Are Digital Marketing Agencies on the Brink of Extinction?
By Derric Haynie
“What do you even do?”
That’s a question I used to get, back when I was running an advertising agency, by prospective clients who needed help with Facebook or Google ads.
As a fellow agency owner, you may already know the answer. Typically, what we’re doing for an ad account client boils down to three things:
And good agencies have a strong process around all of this to ensure the three pieces work cohesively within the budget and goals of the client.
For better or worse, all three core aspects of the “job to be done” by agencies are now completely (not partially, but seriously 100-percent completely) replaceable by technology, and that technology is on pace to get significantly better and more prevalent during the next three-to-five years.
Now, this doesn’t mean that an agency will be completely replaced, but it does mean that your employees will be managing the tools and AI, more than doing that typical day to day grind that we all know and love (or hate). Hey, maybe this will be good for you, as you can decrease costs and keep prices relatively the same or move them down slightly while still improving margins.
Don’t believe me? Let’s dive into each of these three pillars of ad agencies, and I will give you examples of some of the tools putting you and your team out of business.
Segmentation / Targeting
We already know that platforms like Facebook and Google are pretty good at optimizing their own algorithms to find us the most-likely-to-convert customers at the cheapest price, but it doesn’t stop there.
Take a look at this list of key tools for segmentation and targeting already on the market and widely in use today:
All of these tools are, in some way, auto-optimizing your targeting and creating custom segments without needing a human to intervene.
That said, many of these tools still require oversight by a smart human. They do not completely remove the expertise required, but they certainly reduce the time spent and cost of creating all those niche targeted ad sets — or even sitting in a conference room with the client and hypothesizing on which segments to create, etc.
What this means is that a smart internal marketer is more likely to be capable, and have the time, to manage an ad campaign. They have less of a need to lean on the expertise of the agency, since the agency is likely just leaning on the expertise of the AI tool with very little value add.
Have you heard about Miquela? The Instagram influencer who is 100% AI-generated.
No one has actually ever seen her because she doesn’t exist. Yet, she has 3 million Instagram followers and rakes in tons of money in sponsorships and collaborations.
In the next five years, creating content will no longer require a camera. You’ll be able to take whatever products you have, hand them over to an AI engine, and that AI engine will create hyper-realistic photo and video ads with AI actors and AI settings. Want to be under the Eiffel Tower? No problem. On top of the mountains in Aspen? Sure thing. And, of course, that comes with AI-generated copywriting, which can be created at scale, tested against each other, and quickly the AI can come to the right conclusion on which copy/creative/setting/sentiment and hundreds of other factors are contributing to conversion.
In fact, AI can test these aspects of creative and copy cheaper than your content team — and more effectively. They can swap out one word, remove one actor, change a dog to a cat, etc.
Think about it: you don’t need to pay actors, you don’t need to fly people to on-location sets, and you don’t need to hire writers to come up with some witty copy. I’m sure you are thinking, “But, Derric, surely this AI is going to suck, right?”
Here’s the deal, the AI of the last 20 or so years has been pretty bad at replacing humans for things like comedy, drawing out an emotional response from a user, etc. But what these AI-learning tools are doing is analyzing an ad — both visually and from a copywriting perspective — tagging it with dozens or hundreds of attributes, and then tracking that ad's performance. Doing this across billions in ad spend and thousands of companies results in an insanely powerful AI that can, in fact, compete with and even replace content teams and writers.
Here are some of the tools doing this today, and while they mainly still require creative input and censorship/guidance from you or your team, they are certainly streamlining the process of making massive amounts of creative, at scale, without bringing in a design studio or the like:
Krome Image Labs
What this means to me is that your content team is going to be doing less to no work, while other teams like influencer, merchandising (for product photos), and perhaps simply the internal marketing team members who pull review callouts or other elements of social proof, will be the ones adding content into a library, allowing the AI to include it on its whim, and test new content for viability and return-on-ad-spend (ROAS).
Side note: Shout out to the really powerful attribution tools helping to make ROI clearer and thus help our AI make better optimizations and decisions. A shortlist of my favorites:
The last pillar to be replaced completely with AI will be strategy. Clearly there is a lot that goes into creating a successful advertising campaign, and much of it starts before running a single ad. It starts simply with just having a good product and product experience, then moves to pricing, landing pages and product detail pages, offers, discounts, social proof, and any other aspect that you can imagine is “unseen” by the AI engines working to optimize ad campaign performance.
While I can’t tell you all of your business strategy will be fully optimized by an AI in the next five years (it will probably take closer to 20 years for that), I can stand by my original premise that all advertising strategies for any company that has proven to have some sort of mild success with advertising, can and will be completely automated in the next five years.
Many existing tools are already optimizing budgets across platforms, optimizing budgets based on revenue, customers, products — including whether products are in-stock, close to selling out, or close to expiring — or profit margins.
And, yes, there are even tools that will auto-split-test your landing pages/product detail pages, test your pricing and/or discounting strategy, and test your merchandising (these are typically just called “personalization” tools).
The ones that I see doing a great job of optimizing strategy now and into the future are:
Prescient AI: new to the scene but powerful tech, primarily focused on cross-channel attribution and budgeting.
Pencil: primarily focused on extracting strategy (and optimizing for learnings automatically) from the creative testing process.
AdRoll: nearly full automation of the cross-channel advertising process, but currently more on the “helping you do” side, than “done for you.”
Measured: Really great at understanding incremental ad spend by channel and automating the cross-channel budgeting.
What this means to me is that all quantitative aspects of strategy — such as pricing, discounting, and many aspects of split testing on landing pages — are soon to be done exclusively by AI, thus reducing the need for a conversion rate optimization expert, or expensive consulting firms to run customer surveys and tell you more about your customers (I’m sure they will still exist, but they may be less relevant, since we will have easier ways to gather similar information).
Clearly there has been a big shift in advertising and ad-tech over the past 20 years, from the original “helping me see,” to “helping me do,” to “doing it for me.”
Within the next five years, with the proliferation of AI and consolidation (or expansion of feature sets) of all these tools listed here, there will be tools or platforms that single-handedly manage all of your ad spend across all major channels, including creative, targeting, and ad strategy.
And this AI platform will be able to test all of these things, iterate, and improve, without any human intervention.
All the human will need to do is set budget and goals (typically in terms of ROAS or cost-per-acquisition).
What are you going to do to prepare for this? My suggestion: invest heavily in understanding the tools, use and embrace them, and be on the bleeding edge of this transformation. It’s the only way to ensure you don’t go extinct.
Derric Haynie is Chief E-Commerce Technologist for EcommerceTech.io — where e-commerce stores go to research, discover, and buy the right tools to grow their store. He and his team are working hard to map the e-commerce technology landscape and help merchants track and manage their tools, build their technology roadmap, and hold their tools accountable for their performance. He can be reached via email at firstname.lastname@example.org.