A Defense in the War on AdTech: Case 1
Why the need for AdTech is very much alive, especially for content strategists
Last week, I came across a piece by Michael Eisenberg (his Medium blog Michael A. Eisenberg: Six Kids And A Full Time Job is phenomenal), principal investor at Aleph, a venture firm that invests in the blossoming Israeli startup scene. It was written in April of last year, though is still a relevant discussion, if not more so now than then. In it, he calls on all Israeli engineers to stop wasting their time with AdTech and move on to other industries.
I disagree profoundly, and I’ll tell you why below. However, I preface with my unwavering admiration for Michael Eisenberg and his industry perspective. He’s a remarkably successful investor by all objective measures, all while maintaining mindful humility — a trait lacking, in my experience, among too many prosperous financiers. I know this because I read his blog posts on Medium regularly, and have twice been part of a young, doe-eyed audience to whom he’s earnestly and generously spoken about investing, his thoughts on the future of industry, and his professional experience.
With that, here’s his argument in broad strokes:
- AdTech is a business of competing for someone else’s audience’s attention, which can be unpredictable.
- An AdTech company’s margins rely on that unpredictable traffic, which creates risk.
- AdTech companies will be forced to compete aggressively for ad real-estate in order to keep margins afloat on traffic they do not own.
This could all be true of certain AdTech technologies built upon certain business models. But that doesn’t mean AdTech is a bad place for engineers. Let’s start with a big point to note; Israeli engineers on the brink of quits, know this:
AdTech is not just ad placement technology.
Arbor Advisors, Silicon Valley’s go-to investment bank for mergers and acquisitions, published an in-depth, quantitative analysis of the AdTech industry around this time last year, covering 2017 trends (I suspect they intend to do the same for 2018 trends, perhaps still gathering the data). Their analysis of the AdTech market included players as diverse as Contently, Google AdWords, Facebook, and Sizmek. Among these four tools, each one offers radically different services from another. One creates content, the other promotes it on search engines, the other manages placement in a social network platform itself, another enables mobile reach.
These companies are alive and thriving. Content production, MarTech, AdTech, and social media platforms are all behemoths in their own right, and each of the above categories rely on the other to create a smooth, cohesive, consumer journey that ultimately sells products and services.
Mad Men to Math Men
I’m a faithful believer that data analysis and AI can never replace the creative imagination. Poetic prose, comedic commercials, amusing display ads — they all share a sense of humanness, which is why they objectively perform so well. But when it comes to advertising today, it certainly doesn’t hurt to inform this creative process with data as much as possible. We need AdTech to do this.
Here are some AdTech tools that I use all the time, if we stick with a broader definition of the industry. For example:
- A keyword tool to indicate the domains that have the best ranking for keywords the client wants to rank for.
- A topical analysis tool that indicates what competitor content had the most shares
- A tool that indicates what my clients’ competitors dark ads are, across Facebook, LinkedIn, Instagram, Google, Twitter, and Display
- A tool that identifies which publications certain demographics and psychographics consume
- A tool to publish on other publication’s sites.
- A tool from which I can manage ad publication, budget, and distribution on a social media platform.
- A tool for retargeting the same ads to people who’ve engaged, clicked, or interacted with content in some way.
I use a lot of others, but these five tools alone offer incredible insight into how my clients can engage with their potential customers. If I were to plan, create and implement an entire content strategy, I would use the tools above in that order, by gleaning insight from each and then applying it to the next stage of the process. These tools all fall into the AdTech bucket according to many industry leaders, and I would sure be lost without them.
I’d like to vouch for number 5 though, because it’s this sort of placement tool Mr. Eisenberg argues against in the narrower definition of AdTech.
Why I Need Ad Placement Tech, Not Facebook Audience Insights to Discover My Audience
With all these tools, though, there’s a big gaping hole in the content strategy corner of the world. There is nothing (under $30K/year) that can help me figure out before running a single ad who my audience is, what they like to read, what they’re likely to click on, etc., other than Facebook Audience Insights (ish). I’ve looked a lot.
Let’s assume I’m the advertiser. Ad placement technologies can tell me if a highly niche audience likes my product’s ads. That niche audience is defined by the audience browsing a specific publication’s website. By publishing in a range of online publications, I can learn, for example, “Wow, my product performs really well among people interested in running, and I know this because people interested in running read Runner’s World and click on my ads a lot there. Weight lifters, however, are not so interested in my product, and I know this because people who read Muscle Development didn’t click on those same ads.”
Therefore, an advertiser needs ad placement tech to test their audience and ad creatives, because by definition, a publication’s audience is defined by interest in a topic, not by usage of a platform.
(I’ll state this again after some explanation, but that’s the crux of why we need insight that only ad placement tech can provide.)
Case Study: Testing an Audience for a Plant-Based Facial Serum
Let’s delve more into why I, a content marketer, use and need these ad placement technologies. Let’s say I have a client who produces all-natural facial moisturizing serums for women (I don’t, but I do have clients in the natural beauty and health category), I, the “advertiser”, will distribute the ads using an “ad placement technology” company like Taboola, Outbrain, or Revcontent (companies Eisenberg explicitly tells engineers to run from), in a “publisher” or “publication” such as Women’s Health Magazine, InStyle, Allure, and O Magazine by Oprah.
I would do this because each publication covers specific subjects, and therefore owns a unique loyal audience. These audiences are interested in health, fashion, glamour, and lifestyle/culture, respectively. Knowing that:
Each publication’s audience group may share an interest unique enough from the other audience group that leads them to buying the serum (and into the brand of the serum) at a statically significantly higher rate than other audience groups.
I might discover that health people who read Women’s Health really really love my serum because of its natural properties, and beauty people who read Allure don’t care and won’t buy it; perhaps they’re already familiar with too many serums and the barrier to entry is too high. Now I know: go after the health world!
Here you might be thinking, “There are definitely other ways to test audiences against creatives than using native ad placement.” Let’s probe.
How else would I be able to determine whether these four niche groups — health, fashion, beauty, and lifestyle/culture — respond to ads for my client’s serum?
Facebook Audience Insights comes to mind. For those unfamiliar with how Facebook advertising works, when I make an ad for a product, say this serum, I can build a new audience to whom I would like to deliver that ad. I can build it using the Facebook Audience Insights product, or in the Facebook Ad Manager Platform itself. For example, if I were to create the health audience, I can select people who are female, aged 18–34, and who are interested in Women’s Health Magazine and other proxies that demonstrate interest in health, such as healthy brands like Kayla Itsines or Weight Watchers, perhaps. I can also tell Facebook to target to people “interested in women’s health,” as a broad command. Next, I’d make the fashion audience, selecting the same demographics and change selected interests to people who Facebook deems are interested in Vogue, ShopBop, and Revolve, and so on and so forth for each category.
But here’s the problem with using Facebook to find audiences who might like your product’s ads and by extension your product.
Facebook determines and measures an audience interest group only by using Facebook Page Likes. It does not take into account Instagram following and followers, a user’s commenting and post liking — nothing.
So, even though you think you’re discovering how effective your ad is among health buffs, what you’re really discovering is a subset of that: how effective your ad is among health buffs who necessarily “Like” health-related pages on Facebook.
For my hypothetical clients’ serum, which may have a highly aesthetic brand, I have an inkling that my audience is also engaged with health-related things on Instagram, Twitter, and Pinterest.
Here’s an example. The Minimalist Baker is a plant-based foodie Instagram account. Let’s say I’m testing to see if vegans and those interested in plant-based things (a subset of my health audience) were interested in my client’s plant-based serum. If I told Facebook Audience Insights I wanted to run ads to people “interested in veganism”, my ads wouldn’t be served to people who follow the Minimalist Baker. Why? Because the Facebook following is mediocre (compared to other publications). Meanwhile, the Minimalist Baker, a vegan, plant-based foodie account, has 1.4 million Instagram followers. These Insta followers might really like my new product, but I wouldn’t know because my tester ads never got delivered to them.
This leaves me with a huge blind spot, with false insight. Either I could end up with a false positive, where I’d think that health nuts generally are definitely interested in this serum, even though it’s just people who Like the Minimalist Baker on Facebook. Alternatively, I could end up with a false negative, where I’d think that health nuts as a whole are not interested in this serum, even though it’s just the people who like the Minimalist Baker that are not interested in the serum.
So, how do I test an audience that may very well use a range of social platforms, if I’m only getting insight to the audience that uses one platform in one way (likes)?
Ta da! In walks Ad placement tech.
I need ad placement tech to test my audience and ad creatives, because by definition, a publication’s audience is defined by their interest in a topic, not by their usage of a platform.
I can and should publish the serum ads on Minimalist Baker, Women’s Health Magazine, and other health publications too to get broader insight on the health buff audience. I’d do the same for the other categories, which we said was fashion, beauty, and lifestyle. With this, I can accurately measure how responsive the health audience is against my other categories. Great.
If not through a publication’s audience that Ad placement tech gives me access to, and not Facebook Audience Insights, how else could a brand like my hypothetical serum identify and discover their audience? Analysis of Instagram following before running ads? Instagram is now just barely coming out with a more efficient analytics platform, and even still it wont break down followers by interest or type. Sprout Social and Iconosquare are the biggest listening tools, but they wouldn’t really give an audience breakdown either. The only thing I can think of to effectively identify an audience would be using Crimson Hexagon’s audience tools; their tech is extremely powerful, but also extremely expensive, well out of the range of a new company testing its audiences. Rand Fishkin’s new venture, SparkToro, is aiming to in part solve that problem by creating a platform (also S/O, we’re distantly related through marriage).
Still, it will be hard for any audience discovery tool to identify a psychographic group as niche as the literal audience of a magazine, and also enable advertising access to them. Only ad placement tech companies can do this.
So, while ad placement technologies were originally invented to generate product sales, they have the potential to serve even better as an insight tool for strategists like me. All to say:
Content strategists can learn a lot from ad placement technologies, because there’s just not a whole lot of audience insights tools out there that cost less than $30K/year for a subscription (cough cough, Crimson Hexagon). So, we need ad placement tech. Thank God for ad placement technologies. Engineers, we need you working on ad placement technologies!!!!!
TLDR: Ad Placement Tech is a great way to test really niche audiences for product ads. Facebook’s Audience Network is flawed.
The Elephant in the Room
I’ve demonstrated above: ad placement technologies can be invaluably useful for content strategists, marketers, and advertisers. But there’s an obvious elephant in the room: bottom-placement ads are usually really, really bad.
Half of these ads look like an ugly banner display ad for Toys R’ Us that was on my 2002 AOL Kids Desktop browser (that’s a millennial burn, thank you). Users that offer valuable insight into the demographic and psychographic of a potential customer base today will not click on a low quality ad, and by extension, any ad next to it. So those bad quality advertisers ruin the chance for good creatives to perform. In the case above, I actually think Monday.com’s ad is quite good and I’d click on it if I were working with Gantt charts on the reg’. Progressive, poor job indeed. Way to ruin it for everyone else, Flo.
But the ad placement companies are fixing this problem. I spoke with an account manager at Taboola a couple of weeks ago while running one of my campaigns, and she told me their automated and human ad review process is becoming stricter, forcing all advertisers to deliver better user experience.
This is why we need more engineers working at AdTech companies, not less.
With the limitations of Facebook Audience Insights and the clear-cut audience you know you’re getting from a publication, I hope I’ve demonstrated that the need for ad placement technology is very much still present from a strategy perspective. There’s an audience willing and excited to pay for ad tech.
But even if you’re not a content strategist, there’s still a looming gray cloud over the Facebook and Google networks for every single advertiser in the world. Drastic, I know, but it’s true and scary. In my next piece in the Defense of AdTech series, I covered the inherent dangers for advertisers, publishers, and customers when we put all our eggs in one attention basket. Keep reading here.
This piece was published by Aya Abitbul at Studio 96, a firm that produces content for meaningful brands at scale. Check out our company here.