How AI and machine learning are driving real results

How AI and machine learning are driving real results

Published: August 13, 2019 by Scott LoSasso
Categories: B2B marketing, Data, Media

With all the hype and mystery around artificial intelligence (AI) and machine learning (ML) we thought it would be helpful to share some practical examples of how we are driving results for our clients using AI and ML technologies.

To preface, it is important to note that the application of AI and ML is far more than the use of tools and technologies. The changes to the algorithm in platforms that dominate the search, programmatic and social media industries are also of critical importance. In her recent research Joanna O'Connell, Vice President, Principal Analyst at Forrester Research observes that AI has infiltrated all aspects of the advertising process, introducing both new opportunities and new risks. And while AI is certainly the buzzword of the day because of its transformative powers, “buzzwords without understanding can be problematic—if not downright dangerous.”

We are in early days of where AI will take us, but it is already clear that the human aspect of applying these technologies to marketing strategies is arguably more important than ever. At a basic level, understanding that Google prioritizes intent vs. keyword or that Facebook rewards the use of certain types of original content in both organic and paid exposures are examples of a constantly evolving dynamic. In fact, this dynamic has a substantial impact on how marketers utilize modern tools and channels. Studying these trends and understanding their connection to media delivery, consumption and buyer preference is a foundational component of modern marketing and audience development—the practice of cultivating engagement with a target market across a digital ecosystem of websites, social media platforms, ad networks, email, etc.

As we head into the third decade of this century, the frenetic pace of change in the marketing industry continues; but the anchors of new media are relatively entrenched and will remain influential as new tools, strategies and possibilities emerge. Below are a few thoughts on some of the things that are shaping how we support our clients in this era of rapid evolution.

Digital advertising campaign optimization

This is the most widely used application within the marketing industry today. The tools of the trade for digital campaign management have AI and machine learning functions built in. The basic functions of these tools are almost universally used today.

Optimization of media spend

Machine learning technology enables digital media platforms/networks to channel your media dollars to the campaigns, segments, placements and keywords that are driving the highest levels of activity. This is a HUGE advantage over traditional media, which is neither measurable nor adaptable to buyer reaction. Effective use requires critical thinking and active oversight—since AI is not foolproof. In search advertising for example, keywords with multiple meanings, industry-specific terms or acronyms can cause AI confusion and lead to wasted spend.

Creative testing

Programmatic display is an obvious example, but nearly all direct channels/networks are ramping up the development of their own automated optimization tools. Facebook’s Dynamic Creative is a good example. As Facebook describes it, “Dynamic creative finds optimized ad creative combinations by taking multiple ad components (such as images, videos, titles, descriptions and CTAs) and automatically generating combinations of these assets across audiences. The delivery system then optimizes for creative components that deliver efficient results for each impression served.”

We use this for several of our clients—including a client with a complex program that targets multiple segments within the foodservice industry with varied messaging. In this case, minimal creative asset development paired with the Dynamic Creative capability allowed for roughly 4,000 distinct ads (combinations of headlines, images and text). The results allow us to evaluate both quantitatively and qualitatively to drive short-term creative adjustments and long-term planning strategy.


Ad platforms allow you to upload data such as email lists or first-party IP data, so that you can use AI to create look-alike audiences. These are segments of buyers that can be reached over the platform that closely align with the interests, demographics and behaviors of your current customers and prospects. This is an extremely effective way of sharpening your targeting and messaging to get higher engagement and ROI on your media spend. However, this is not a simple and fully automated process. True success with this tactic requires critical thinking and expertise: If you don’t effectively segment the data that you upload, or think through the desired outcome, this won’t work nearly as well.

Content curation and evaluation

The world is overflowing with content—and it isn’t all good. However, building a connection to your customers and prospects by serving them content that aligns with their needs is an extremely effective way to build trust and awareness for many types of brands.

Content curation

Content comes from a wide range of sources—finding and evaluating it can be a massive undertaking. This is the genesis of content curation tools. These are tools that search volumes of content and filter recommendations based on parameters that you control. The more you work with the tool, the more effectively it will screen content for you—and the more you circulate and monitor audience engagement, the more you can tailor the tool to what your market is hungry for. Here again, it is the combination of a human insight and expertise with AI and machine learning technologies that create new advantages and possibilities.

Content ideation and research

Machine learning and AI come in handy for content ideation and research as well. We use tools that automatically aggregate, analyze and present popular existing content and new topic ideas that align with specific keywords, industries and audiences.

Email optimization

Other than paid digital optimization, email is probably the most widely adopted use of ML and AI. The low-hanging fruit is simple subject line testing and message testing. These are both common and effective. In-fact even as our inboxes continue to overflow, smarter use of email is one of the most effective tools in the toolbox for B2B and ecommerce brands. Marketing automation and lead nurturing are more advanced but also widely applied. The combination of email and media placement is an additional layer. In fact, your email database is one of your most valuable first party data assets. We use our client’s email data to build media segments for them across display and social networks; in turn, gated content generates more contacts for our nurture streams and ongoing email communications.

AI and ML can also help you avoid spamming or fatiguing your list, and there are continually evolving ways to drive value within your email programs.

The fundamentals of marketing are the same as before—the more insight you have into the wants, needs and processes that drive buyer behavior, the more effectively you can market your brand. The universal challenge is doing it all with the same budget. There are more bases to cover today than there used to be, and this requires a different approach to planning. The good news is your competitors are in the same boat. Marketers that are leveraging these tools are making decisions driven by data, and they are quietly building a sustainable advantage. When you can reallocate budget with data-driven confidence you can be more agile—and there is no question that agility in today’s marketplace is a strategic advantage.

We’re interested to learn how others are applying these technologies, and welcome comments and discussion from clients, peers and prospects. How can we can put tools to work to make us all better, more efficient and more productive? Leave a comment or reach out directly, we would love to hear from you.