Nobody Knows How the Internet Works and it Shows.
On the Absurdity of Precision in Digital Marketing
There are only three ways for internet-focused firms to make money: services, entertainment, and advertising.
The first route to revenue consists of actually doing something that creates value for other companies. You can think here of Shopify, Salesforce, and AWS. Sure, they are not stereotypical ‘internet’ operators but their services and products require an internet connection to work. The revenues here consist of a subset of someone else revenues in the real economy. So long as the unit economics work out there’s enough money to go around for everyone. Classic vendor-buyer relationship.
The second grouping consists of the likes of Netflix, Spotify, Steam, and even something like Xbox Live. These offerings help numb the pain of existence and wile away the hours in the day. Folks are happy to pay for pleasure and that’s where these folks make their moolah.
I am not going to talk about these first two groups though. Instead, I’m going to focus on digital advertising. Why? Because it’s an absolute shitshow of an industry.
At its core, digital advertising is an oxymoron in terms of personnel. ‘Digital’ denotes the technological medium of the field. Without the metaphorical cogs and pipes of the internet, you cannot have digital advertising. Therefore, the bones of the industry consist of technical products and solutions created by STEM professionals. By contrast, the ‘advertising’ part of the industry consists of folks who overwhelmingly went to school to be Madmen-esque whiz-bang professionals.
They dreamt of drinking highballs in their corner offices and rubbing elbows with the creative directors of prestigious companies. In reality, these folks spend their days doom scrolling through spreadsheets hoping that they can bill hours despite the complete glazing over of their eyes.
To be fair, there are creative types in the digital marketing space who come up with shiny pictures and nice words that are the advertisements themselves. However, those folks are dwarfed in number by the media planning people. The planning crews exist in-house at brands and at external agencies. Typically the agencies do the ‘planning’ and buying of ad space while the in-house folks act as a QA/QC team.
The planning of digital advertising is a shit simple process. There are vendors – in the form of platforms – who have the ability to place creative assets in front of a target audience’s eyeballs. A media planner decides for, or recommends to, a client which vendors they should use and when. Once selected and contracted, the planners then track vendor performance and will haggle over missed engagement guarantees. Data will flow between the vendors, agencies, and clients and eventually everyone agrees that something of value happened.
It's a simple enough setup right? Not quite. Why? Because the personnel disconnect mentioned above means that nobody at the agencies – and at the client brands – ultimately has much of a clue about how the system actually works. Nor is the industry set up in such a way as to recognize the nature of the business.
Allow me to walk you through digital advertising from the perspective of an ad. An ad is created by the creative agency and turned into a trackable unit. Typically, the unit is tracked using a platform like Google Analytics (GA). The ad sits on a vendor platform for a given set of flight dates and is then killed (i.e. turned off).
GA allows a media planner at an agency and/or client to see the performance in relatively real time. In theory, this should be fine and dandy except that for non-technical people there’s a gap between reality and their perception thereof.
In reality, the internet is a very busy place. If you ever have the opportunity to look at the backend of a platform’s webpage you’ll see a lot of traffic all the time. Some of this traffic is legitimate human stuff, but there are also a lot of bots out there in the wild. Not all bots are malicious. Some are from Google, Amazon, Facebook, et al., and their job is to index sites for search engines and big data shit. However, some bots are less than benign. Anyways, a platform like GA will filter out some of these bots. That is because either they recognize them as bots or the bots themselves self-identify as such. Nothing crazy really but the process is certainly not perfect. There’s the chance of real eyeballs being filtered out and non-self-identifying bots being left in.
The reality of the situation though is that 99% of media planners do not know this. They think that the GA dashboard they see is 100% of all activity. Additionally, when GA does tell them there are bots it’s not overly helpful because either GA has already removed them from its counts or the presence of bots does not negate the presence of real people.
If a planner wants 100 ‘real’ engagements with their assets and GA is filtering out bot traffic from its counts then it should not matter that in reality, those 100 engagements occurred alongside 1000 bot engagements. Nonetheless, the media planners begin to shit their pants when they hear anything related to bots. Why? Because as soon as someone says ‘bots’ a marketer hears ‘fraud’ – even though in the scenario above, no fraud has occurred whatsoever. It’s just how the sausage gets made.
A way that media planners ‘try’ to get away from the possibility of bots is through the use of what the industry calls ‘walled gardens.’ These are platforms that are not immediately open to the world. Think of Facebook or LinkedIn. To one extent or another, you have to be a ‘real boy’ to access these platforms in their entirety. One upside is that you then get added demographic information. At one point or another, we all gave our age, gender, and location to Facebook. Additionally, these platforms track user behavior and can derive further information – such as whether or not you’re interested in Italian ties – that aid in shaping targetable audiences.
The problem with walled gardens is that they control the data. If you want to show an ad on Facebook, then Facebook is going to tell you how it did. The incentives are entirely skewed towards Facebook making itself look good by any means possible. Therefore, it should come as no surprise that Zuck’s Frankenstein has been repeatedly caught inflating ad numbers.
The trade-off is that while walled gardens allow you to target your audience in some allegedly meaningful capacity, you have to take the platforms at their word. This leads to an incredible conflict of interest.
The platforms are in the business of selling advertising performance. The media planners are in the business of purchasing performance on behalf of their clients. The clients, in turn, have internal media teams who are in the business of showing that they are actively engaging the end consumer. If a platform inflates performance figures no one in this chain of media is incentivized to say anything. The platform will continue to receive media spend, the planner looks like they are great at their job, and the internal team can go to senior leadership and say that they’re knocking it out of the park. It’s a perfect storm for malfeasance. Who watches the Watchmen? Nobody. That’s who.
If you are a digital marketer you’re essentially stuck between the fear of bots – regardless of their impact on activity – and walled gardens that are incentivized to lie to you. What’s wild – to me at least – is that digital marketers run towards the walled platforms. They are happy to send money to these platforms because of the calming story that they tell. Who is going to get fired for advertising on Facebook? No one. Even if the numbers are inflated there is real activity underlying the false numbers. On top of that, Facebook is not going to scare them by mentioning bots. We all know that the platforms are full of them but marketers look the other way because the numbers just look that good.
Digital marketers simply want to insert numbers into spreadsheets so as to satisfy their clients. If Facebook et al., are willing to provide the right numbers then the marketers are happy to pay for them.
All this points to an underlying problem which is that nobody knows how the internet works. You have digital marketers who pay their rent by pretending to be experts in the tech space yet cannot articulate the processes underlying the data they hang their hats on. The fact of the matter is that any platform, whether it be digital, print, television, or in-person, comes with unresolvable ambiguity.
Sure, Nielson will give you an estimate of how many people watched NCIS last week but they can’t tell you whether or not they actually saw a given ad, cared to pay attention, or were just drunk on the couch. Ultimately, the data used for advertising purposes is a defensible guess given a laundry list of caveats. No one – at least no one should – imagines that Nielson ratings are scientifically precise. Instead, they are relatively agreeable. It might be shit, but it’s industry-recognized shit.
The problem with digital platforms is that they provide a false sense of precision and accuracy. GA gives you a particular picture of user activity. It filters out the less believable aspects, but it’s a best guess using algorithmic approaches. Similarly, walled platforms know that they have advertisers’ balls in a vise grip. “You want access to our audience? Okay, but you’re going to pay what we tell you to.” It’s the equivalent of paying a carpenter to build you an up-to-code house but not being able to send an inspector in to confirm compliance.
All this might seem trivial – it is just digital marketing – but worldwide companies spend upwards of $ 180 billion a year on this shit in the US alone. It’s a big fucking industry. And for what? Numbers that are too precise to be accurate and metrics too vague to be meaningful.
The peak absurdity of this situation was exhibited by a senior partner at a major marketing agency who came across my LinkedIn feed. Long story short someone had put out a white paper comparing marketing spend changes and share price movements. Admittedly, not a great comparison given inefficiencies in market dynamics and ulterior drivers of share price outside of sales. Not to mention that marketing efforts can have long tail lag times. However, it’s not entirely ridiculous to expect that [assuming reasonable quality] more marketing efforts should lead to an increase in sales which – in one form or another – should lead to a higher share price.
Instead of taking the work for what it was the marketing partner railed about how the role of a marketer was to engage an audience, shape a corporate brand, and capture mindshare in the marketplace as opposed to drive sales and thereby boost a share price. What the partner had deluded themselves into believing was that marketing exists for its own sake and its own metrics. For this person, so long as the ‘objective and scientific’ engagement data looks good, then marketing has done its job. Call me crazy, but I can’t pay my rent with engagement data.
I’ve been told that I should include tangible takeaways in these pieces. Insofar as there is one here it is that you should pay your salespeople handsomely for performance and treat the marketing team as second stringers too scared to get their hands dirty in the business of making the money that facilitates their lifestyles.