How Automation is Transforming Growth Marketing

The rise of more automated advertising products poses both conveniences and challenges for marketers. Whether you fully embrace this evolution or are hesitant to give up control to ad platforms like Meta and Google, automation is here to stay – and it’s already changing how you work. Read on to learn what to expect.  

‍

Fewer Optimization Levers 

‍

In the early days of paid digital advertising, marketers controlled almost every element – from copy in ads, to granular audience selection, and customized bidding strategies. 

As quickly as these levers came, they are on their way out, with major players like Google severely limiting their options in favor of AI and algorithmic buying products. Google’s new campaign type “Performance Max” or “PMax”, for example, uses machine learning to automatically generate paid search ads from provided product information. It determines where the ads are served, and adjusts the ad creative and content to fit the placement. Similarly, Meta’s Advantage+ shopping campaigns automatically designs dozens of creative combinations and decides which audiences they’ll be served to.

While optimization levers are shrinking, that’s not necessarily a bad thing. Leaving your ad buying to a tech giant like Google or Meta lets you leverage their rich data resources to make automated ad decisions, while removing human bias from the equation. 

That’s because these platforms have their own massive first-party user data they use to determine ideal audiences for your ads. Machines can create optimizations at a frequency and with a consistency that humans can’t replicate. Using this technology to your advantage ensures data-driven decisions are prioritized, and the process can move along effectively without manual intervention. 

‍

Evolving Growth Marketer Responsibilities  

‍

The time freed up by machine learning is leading to different opportunities for growth marketers, who are now shifting their attention towards strategic planning, creative, and experimentation. 

Daniel Pearson, CEO at leading digital agency Bamboo, says 

“We've definitely seen the role of media buyer evolve from one that was extremely ops heavy to one that can be more strategic and analysis heavy instead. It's way higher impact to spend time analyzing how the ad accounts are driving (or not driving) business results than it is spending time on campaign build outs and flighting. At the end of the day, our mandate needs to be growing the businesses we work on behalf of rather than managing their ads.”

‍

• Strategic Planning: Marketing is quickly becoming a much closer partnership with finance. “Media buyers must understand the financials of a business in order to understand and determine the media mix and the strategy behind each channel” says Mike Lukashewich, founder of growth marketing agency, Online Impact. As a marketer, you need to understand the full accounting of your business, not just top-line revenue.

• Creative: Even the automated ad products require you to provide your own product information and/or design assets. That means, while the platforms will ultimately generate the ad layouts and placements, you still have full control over the concepts and the raw materials it uses. Over half of the marketers we spoke to said that they are spending more time on creative with the rise of more automated ad products. 

• Experimentation: Running experiments on your key marketing channels & tactics will enable you to validate and accurately quantify marketing’s contribution to your business. This is especially important as we see increasingly black box solutions from Meta & Google. Controlled experiments or incrementality work on any ad platform or product. By simply comparing an audience that was exposed to your marketing with one that wasn’t, you can continue to test the performance of your marketing approach. 

‍

Scaling Your Experimentation Practice  

‍

On the surface, automation makes it harder to understand what is driving ad performance. After all, when PMax and Advantage+ are determining strategies for you, you have a lot less visibility into what is actually working. Additionally, the conversion results these platforms claim likely have a certain amount of bias in them, since they’re essentially grading their own homework. 

Incrementality testing allows you to run controlled experiments in order to detect the true impact of any given media channel or tactic. Incrementality experiments use the scientific method to test a single marketing variable and measure its lift. In an incrementality experiment, there is a control group and an exposed group. The exposed group receives the marketing variable you are testing and the control group does not, but all other factors remain constant. Any lift in conversions in the exposed group indicates the true impact of that variable. If you aren’t familiar with incrementality, we’ve created an ultimate guide to get you started. Check it out to learn why this is the most accurate type of marketing measurement in light of automation and privacy changes. 

Incrementality experiments provide a new lever for ad optimization in light of the reduction in traditional levers. For example, you could run an experiment to determine if PMax is driving as much or more incremental value as non-brand search or to determine the optimal spend level to allocate to each tactic to maximize efficiency. These automated campaign types are more black box than ever before so it’s important to test these tactics to get a true understanding of their business impact.

These experiments can be run with a testing platform like Haus to ensure the results are accurate and actionable.

‍

Start Incrementality Testing with Haus  

‍

Incrementality experiments may seem simple in practice, but carrying them out accurately and precisely is complex and time-consuming. To ensure accuracy and precision, you need economists, data scientists, engineers, and a powerful experimentation model. That’s where Haus comes in. 

Our state-of-the-art experimentation platform adds the convenience of automation without the loss in control. We create the model, you design the experiment in the Haus app, and we deliver you actionable results in weeks, not months. Want to learn more? Reach out today to request a demo.