A few weeks ago on the show, we talked to Emily Ho, the marketing strategist behind Authentically Social and a fashion influencer in her own right under the monicker Authentically Emmie. In that episode we talked about influencer fraud, comment pods, buying followers and the like. I even mused, “When is someone going to create a software to detect comment pods?”
Well, low and behold, I found one that not only has, but started its entire life as a tech startup to do just that for businesses and brands.
HypeAuditor has been on my radar for some time now. They’re an influencer discovery tool that allows you to search for relevant influencers across Instagram, YouTube and TikTok. I found a research report they published on influencer fraud. Here are some high points:
- In the U.S., 61 percent of influencer accounts show some type of fraudulent activity. That number is only slightly lower in the U.K. at 58 percent.
- Oddly enough, even micro- and nano-influencers aren’t immune. Almost 50 percent of those with 1,000 to 5,000 followers show some type of fraudulent activity.
HypeAuditor did the research, so to HypeAuditor I went. CEO Alex Frolov joined me from St. Petersburg, Russia to dive into the research, explain the methodology and chat about the issues at hand. Give it a listen, then share it with someone else who could use the information!
Find HypeAuditor online at HypeAuditor.com. You can find Alex Frolov on LinkedIn. The recent research he discussed that reveals the scope of influencer fraud out there can be found on HypeAuditor’s website as well.
Winfluence Transcript – Alex Frolov – HypeAuditor
Jason Falls
Hello again friends. Thanks for listening to Winfluence – The Influence Marketing Podcast. A few weeks ago on the show we talked to Emily Ho, the marketing strategist behind Authentically Social and a fashion influencer in her own right under the moniker Authentically Emmie. In that episode, we talked about influencer fraud — comment pods, buying followers and the like. I even mused then, “When is someone going to create a software to detect comment pods?”
Jason Falls
Well, lo and behold, I found one that not only has, but actually started its entire life as a tech startup, to do just that for businesses and brands. HypeAuditor has been on my radar for some time now. They’re an influencer discovery tool that allows you to search for relevant influencers across Instagram, YouTube, and TikTok. The software is Russian born and while I hadn’t done a demo of the product until recently, they provide interesting content on their blog and via an email newsletter. So, I knew who they were just not what their tool actually did that others might not.
Jason Falls
I then spotted an interesting research report they published on this very topic of influencer fraud. Here are some high points: In the United States 61% of influencer accounts show some type of fraudulent activity and that number is only slightly lower in the UK at 58%. Japan had the best rate with just 46% of accounts, that are showing fraudulent activity. Oddly enough, even micro and nano influencers aren’t immune, almost 50% of those with 1000 to 5000 followers show some type of fraudulent activity. Now keep in mind this research is data analysis. It’s not a survey of people verifying they’re doing it. It’s an algorithm that finds fraud-like behavior. Some of the influencers may not be intentionally fraudulent, but 60% in the US? Where there’s smoke, there’s fire.
Jason Falls
HypeAuditor did the research so to HypeAuditor I went. Alex Frolov is the CEO of the company. He joined me last week from St. Petersburg, Russia, to dive into the research, explain the methodology and chat about the issues at hand. Now, if you have a little trouble following along — some people, including me, struggle with accents of those non-English native speakers and as you might imagine, Alex is Russian. So, if you need a little help clarifying what he says at one point or another, know that all of our episodes have transcripts on the website at JasonFalls.com. In fact, if you ever need to get to an episode quickly we have an easy short link formula for each show. It’s jasonfalls.co/ and the guest name. So this episode will be jasonfalls.co/alexfrolov. That’s A-L-E-X F-R-O-L-O-V. Or you can just go to JasonFalls.com, click on blog or articles and find the episode in question. I’ll also have links to the research there. So you can see it all for yourself. Influencer fraud, the size and scope of the problem and at least one way to fight it coming up with HypeAuditor’s, Alex Frolov on Winfluence.
Jason Falls
So Alex, I’ve seen some of the research that HypeAuditor has put out. How big of a problem is influencer fraud in the marketplace today?
Alex Frolov
Oh, actually, it’s quite big. And it’s still, because when we ran our business in the February of 2018, I think that we can solve the problem of influencer fraud quite fast. But I was pretty naive. And the influencer fraud problem is still quite big. I would say numbers that around 45% of all accounts in Instagram, bots must follow. So, inactive accounts. So basically, it’s quite big. And in terms of engagement rate. Let me say two out of the five flags that you see on Instagram is made by all those bots programs.
Jason Falls
Wow. So the I think I have a list of kind of the fraudulent activities in my mind that might fool business into dealing with a fraudulent influence or why don’t you take me through the the two or three major things that we need to understand and watch out for so that we don’t become susceptible to that.
Alex Frolov
Basically, when we say that we need to watch out something, you should understand that the most of the fraudulent activities is quite hard to detect by your own eyes. And you need to have specific technologies as HypeAuditor, because sometimes you will just see no difference. But speaking about most popular activities, it’s definitely buying followers, when people just try to engage their followers base by buying bots, for the different programs, and software. And another one is fake engagement rate, and you can basically buy everything … likes, comments, story views, and like the smart guys, let me call them so. They buy followers and engagement rate at the same time to look like very natural. And when the engagement rate is quite good followers is growing smoothly. And without proper technologies, with your own eyes, it’s quite hard to understand where fraudulent activities is.
Jason Falls
So you know, we’re talking about people who are, you know, buying followers, which I think people understand from, you know, years back, that became kind of a problem on Twitter and other platforms. But we’re also talking about people who were actually, there are bots out there that will go into an influencer’s account, and post fake comments. Now, in my experience, and tell me if I’m right or wrong in this, the way that I look at an influencer’s account to see if they might be suspect, or they might not be? I’m looking for one word comments, like “awesome” or “Wow.” I’m looking for comments that are just emojis, which certainly some people naturally can comment that way. But it tends to be that these bots are driven by these sort of very generic automated comments. Is that fair to say?
Alex Frolov
Actually, it’s interesting theme, because sometimes people will see the thing that we analyze only content of the comments and say, are they fraudulent or not. But the reality of data analysis is much more complicated. We analyze different patterns and account’s behavior to understand authenticity of the comment. Because right now, some influencers are participating in comment pods, where the real people leave them comments, and it’s automated comments is not real activity. But it still looks totally like real activity. And also software that make these comments on these, bot programs also become smarter. And you can type in many variations of different comments. Not only are some “awesome,” or smiles, so they also become more non-diverse. There are tons of different options of comments you can get. So basically looking only on the text, probably, you can detect only spam comments when it’s obvious when someone like advertise their shop or their account. Now, there’s a lot of such comments in the in the big influencer accounts, you they’re obviously to understand but when the influencer itself is trying to increase their activity, and raise up their engagement rate by those fake comments, it’s very hard to detect, actually, and analyzing all the text will not help you.
Jason Falls
So the research that I stumbled across that sort of alerted me to the fact that you that HypeAuditor had published some some data that says, hey, this is how big and bad the the fraud problem is, was that research based on I mean, there’s, there’s I’ve seen, you know, a dozen different ways to do research reports. And one might think instinctively that, hey, they’ve gone out and surveyed a bunch of influencers or a bunch of companies and come up with these numbers, but your methodology, I think was was it not to actually do the deep, you know, sort of big data analysis that you’ve done to say, What do our algorithms tell us are fraudulent about accounts that we’re looking at? And give us a little bit of the scope of, of how you you built this research?
Alex Frolov
Yeah, definitely, we do not use any service, maybe we start to use them at the latest researches, but just to get the opinion of the market, and what they think about trends or something like this. When we’re speaking about numbers, they’re purely based on the data analysis. And there is no opinions, they’re just algorithms. You know, we have a database of 19 million accounts of influencers, it’s really big. We analyze really huge amounts of data. And we understand overall pictures like we have reports, for all those accounts. We know the quality of their audience, like from what reach people it consists, how fraudeulent it is, and we aggregate those numbers, and we use it to make all those researches. So basically, we can say that we analyze all influencers in free platforms like Instagram, YouTube, and TikTok, and get information about them. For sure that if we talk about fraudulent activities, and its problem, it’s most important for Instagram, because in YouTube situation is much better. And TikTok is just developing. So the biggest problem is Instagram, but we get our data about all free platforms. So we have a lot of numbers to operate in our research to get from this data analysis.
Jason Falls
So did I see correctly that overall fraud is actually going down gradually in the top countries you focused on? And if so, why do you think that is?
Alex Frolov
What we see on the market, the fraudulent activities is definitely changing. It’s hard to say are they going down or not? Because we still see a lot of fraudulent activities and what is important, like they’re fake followers and fake engagement, they get previously is not disappeared. It’s still in Instagram. And in 2019, we see that Instagram starts to find some kind of fraudulent activities like for example, mass following using special software to follow-unfollow different accounts. And by those the way they get new followers has become much less popular because Instagram, make new restrictions make new limits, and it’s become harder to use those methods and you also get a risk to get your account banned. So some methods is not more actual, but there are still some accounts who use mass following but much lesser that it was like two years ago, for example. But every day we see that they new ways of fake activities created like for example not so long ago, we see when the fake Story views become live. There was no previous such method of fraudulent activity and now it’s real. So I would say that the fraudulent activities is always changing. We see that things Instagram starts to find it some way but it’s still very big.
Jason Falls
interesting I was also surprised that the those with a lower number of followers actually had a higher percentage of fraudulent activity I would think that the larger the follower count, the more that count might be that account might be fraudulent. What does that say about micro and nano influencers which are all the rage right now we’ve got to be really careful with them don’t we.
Alex Frolov
Oh, definitely Yeah, you should be quite careful when you make influencer marketing but it’s a true for all size of accounts. You better know know with whom you’re calibrating. Now, but speaking about those segment of influencers. Now, nano influencers is quite authentic one. Usually they have biggest engagement rate through all other categories. And they are quite authentic. And we think that it’s based on the … Many influences they have natural, small community. They already can be called influencer. They have from like, from one to five thousand of followers, and they have a good connection with those followers. And many of them doing it naturally not trying to become a big influencer, not trying to get the money through their influence. Na .. and there is a lot of authentic influencers, that is nano influencers. But when we’re speaking about small influencers, that is have from five to 20 thousand followers, it seems to be the moment when the influencer decide, okay, I want this to be professional kind of stuff. I want these to bring me money. And when they decide, to grow their account, they start to find … search for an easy way, how they can grow it. And they have not yet experienced it. And that’s how they find all those software that sell them authentic, like following engagement. And they do not yet know that all this is just a cheat, and they will get a dance of bots account not real engagement and so on. And it’s also cost quite cheap. That’s why so many people decide to Okay, why not to try and use those inauthentic services to grow my account. And basically, those guys who are on those businesses, they just cheat and lie for those influencers. That’s why there are so many fraudulent audience for these small accounts, because they still do not understand how to be real influencer, how to make their real audience engagement but they already start to use all those methods to grow their account.
Jason Falls
So I really have to tell you, I mea culpa, because I didn’t know much about HypeAuditor a few weeks ago, I had a conversation with someone. And basically, you know, said “When will someone invent the tool to detect comment pods and inauthentic comment behavior? And it seems to me, that’s what HypeAuditor does. I’ve been playing with it for the last couple of weeks. And you’ve got a you know, this this information that you’ve surfaced in this research is, those algorithms are available basically, through your quality score, tell us a little bit more about HypeAuditor and the gap that your solution fills in the marketplace.
Alex Frolov
Actually, we started from those main features to detect fraudulent activities. And we created this unified score that we called audience quality score, which helps our clients to decide with whom they should work in a very quick way, they just look at the number from zero to 100. And say, then, very fast how general quality of the audience of influencers look like. But it was the very beginning point. It’s still one of our unique value proposition we’re still best in terms of fraud detection on the market, but from the start of our company, our product became much more complicated and covers more client needs. So basically, now you can use HypeAuditor to find influencer using influencer discovery over our database of 19 million influencers in free platforms, Instagram, YouTube and TikTok. Also, you can get those deep analysis of accounts. You can compare accounts, one to each other. You can track results of your influencer marketing campaigns. So there is a lot of tools to make your influencer marketing effective. And the latest tool that we launch. It’s also pretty interesting. It’s competitive and market analysis too, where you can get information about your competitor’s campaign, how many they spend, which influencers do they use. And also you can get information about market trends, what is growing, what is most popular at the moment. And it’s also very interesting. And we use these competitive analysis tools to analyze all brand integrations in the Instagram and YouTube, and actually calculate the size of the market. And I think we are the first one who do this. And this is also pretty awesome, because previously, there was no measures only assumptions. And now we get those numbers right through the data. So there is a pretty bunch of tools that can help our clients to make their influencer marketing much more effective.
Jason Falls
And I’m really looking forward to getting into playing with the market analysis tool. I know I’ve seen it in the platform, but I’ve been very focused on looking at how you, you know, identify influencers and playing with the filters so that I can, you know, weed out the fraudulent ones. So I really do appreciate the tool, the platform that you’ve built here. You did mention that you’re focused on Instagram, YouTube and TikTok, why limited to just those three?
Alex Frolov
Ah, actually, we’re not limited. We started from these platforms, because they’re the biggest ones, Instagram and YouTube, is the two biggest platforms, where influencer marketing goes, and TikTok is growing very fast. For the last year it grew 60 percent overall in terms of users and it’s already in the top three networks in terms of influencer marketing. So we started from covering the main three, but we actually go into other social networks also. The next two will be Twitter and Twitch. But we are not going to stop on these. And like over one or two years we will cover all social networks where influencers exist.
Jason Falls
That’s great. Alex, this is so much fun to talk to you about this. And I really appreciate the work you’re doing to help people understand influencer fraud and how to identify it. Certainly your tool is a great option for them on how to do that. How can people learn more and find you online?
Alex Frolov
Oh, actually, it’s super easy. Despite our competitors, we are a very open platform. We like speaking free about technologies that we create. We are doing our researches. And you can basically go to HypeAuditor.com and find there a lot of examples of our reports. And definitely, the demo is the king. So you can just sign up for free and request a demo. And you will get the full product overview for free. It will take not so long, around 20, maybe 30 minutes, and you will get all the information and demonstration of how our products work. I think it would be really awesome time because sometimes our clients couldn’t even imagine what we can really do. While they do not see it live. So I highly recommend to all who’s interested to get through this demo.
Jason Falls
Well, and I can attest to the fact folks that the demo process is fast, effective and certainly revealing. I’ve been having a good time playing with the tool for the last couple of days. Alex, thank you so much for your time. We really appreciate it.
Alex Frolov
Thank you very much. It was my pleasure. Thank you.
Transcribed by otter.ai
The Winfluence theme music is “One More Look” featuring Jacquire King and Stephan Sharp by The K Club found on Facebook Sound Collection.