Reading Time: 4 minutes

Over the past several years, Google and Meta, the dominant online ad platforms, have invested heavily in using AI to make advertising more effective—and when embraced, we’ve seen these investments help advertisers succeed. However, driving strong performance in this new AI-driven world of advertising requires a different approach than what has historically been successful, and we’ve observed many brands struggling to adapt. This article outlines the changes that have occurred (without many teams noticing), and the new way you need to think about managing and optimizing your paid media.

Mindset Shift

AI models require a large amount of training data to be effective (think of ChatGPT). Training data for ad platforms is primarily user behavioral data—which ads are clicked, which websites are visited, for how long, was there a purchase, etc. The models are attempting to predict interest and intent, and match cohorts to the desired actions others have taken. The ultimate goal of a performance-based campaign is to drive conversions (purchases, leads, etc.), so those behaviors are weighted much more heavily in the AI programming.

Unfortunately, using training data to predict individual buyer behavior is not how advertising has traditionally worked; a mindset shift is required. Imagine you’re selling high-end mattresses—historically, you’d research who the likely buyer demographics are, write an appealing ad, target that audience with bids based on keywords for Google and audience attributes on Meta, then monitor number of impressions, CPM, etc. to determine return on ad spend.

Control is Being Taken Away

Today, ad platforms’ AI has become so powerful that they’re removing the majority of targeting controls. Instead, AI identifies people who need a new mattress based on their behavior. As an advertiser, it’s still on us to create a persuasive ad. However, even that is powered by AI. Google’s text ads are driven by Responsive Search Ads (RSAs), which mix and match 8 headlines and 4 descriptions to find the optimal combination. RSAs and the like provide newfound creative power to better understand what the customer wants—by trying a variety of approaches, we can quickly see which is most compelling.

With a Lack of Control, Your Customer Experience is Paramount

I can’t emphasize this enough. With changes wrought by AI, site experience is much more important than it was previously. The ad platforms know that advertisers getting a good return will spend more. It’s challenging to secure a good return when your site experience makes it difficult for prospects to research the right option for them and make a purchase. Therefore, ensuring your site experience is optimal is critical to success. The only way to achieve the best site experience is through experimenting with what works. Today’s most successful companies (Amazon, Netflix, eBay, etc.) have known this for eons and never redesign their site. Instead, these leaders run hundreds of controlled experiments focused on better understanding what their customers want. In other words, their websites are incrementally redesigned over time through winning experiments backed by statistically significant results.

Effectiveness Can Be Misleading

Given the power of AI to find more similar people when given proper signal, data infrastructure is important to an effective bidding strategy. Many advertisers are still using a cost per conversion (CPC) strategy, but value-based bidding is much more effective at taking advantage of platform AI. Not all customers are equally valuable: advertising to a repeat customer already familiar with your brand may prove less valuable than acquiring a new customer likely to have a higher lifetime value than your existing base. As a result, setting up appropriate attribution and sending results to the ad platform based on likely profit potential will push the AI to find potential customers in proportion to the signal you send. This is how smart advertisers drive ever-elusive incrementally. Below is a high-level illustration of how to think about this.

Optimize for Value-Based Bidding

You may have dabbled with value bidding or are running campaigns yielding decent results, but not driving the return you’d hoped.

As much as the ad platforms have removed controls in favor of AI-driven bidding, you still have many levers to strategically influence. For example, if you’re running an e-commerce business, clearly define your objectives, whether it’s maximizing revenue, increasing profitability, improving return on ad spend (ROAS), or other specific targets. This will guide what value signal you send back to the ad platforms so they can optimize against those KPIs. (Again, this is why data infrastructure is key.)

If you’re a B2B company optimizing for leads, integrating your CRM, capturing the lead source (Google, Meta, etc.), and assigning values at each stage of the deal (i.e. MQL, SQL, Proposal, etc.) will best train the AI on who your most valuable prospects are and when they close. Despite my earlier comment about needing large amounts of training data, Google tweaked its algorithm for lead gen so that surprisingly little data can make a big difference… even if you have only dozens of leads per month and longer sales cycles.

Experiment matching customer intent to the landing page experience. This is often achieved with symmetrical messaging (ad-to-site). Other ways to personalize the experience include leveraging what stage the visitor is in the funnel (awareness->consideration->conversion->retention), such as guiding them to a desired next step based on that stage. Examine which on-site experiments are driving better results from your paid campaigns, and run iterative experiments from winners to seek further gains. The ancillary but significant benefit is you’ll also learn more about how customers in each stage of the funnel respond to different messaging, creative, promotions, which can then be experimented on other marketing efforts such as email or SMS. You’ll better understand your customers and be better equipped for ongoing marketing and driving maximum lifetime value.

This has been a winning strategy for our clients at Cro Metics. As marketers increasingly lament about how costly and competitive advertising has become, it’s behooves you to work with the ad platform innovations while maximizing and optimizing control over the things you can—messaging, creative, and the customer experience.

We at Cro Metrics want to be your strategic partner in maximizing results from the ever-evolving digital advertising space. Let’s talk about how we can help your company grow.