If you’re a New Yorker, you know what I’m talking about. If you’re not – trust me, it’s a thing. Minimalist styling, single-tone backgrounds, basic animations, and near-identical fonts.
For companies interested in advertising in NYC’s subways, the price tag for the most modest ad campaign starts at $250,000 and goes up significantly from there depending on all the encouraged frills, add-ons and full cart takeovers.
What you’ll immediately notice is that some of Silicon Valley’s darling unicorns from the technology and Consumer Packaged Goods sectors (Monday, Seamless, Casper, Ro, Brex, Spotify, etc.) jump headfirst into spending almost half a million on these physical ad campaigns. This is very peculiar and unusual – in almost no other physical ad inventories do you see digitally native brands competing so fiercely and bidding so high for physical ad space.
These type of direct-to-consumer venture-backed startups are encouraged (by their investors) to invest aggressively into growth... Usually too aggressively. The majority of this investment ends up getting funneled into ad spend for digital advertising - that’s how almost forty cents out of every venture dollar ends up in the hands of Facebook, Google, and Amazon (according to Chamath Palihapitiya).
However, we know NYC’s subway systems get in front of one of the most lucrative and ‘relevant’ audience groups in the world, and there’s certainly something to be said for the fact that having a presence underground comes with a hefty ego boost in front of all the others – one of those hallmark ‘we made it’ signals to the rest of the community.
So let’s get into it: here’s what we know for sure – the success of your ad campaign depends overwhelmingly on what your ads look like. What does the ad’s creative show and say about your product/service? How does it do that? How is the ad styled? Etc.
As marketers, we know that even small, nuanced changes in an ad’s creative – like changing the background color, or replacing a palm tree with a cactus – can have a massive effect on an ad’s performance. In fact, it almost always does.
This observation poses the leading question - how are these decisions made? How do companies decide what their ads look like? Here’s a fun one – next time you meet a business owner, ask them this – ‘What information do you use to determine what your ads look like?’ The answers, or potentially the lack thereof, might surprise you. Today’s process for producing advertising creative looks a lot like it did in the Mad Men era fifty years ago, and it’s certainly not the fault of the creatives.
I used to look at ads and think they were masterminded in ways that were beyond the conscious grasp of the typical consumer. When I started a digital marketing agency, we were given the mandate to create ads. Here’s how we did it:
Step 1: Creative Prompt. What I did was hire some of the best designers/creatives I could find, and I’d give them loose prompts with some guidelines and told them to go. They would turn around work that I usually thought looked great. There is no systematic process here – the job of the creative is to be a taste-maker that creates something novel and captivating based on cultural context and the task at hand.
Step 2: Feedback. We would then present this work to the client and gauge their feedback. Almost always, the client would have opinionated commentary and propose edits. This would kick off an iterative process where we would go through several versions until we reach an end-product we both like. This feedback loop is usually just as arbitrary and rooted in personal preference/taste as Step 1.
Step 3: Launch & Learn. We would next launch the ad and see how it performed. If it performed poorly, we could take guesses as to why it didn’t resonate, but we didn’t really know why for sure. We could say ‘people just didn’t like that message’ or ‘there were some negative comments about that actor’s man bun,’ but nothing truly conclusive. If it performed incredibly well, we were happy, but still as much in the dark as to why. All we would really learn was if it did well or not – not why.
This is what the creative production process looked like fifty years ago, and primarily what it looks like today. While an agency iterates back and forth, the client pays for time. This process is time consuming (and as a result, expensive) due to the inherent built-in frictions between technical talent and creative talent, represented here in one of the visuals we created internally:
The frictions here are a result of seemly opposite modus operandi for the two primary stakeholder groups involved in this process. We’ll call this ‘Technical-Creative Friction,’ and it’s an age old business problem. It might not be surprising to learn that the number one reason why clients break up with their agencies is due to dissatisfaction with the quality of their creative.
The Technical Talent refers to the execs and business people who see the financial need to make good ads, commission the creative for the ads, and ultimately end up buying and investing in the end-product.
The Creative Talent refers to the taste-makers responsible for making good ads. Their world is filled with constraints – they often work under tight budgets, resources, and exceedingly vague expectations. There exists a substantial language barrier between the Creatives and the Technical folks, wherein ‘the Technicians’ want decisions rationalized in the language of data or money, but the Creatives can’t use those measured, objective terms to explain their aesthetic decisions.
Leonardo didn’t have to justify to investors why Mona Lisa’s arms were folded as such, but we can surmise that his paintings would look a bit different if he had to justify every brushstroke to a counterpart who’s task it was to sell paintings.
So what happens in a world where creatives are forced to rationalize their work to ‘the technicians’? You get an ecosystem of creatives that end up looking over the shoulder of those who everyone assumes have the most successful campaign or best company.
This copycat mentally is defensible for both sides. For instance, if a successful company’s ads have minimalist styling, you must think that those folks either did a lot of homework to arrive at that decision or they got lucky. As a business decision-maker, why spend your own resources or roll the dice? You could just imitate their ads. For creatives, it’s rarely worth the above-and-beyond energy and the risk involved in convincing businesses to do something different because if your bet is wrong (if your creative fails to perform well), then your reputation takes a big hit in an industry where credibility and notches on belts are valued the most. If the ‘copycat creative’ doesn’t perform well, all the stakeholders involved can save face and point the finger of blame elsewhere.
NYC’s subway ads are an outcome of this system of friction. This friction has existed since the beginning of the mass advertising era - it’s the same reason why all the alcohol and tobacco ads in the 60’s looked the same as well. In high-stakes advertising environments, you typically see ‘The Subway Effect’ emerge.
The billions of dollars that are funneled into producing and promoting ad creative (to exceed $285 billion by 2020) is significantly stifled by the lack of… creativity. The lack of experimentation and tinkering in advertising creative leads to plateauing results, which is deadly for brands in a world where advertising creative is such a fragile and quickly-expiring asset that’s so expensive to produce.
It’s clear to folks in the industry that this dinosaur system needs to be overhauled, which is why CMO’s are very vocal about their demand for technology like AI, real-time ad assembly/versioning, multivariate creative testing, and dynamic creative optimization. These tools, only recently made feasible at scale through a much lower threshold of accessibility to machine learning software, hold the promise to solving this problem by arming creatives with creative data.
Creative data is a new emerging category of business intelligence that informs companies on what their best creative will look like in the same way that audience data informs companies on what their best audience segments look like. This type of data is unique to each brand and product, so it must be created in a bespoke way for companies. I initially built Marpipe to solve my own problems as an agency owner – by collecting creative data for our clients, we were able to solve the issue of Technical-Creative Friction by implementing a 100% data-driven approach to creative. No more arbitrary aesthetic decisions or taste-making – everything we put forth was justified through testing results, and we knew it would work. Our creatives were extremely empowered through data, and our technical folks were happy with the new scaleable system driven by performance metrics.
Since then, we’ve completely pivoted to an enterprise software business (Marpipe) because the demand for creative data is overwhelming and driving rapid growth (primarily through word-of-mouth, we have yet to activate scaleable growth channels). How it works is simple – by isolating visual variables within creative (like objects, backgrounds, people, etc.) and testing them through sound experimental design, we can quickly identify the components of the visual (in videos and images) that impact engagement the most. As a result, brands can learn exactly what their best creative should (and shouldn’t) look like, but beyond that, they’ll never have to worry about where their next best-performing piece of creative is going to come from ever again.
Through our data gathering, we find that a lot of the creative assumptions that brands make through ‘the subway effect’ are associated with negative performance once measured against many other variants. For instance, a popular food brand we work with overhauled their ad campaigns and their website once they learned that their styling assumptions were linked with negative performance and they discovered an alternative that was a breakout performer. If all the companies advertising on the NYC subway were driven by their own creative data, all of their ads would look completely different and each company’s ads would yield better performance – creating a much healthier advertising ecosystem devoid of friction and aspiring to superior results rather than maintaining the ‘status quo plateau.’