September 14, 2017 | Austin, Texas

Influencer marketing, in some form or another, is here to stay. Budgets don’t rise by 400-600% in just a couple years on marketing fads and influencers are now gaming the system – a bullish sign. But where is influencer marketing in its lifecycle? Is it as big as it will get or a pimply-faced, socially-awkward teen? Maybe it’s still a cute toddler? Evan Asano, the founder of Mediakix, did an experiment: Create fake Instagram influencers, then see if these fakes could get payment from brands for posting on the brand’s behalf. You can guess what happened; the platform he used didn’t identify them as fakes, and each was paid a few hundred dollars. As a brand, should you be worried? Do you think Facebook is wringing its hands?

Asano’s test gives us a clue where influencer is in its life; closer to the meddling teen than innocent toddler or professional young adult. The test also spotlights “programmatic” influencer platforms which, if Influencer is ever going to scale, will need to succeed. Imagine where digital marketing would be if brand marketers had to create an ad, then submit it to every web page’s inventory for display. Asano’s test might lead you to assume the fault of fakes rests with these programmatic platforms – but you’d be wrong to think that.

To determine who’s really at fault for fake influencers you should know what’s under the virtual covers of most influencer platforms. Back in the old days (2010-ish), providers like Mediakix and Mattr could have API access to just about every interesting bit of data Facebook, Twitter, YouTube, and Instagram could offer. This means that computers could bring in selected data right from Facebook’s computers automatically in real time or in big chunks. If you had the technical ability and resources to bring all that data in, spin it around with “if this, then that,” logic and then make analyses from it, you could present a brand with all sorts of valuable peepholes into a consumer and their audience. But in 2014, Facebook turned off the data pipe. Then, as Instagram was more fully integrated into Facebook after its acquisition in 2012, API access was cut off in June last year, about the same time the algorithm-based feed appeared.

The reason Facebook cut off this access is defensive. But a side-effect is that companies like Mattr have had to get creative and very smart about detecting fraudulent Instagramers or Instagramers with fake audiences. And Mattr and its competitors have risen to the occasion. For example, Mattr servers can look at the following image (an audience member from one of our local influencers) and inform our segmentation algorithms that there’s a happy 21 year old woman with a happy 28 year old man on a boat on Lady Bird lake in Austin.

There are thousands of calculations that we, with very reduced access, run and very accurately segment while verifying that an influencer’s audience is at least 95% valid. Now, here’s where you need to ask yourself, “if Mattr can do it, why can’t the programmatic platforms?”

They could, but they’re mere cut-outs in the larger picture. Going to the source, CNN reported in 2015 that 8.7% of Facebook’s profiles were fake. If you extrapolate to 2017, this calculates to about 174 million fake profiles. Now assume good guy Facebook is cracking down and has halved fake accounts by 50%; you’re still left with 87 million. Is it a concern that you could be overpaying Facebook and Instagram for ads (or influencers, if you pay by the like) by 4-8%? Do you have a choice? Now look at it through the emerald green visor of Facebook’s CFO. It’s much worse for Facebook to have to report that up to 8.7% of their revenue for 2016 could be wiped away. Not a big deal, you think? Consider Wall Street’s marker, earnings per share. If Facebook were to incur a reduction of 4-8% revenue, their 2016 pro forma EPS would reduce an estimated 15-23% ($3.49 to $2.98 – $2.67). This would be an ugly earnings hit to Facebook. (This is an extremely conservative estimate because Facebook makes about $20 per subscriber per quarter.)

With such a substantial loss to be realized, how can the problem of fake social profiles be solved? Government regulation and oversight? Maybe. Self-compliance? We’ve already seen how far platforms will go to avoid real compliance. The only recent analog I can think of is telecom and cable’s dominance of the communications infrastructure. Resellers have technically been permitted since the mid 1990s. But they’re mostly all gone. The only competitor who’s made in-roads into selling communications services to consumers is Google Fiber. And they have done it by laying their own, superior infrastructure. And yet, this is a poor example because creating a popular social network that could compete with Facebook and Instagram has proved difficult – you can’t just throw time and money at it like a fiber network.

However, there is a solution, I believe, and it’s probably going to be the future of all advertising: CRM-driven ambassadors. With this model, a “verified customer” would promote the product on their social channels on behalf of the brand. This goes well beyond what Amazon does for product reviews, which is helpful but an under-ambitious attempt at an effective advertising strategy.

Influencer marketing, one way or another, is the future of how brands will get the word out. But the players are unclear at the moment. Social platforms are behaving like ham-fisted, protective cable companies. Upstart social platforms have what looks to be an impassable mountain range to cross to compete. But enterprises are beginning to insource influencer marketing, which possibly signals the beginning of CRM-based influencers and may represent an existential threat to Facebook’s budding influencer program.

Should you be worried? If you rely on programmatic influencer platforms and are intolerant of bloated results, yes. Regardless of how tolerant you are, you can bet Facebook views fake profiles with a wink and a shrug.



MATTR is the only full-service influencer marketing provider with detailed audience insights from PersonaMesh™. We go beyond demographics into psychographics such as values and interests so that your influencer campaigns align with your campaign targets.