unsplash robot hand

Your Top Questions About AI & Machine Learning Answered

Both artificial intelligence and machine learning are trending for 2019 with absolutely no signs of slowing down anytime soon. While this new era of technology has a tremendous number of benefits, it can sometimes be difficult to tell fact from fiction.

Recently, our Co-founder Marc Poirier, along with PPC pros Brad Geddes of AdAlysis and Jeff Allen of Hanapin Marketing hosted a webinar to address these questions. During the webinar they covered many questions on AI and machine learning, and we’re sharing the hottest questions in the round up below.

Every day, we’re seeing automation on the search engines, platforms and third-party tools. Should we be changing our marketing strategy to address this?

Brad shared how over the past five years, the two biggest things in paid search have been machine learning — which is all about automation — and then audience targeting where we want to get specific about each user group.

“These two things work at cross-purposes, and if we rely too much on machine learning, we risk losing the conversation with our users. The goal needs to be using machine learning for automation but to stick with our strategies around brand voice and messaging.”

Marc went on to add that having a solid strategy remains critical, and the tools are there to help execute on the strategy.

A lot of people seem to fight against or be fearful of AI. Why do you think that is?

For years people have fought against automation, and according to Brad, they now realize that math is math and if a tool can handle it, that’s helpful. “Where we do see people fighting AI is when it comes to their brands as they have a lack of trust in the algorithms. They want to know exactly what’s happening.”

He went on to explain, “We’ve seen people pushing back against Google’s automated bidding as it’s messing with their campaigns, and that’s seen as a problem with the automation. Really, it’s more of a challenge with how things are tied together on the platform.”

From Marc’s perspective, there’s a clear reason we’re resisting AI, “A lot of the challenge is explaining the principles of what we’re trying to accomplish with AI or machine learning but we can’t always explain why a specific decision was made.” The machine analyzes the data, makes decisions, and continues to learn as it goes.

What’s the difference between automation, machine learning and deep learning?

Marc walked everyone through the differences, “Automation has been around since the 1950s, and it’s simply a way to get something done automatically so you can create business rules based on logic. AI is tied to automation procedures, and you’re trying to get a computer to think like a human using those business rules. Machine learning emerged in the 80s, as computers were able to deal with larger datasets. Over time, researchers noted that the machine would learn on its own and improve.”

“Over the last five years, there’s been a significant acceleration in the different types of machine learning, including deep learning, which is a way to analyze a very large dataset rapidly.”

PPC professionals are having to get more and more comfortable with automation. What’s your comfort level with what’s out there right now in terms of AI and machine learning and its accuracy?

In response to this question, Marc shared how professionals need to ask themselves what problem they’re trying to solve and if the tools are doing better or worse than a human would. From there, we need to assess if, at scale, you can solve the problem in a cost-effective way. It really comes down to figuring out if you’re seeing results.

Brad went on to share an example of using automated bidding, “Did you get more conversions for less money? A lot of the decision making comes down to your risk tolerance with different parts of your campaign. Ask yourself, is the bid completely wrong or does the problem lie with your creative?”

He pointed out that as PPC professionals, we need to look at if the tools, on the aggregate, are working as well and how machine learning and AI should act as a compass for everything else you’re doing.

“Everyone needs to determine the level of risk they’re willing to take on, based on the potential rewards of using AI,” added Marc.

How much about AI and machine learning should marketers really know?

Brad shared how a lot of marketers feel like they need to get a deep understanding of what’s going on when really it’s about auditing the outputs.

“We should use the tools to get recommendations — and then accept or decline them — and that will require a fundamental understanding of how things work. A PPC marketer doesn’t need to be a developer who’s able to write script. The focus should be on the creative, writing and the overall strategy. It’s about getting it right at scale with machine learning and AI.”

The story is a little bit different from Marc’s point of view, “For agencies, I do think there needs to be someone on the team who’s well-versed in data science, so you can explain to clients how things work.”

Does this mean the agency model changes when it comes to PPC?

“Typically, you hire an agency because you don’t have all the skills you need in-house, and you don’t want that headcount,” indicated Brad. “The agency is really about the strategy, but it’s likely that job functions within the agency will change. Account management and reporting about what the machine is doing will be key, as it’s still in human hands to manage what’s happening.”

Expanding on Brad’s thoughts, Marc argued that the agency offering will need to evolve. “There’s likely a feature change where the offer evolves so there’s expertise in data science included. That way, the agency can tell the client what the data means and what they should be doing with it.”

As an agency owner, Jeff had valuable insights on how some things have become simpler and others more complex. “We’re dealing with multiple platforms that we need to have expertise in, and we need to deliver strategy and results for clients. It used to be about making things like Google Ads simpler for clients; now it’s about taking our clients’ complex business models and making them work within the ecosystems that are out there.”

What type of math do I need to have training in to better understand PPC and machine learning?

The good news is that Marc doesn’t think you need to be a mathematician to excel as a PPC professional. He did outline what you need to know to succeed, “What you need to know or get trained on is statistics 101 so you understand what tests to apply in what situations. You need to have a working knowledge of the variables at play and what degree of confidence you’re seeking. There are tons of courses you can take — including free online ones. You may want to check out Linda.com or Khan Academy.”

When do you think would be the perfect time to begin a test with a new machine learning tool?

Unsurprisingly, Brad encouraged the audience to start testing as soon as they want to get better because there isn’t really a bad time to start.

However, he did review some tips for getting started, “I wouldn’t try it on a brand new account as there’s no data, but if you’ve got a bit of data, and you’re happy with your current volume, then you have what you need to get started.”

He pointed out that what you want to look for is consistent data with no outliers. “If you’re a flower company, you likely don’t want to run a test in the lead up to Valentine’s Day as your results will likely be skewed. You want to use repeatable, consistent data.”

What do you think the impact of machine learning and AI has been on user experience?

There has been a definite impact on user experience — but it’s not AI that’s causing the issues.

Brad got to the heart of the problem quickly: “It’s marketers not setting up campaigns correctly. For example, excessive retargeting ads that don’t have a frequency cap or negative audience. That’s definitely a marketing problem, not an AI problem. It’s not the machine, it’s what people are doing with it.”


Whether you choose to jump head first into using AI and machine learning or take a more measured wait and see approach, the fact is that these technologies are here to stay. Understanding what they do, how they can help your business and separating the facts from the fiction will be the first step towards adopting these technologies over the long term.

If you missed the webinar and want to hear the discussion in its entirety, you can check it out right here.


Image Credits

Feature Image: Unsplash / Franck V.

Chandal Nolasco Da Silva

Chandal Nolasco Da Silva

With nearly a decade of digital marketing experience, Chandal has created content strategies for both the biggest and sometimes the most unexpected markets, while developing strategic relationships with editors and publishers. Chandal contributes to some of the highest authority industry publications, has been featured in industry events and is thrilled to be Acquisio’s Content Director.

The First Machine Learning Marketing Platform
Built to Scale Search for Local Resellers & Agencies

Automate, optimize and track more campaigns, more profitably.