing.
Scott Ayres: Well, there’s a lot of people right now, I’ve had people actually come to me and say, “Hey, I want to test putting the link in the first comment on Instagram,” I mean, on LinkedIn, because they’re seeing better results from it. And so maybe there is a little trick, but again, that’ll change as soon as LinkedIn figures it out.
John Jantsch: So Scott, where can people find out more about the Social Media Lab and AgoraPulse?
Scott Ayres: Yeah. Our little room in the house on AgoraPulse is agorapulse.com/socialmedialab. You can find us everywhere on all the social media sites. We’re @AgoraPulseLab. That’s a brand new social media account we’ve created here recently on all of them. So the follower count’s low on them, but we decided to finally get our own accounts here in the last couple of months. So AgoraPulseLab, you’ll find us everywhere, or you can search Social Media Lab on podcasts, wherever you listen to podcasts.
John Jantsch: Awesome. Well, thanks Scott. I love Social Media Lab. A lot of fun there and appreciate you stopping by, and hopefully we’ll run into you soon someday down the road.
Scott Ayres: Yeah, thanks. Thanks for having me on, John. I really appreciate it.