Key Article Takeaways:
- Audience measurement and avoiding unnecessary budget sinks are prevalent topics of discussion in brands/marketers running influencer
- Wanting to hone in on specific audiences isn’t reason enough to switch all funds from organic influencer to paid
- If a brand wants to work with a specific influencer but their audience doesn’t completely match what they’re looking for, brands can use that as a negotiation tactic
I recently had a fun and interesting conversation with a large brand where it occurred to me that we were wrestling with some topics that felt incredibly complex on their face, but with some relatively simple math could be easily quantified. The issue was that several of the influencers we were looking at had pretty wildly divergent audiences demographically, across gender, age and geography…and all of those factors were important to the client.
There isn’t going to be anything complicated about the math we’ll walk through here, but if not already doing so, hopefully you’ll be encouraged to apply similar techniques to your own influencer marketing executions. The tactics can be applied in a number of ways and if nothing else, your manager is likely to be impressed by the rigor you are bringing to your work.
If you haven’t done so already, you may want to read the post, “Why Is Influencer Marketing Usually Priced based on Reach?” as it is a foundational piece for this post.
The conversation with the brand was focused on picking 3 influencers out of a Lookbook with around 10 or so options. The client had given us a couple of different behavioral targets to investigate and in order to go deep in those niches, we had surfaced influencers that had larger than usual diversity in terms of primary platforms, size, age, geographical and even gender following profiles. The unintended consequence was a bit of a search for a unicorn that had the perfect mix of attributes. Inevitably, what started slipping away (for us and the brand; this is a “how the brain works” issue, not “what was the brand doing wrong” issue) was the ability to maintain a sense of scale & even more so, any sense of relative value. We weren’t even providing the cost metrics to allow the brand to consider those factors.
At the top level, the brand was honing in on the geographic breakdown of each influencer’s audience. The brand didn’t want to “waste” money on a non-US audience for a North American product launch. Perfectly logical. And worth noting that this line of investigation and potentially avoiding a problem has been made possible by the incredible data available in the best discovery platforms available today (gen.video isn’t really in the discovery business, we license tools to get this data, so this is not a self-promotion). Given that, two things jumped out at me before we could even get to thinking about this problem more deeply:
- Audience measures (at least the one we were looking at) look at subscribers or followers, not viewers. Yes, it is intuitive to believe that audience and viewer characteristics are going to be correlated but in a world where the average “view-through” rate is 5 – 15%, there’s also a pretty good argument that viewers of a post might be materially different demographically then the overall subscriber base. In fact, I think we can almost assume that the social media platforms are really *REALLY* good at targeting a specific post to the relevant segment of the follower base.These algorithmic nuances are the addictive special sauce that keep us coming back for more and why the winning platforms have “won.”
- Fear of monetary waste is a very human characteristic that often leads to irrational decisions. Humans are notoriously wasteful ordering in restaurants and with other material things. Yet at the same time many of us, finding out that we have spent a dollar more for something at one grocery store than it was on special somewhere else will drive back to the first grocery store, return the item and go buy it at the second store…despite the loss of time and burning of expensive gas. In the example above, an audience that’s 50% outside the US sounds very wasteful but since we weren’t even talking about the costs, that’s a premature assessment.
The simple thing to do here to take out the emotion and the increasingly desperate search for “unicorns” by shrinking down the reach of each influencer based on the percentage of the audience with the right criteria. If we only care about followers in the United States simply multiply the total subscriber base by the percentage that comes from the US. We are also targeting women 25 – 44 only? Great! Take the geographically “reduced” audience and multiply that times the percent of the audience by the 25 - 44 segment. This math can be extreme: we might take 1 million followers down to 100,000 by the time we’re done but that’s ok so long as we apply the rules evenly to all of the influencers.
To be clear, this is not a perfect science as it includes two huge assumptions that are unlikely to be precisely true:
1. We’re assuming the data from the discovery platform is accurate,
2. We are assuming all of the attributes are evenly distributed in the population…i.e.., there are the same gender and age distributions in each country.
There is also a danger that this analysis will not only heighten the client’s ‘aversion to waste’ but will drive them to want to abandon an organic influencer campaign in favor of a targeted paid media campaign. Here’s how I would approach that conversation:
- This is an exercise in bringing more quantitative rigor to a facet of influencer marketing that can benefit from more rigor. But it needs to be ‘married’ to the actual results that the influencers have been able to demonstrate in other campaigns in terms of quality content, engagement from a relevant audience (by reading the comments and looking at the profiles of commenters you can gain a qualitative lens on attractiveness), and click level behaviors. The math is a signal, not a prescription
- That said, if everyone does become convinced that the applicable audiences really are whittled down to the point where the campaign doesn’t look interesting, maybe the client should do a paid media campaign (potentially still using content created by the identified influencers). Effective targeting is, to my mind, the single most compelling reason to include or even favor paid impressions over organic when using influencer content.
Another interesting opportunity raised by this approach is that in theory, you could go back to the influencers and say, “hey, the brand is really interested in working with you. However, a lot of your audience is outside our targeting and so it is making the economics difficult. No doubt you are worth every penny of your rate card in most situations, but can you be flexible here given this dynamic?”
Here’s A Graphic Of How This Mathematical Exercise Would Work:
In each column, I’ve highlighted the three most attractive influencers based on that single attribute or criteria. I’ve sorted the sheet based on the three lowest basic CPMs (in reality, we’d also at costs based on estimated impressions, but for simplicity and consistency let’s stick to the audience criteria). Then using the method above, we systematically reduce the audience by the relevant percentages of each criteria (advanced idea: you can also weigh these criteria if they are not equally important to the client. If not obvious how to do that, feel free to ask in the comments or write to me privately). As you can see, by the end none of the influencers has ‘retained’ more than 40% of their audience. Now we recalculate the CPMs and two of the three original “top picks” would in theory get swapped.
But saying it that strongly makes even me uncomfortable. Rather, I’d like to see you look more closely at Elizabeth and Elaine. Does this push them into the top three? For Elizabeth, I’d say heck yes: she’s the biggest influencer of all and scores so highly on both the US and female audiences that she becomes sort of obvious. But for Elaine, I think it just puts her in play but unless her content is amazing, this signal may not be strong enough to make a difference.
In short, the math really just simplifies one of the myriad variables that go into these decisions and hopefully makes it easier to make a great decision with confidence. Good luck!