You’re Doing It Wrong
The Problem with Programmatic Advertising
Programmatic advertising is for lazy marketers who don’t really care about the reputation of the brands they represent.
I don’t want to say that. I want to believe in the promise. Just think, all that scale with all that personalization! But believing programmatic advertising is the answer to the age-old advertising conundrum of delivering the right message to the right person at the right time, is like believing the Keystone XL pipeline will solve the Venezuelan inflation crisis.
Perhaps the strongest evidence indicating a shortcoming with programmatic lies in the fact that Facebook and Google (both of whom do not truly participate in a programmatic advertising) earn over 60% of digital advertising revenue (which is still only a small portion of overall marketing budgets). Even more indicting, there are reports that claim 100% of digital ad growth last year was driven by this duopoly. It’s not coincidence, it’s simply a matter of getting better results.
What is Programmatic Advertising?
There are several shades of complexity to the definition of “programmatic advertising,” but for the purposes of this article let us say that programmatic advertising is any digital ad purchasing done via a portal to a distributed network of advertising websites, like CNN, Galoremag, Breitbart.com, but excluding closed networks like Google Adwords or Facebook. Digiday provides a pretty good explanation about programmatic advertising if you’d like to dive a little deeper down that rabbit hole.
The appeal of this system is that advertisers can use one portal to buy digital advertising at a set budget across thousands of different websites, while still being able to segment an audience based on target demographic preferences like location, gender or age.
Over the past three years, digital advertisers have moved towards programmatic ad buying in droves. Several sources estimate 84% of all digital marketing is going to be done programmatically in 2017 (although this is a misleading statistic, since that estimate includes ad purchasing on Facebook and Google).
The media, powered by programmatic PR dollars, is just as eager to encourage this behavior. Every few weeks, another puff piece is published extolling the advantages of programmatic ad buys, and so spins the wheel.
Recently, Hallmark shared some positive results about their programmatic advertising, notably a substantial decrease in their ad campaigns “costs per thousand” (the most common formula for trading digital ads, more commonly referred to as “CPMs”).
Theoretically, this should be the result ofa low-cost, competitive supply chain that delivers the right ad to the right person at the right time, but–
that’s rarely the actual result.
The False Promise of Programmatic
Hallmark’s decrease in CPMs is most likely not due to the programmatic bidding network, but rather several cheats and shortcuts that underlie the digital ad landscape.
The first is a tendency for networks to depress their CPMs by displaying ads on low-value sites, rather than targeting more specific buyers with less impressions.
You know these sites. They are the kind you land on if you click one of those paid content links at the end of an article.
Should you actually navigate to one of these low-value sites, you’ll see they are covered in ads, like stickers in the bathroom of a dive bar, and generally only have a sentence or two on each page.
The Broken Supply Chain that is Programmatic Advertising
The article also mentions the value of 1st party data– that is, the information a company has about their customers, such as the products they purchased, or the dates they made their purchase.
First-party point-of-sale data is changing the economics of targeting individuals at scale. If a mother buys a birthday card for her son or an anniversary card for her husband at a Hallmark store, for example, Hallmark can begin building a time-based element into its targeting strategy.
This is great in concept, but it is incredibly challenging to execute successfully. Not only does Hallmark have to reliably store and consistently attribute data to the right customer, it also faces the enormous challenge of passing all that data through a fragmented system of serviced providers, exchanges, attribution validators, DSP, SSPs and a bunch of other acronyms that make up the “display advertising technology landscape.”
All of these components are meant to provide a valid service (and many do) but at each “value add” hand-off, there is the opportunity for things to get lost. Databases don’t match up. APIs aren’t integrated properly. Improper vendors are chosen. The result is a supply chain so disorganized that the all-important 1st party data is often dropped on the floor before it can be applied to the ad an individual is being shown.
The combination of these two factors means that instead of delivering highly targeted messaging to a captive, engaged audience, brands end up showing up in some strange places.
Lookalikes: How the Internet knows you but it doesn’t really know you
Additionally, programmatic advertising depends on matching the demographic information from an advertiser’s 1st party data to a certain segment of the audience on a publisher’s website, a practice known as “lookalike targeting.”
This in principal isn’t a problem, since advertising campaigns tend to target demographics rather than individuals anyways, but the data networks use to establish these “lookalike” comparisons are sketchy at best. For instance, it’s not uncommon that a network will classify a user by gender or age based on the type of apps they use, rather than actual user information. That means that if you’re trying to deliver a message to him:
You might end up showing an ad to him instead:
The Almighty $
There are only 2 ways to drive down the cost of digital advertising, improve your targeting or expand your inventory. Because there is an ever-expanding universe of advertisers and publishers, it is far easier for programmatic networks to invest in the latter rather than the former, and that is what they have done, continually including shadier and shadier websites in their network.
Much of this expansion has gone unbeknownst to their clients, which is why so many brands were “outraged” to learn their ads were showing up on sites like Breitbart (an epiphany that speaks more to the laziness of marketers than the shortcomings of the technology). And while it’s true that advertisers can try to circumvent this problem with whitelist/blacklists, such a tactic does not address the core issue, is hard to manage, and creates a similar level of manual effort that programmatic was suppose to solve in the first place.
The New SAAS: Stories-as-a-Service
There is an alternative marketing tactic out there:
Stories as a Service, a way of marketing in which brands hone their particular area of domain expertise through dedicated channels, rather than hawk their wares to the masses like a bleacher vendor.
The underlying premise of Stories as a Service is that while content is still king, the consumer, overwhelmed with an ever-expanding array of content, and unable to distinguish which content is true and which is fake, will choose instead trust only a select few sources.
As such, advertising becomes less important than grooming an engaged user base that trusts your content and seeks you out for your particular domain expertise.
We can already see evidence of this strategy taking root. In Q1 of 2017 the New York Times posted their largest growth ever in digital subscriptions, which helped buoy a 7% decline in advertising over the same time period to give the company a 5% YOY revenue increase.
Brands that are engaging in Stories-as-a-Service generally rely on closed distribution systems, where the entire consumer experience is controlled by the brand. This means that these companies must take the time to actually know where their ads are showing up (if they advertise at all), and what the overall consumer experience of those ads are. This is why the vast majority of digital advertising dollars are spent on Google and Facebook, two publishers who have uniquely custom and comprehensive data about users. While this doesn’t necessarily allow for the same scale and cheap inventory as programmatic, it undoubtedly does a better job preserving brand integrity, and makes for a better consumer experience.
It’s a model similar to what Disney has employed over the last 10 years in developing and producing their IP. Think about all those Marvel and Pixar movies that have come out recently — from casting the next Spider-man to producing the souvenir plush dolls — Disney owns the entire damn chain.
This tactic has led to remarkable growth within the company and is the result of many things, but a key component is adherence to 3 unique principles:
- Be specific about who you partner with.
- Build custom engagement experiences.
- Take the time to know where and how your message is being consumed.
Can’t we all just get along?
Programmatic advertising is not all bad, but it should not be trumpeted as the panacea that some make it out to be. Many marketers are too quick to wave their hands at the integrity of programmatic ad buying, but knowing the mechanics behind each piece of technology in your tech stack, and understanding how each component relates to the greater whole is critical to a successful digital marketing strategy. After all, it is the deliberate, mindful application of technology that will yield better results than what any single bidding system can provide.
