Sniff out fake Influencers with analytics

Efficacy of Influencers — a Nacho Analytics Case Study

When Mike Roberts first demoed Nacho for me, I’ll admit to being more than a little jealous. How could something so powerful be created less than a couple miles from my own office?! It’s a natural fit for the makers of SpyFu though, considering it is a truly unique way of conducting competitive research. SpyFu has been in the competitive analysis game for a very long time and remains one of my go-to tools when building out a growth strategy. Nacho changes everything for us as marketers.

It would be easy to focus on the simplicity of Nacho as it pertains to making money. Indeed, one of the first use cases that popped into my mind was using this for stock market gains. That process only really involves finding an industry where the majority of economic activity amongst the companies exists on one website, such as buying flowers. Then, after collecting the S-filings of those publicly traded companies, allowing Nacho to run for several weeks with a couple goals created to gather enough data. At this point, one can begin to see patterns and should be able to determine whether sales are trending up or down, as an industry and for individual companies. Even an old marketer like me can see how to trade based on such information.

However, that is probably not why you clicked to read this article. I run Intellifluence, an influencer marketing platform, and I want to share with you how I’m using Nacho to train our own backend determinations of influencer efficacy. In English, I wanted to know if there was a way to use Nacho’s traffic data on Instagram (yes, I got to see Instagram’s data!) to see if traditional measurements of engagement (in this case likes and comments) are a valid way of roughly screening for influencer fraud.

It doesn’t take much money to purchase a few hundred thousand followers, but surely an influencer wouldn’t also go through the trouble to consistently purchase likes and comments on posts would they? Does look odd when compared to traffic data?

Let’s dig in on the data.

Behavior flow for Instagram

Keep in mind that we’ve only had ~7 days of data at this point in the analysis. Also, for those that are wondering why traffic appears to be low, Google Analytics basically breaks on giant sites. The factor is off by 1000x. Here’s some selective data for you (I’m feeling greedy and don’t want to tell you everything; you can buy a subscription for that).

instagram influencer data
The highest traffic on the list belongs to @sudoooooo

instagram page for sudoooooo

Is it strange that such an account should have the most traffic? Yes. Yes it is. With 82,33 followers, one wouldn’t generally expect to have what amounts to 4M pageview. Of course, we can see pretty quickly that this is fake, as 100x of that is not unique. The # of likes as an engagement seems rational and the # of comments is also realistic, but the traffic is simply weird. If I were being charged based on traffic from the account, I might be concerned.

The highest unique traffic on the list belongs to @_agentgirl_

 

Instagram page for _agentgirl_

While the unique traffic is one third of the overall traffic, this influencer passes the proverbial sniff test more favorably with a better “like” engagement on those 8.8M followers, and a very solid “comment” engagement. While it would not be impossible to do, maintaining numbers like this over a long period of time would be rather difficult. She’s welcome to join Intellifluence as an influencer!

Who has the most followers? This won’t surprise anyone that follows the influencer space. @kimkardashian edges out @kyliejenner 116M to 113M.

 

Kardashian sister showdown

In both cases, the “like” engagement is extremely strong, occasionally over 4M, and “comment” engagement with Kim is second highest of the entire group. Of course, we know these ladies are social celebrities in their own right.

Which influencer had the best “like” engagement? @_grisha_man_

 

Instagram page for _grisha_man_

If I were a brand, I’d be wary. 50% of the followers like the post, but the pageviews cannot possibly be real; 1.5M views on 106 followers makes me nervous. Without Nacho, we could not have known this before engaging.

Which influencer had the best comment engagement? @chrisssandresen

 

Instagram page for chrissandresen

Once again, something is going on here. The traffic numbers are not grounded in reality, and looking at data only found using Nacho, I could see that there appear to be takeover attempts by this account of other accounts. Personally, I’d steer clear.

From a brand manager perspective, it would be very handy to have this type of data available when engaging with influencers. While one should never rely solely on one glance of information to tell a full story, it can be helpful in order to improve one’s analysis on the expected efficacy of an influencer. As an influencer platform owner, I know I have a very long list of influencers to reach out to now thanks to Nacho giving me the ability to screen out possible bad actors.

***

Joe Sinkwitz‘s favorite food is, coincidentally, nachos. With 20+ years of digital marketing experience for some of the most competitive industries on the planet, he is CEO of influencer marketing platform Intellifluence and Principal of boutique search agency Digital Heretix.