Using Machine Learning for Enhanced Marketing Results.

#AdobeChat — Machine Learning

When we look at today's digital marketing landscape, it's easy to get overwhelmed with a case of technology overload. Marketers are being bombarded with new tools and strategies every day. Let’s dive into one of them — the hot topic of the week is machine learning.

With machine learning, marketers are able to change their entire targeting and creative processes — from how they attack simple tasks to how they handle the much more difficult task of brand storytelling. The role of machine learning will continue to grow over time, and we think it’s safe to say that machine learning is just taking off in a big way. Let's discuss.

Q1: Machine learning: What is it?

If you’re not already versed in the concept of machine learning, it’s time to give yourself a crash course. At a basic level, machine learning handles the hard data, allowing digital marketers to truly understand consumers — their interests, personality, habits, online behavior, etc.

In the age of the consumer, this data is incredibly important for a brand's success. Marketers can’t operate the same way anymore. Machine learning moves the needle from traditional, generalized demographics to highly specific, individualized information. So... machine learning: What is it?

Machine learning is: 

  • Automating automation. Getting computers to program themselves, similar to data mining. Let data do the work.
  • Using algorithms to learn from data so computers can make decisions without humans.
  • The ability for the knowledge of software to be expanded without a human interfering.
  • The ability for computers to learn from data without needing additional programming.
  • The ability for a machine to learn input tasks and use them in the future for automation.

Q2: How can machine learning be most beneficial to marketers?

The constantly evolving role of marketing technology has made it both exciting and confusing for marketers to operate internally, as well as how to reach their target audience. The exciting part is that this technology helps marketers understand their consumers in ways that were never before possible.

The confusing part is trying to understand how the technology works and how to implement it within brand strategy. Let’s focus on the positive. We see many ways that machine learning can benefit marketers. How do you think machine learning can be most beneficial?

  • Machine learning gives marketers a clear vision and understanding of their audience.
  • Machine learning can help connect the dots between what consumers say they do and what they actually do.
  • Machine learning collects and analyzes consumer data for hyper targeted and personalized marketing. 
  • Machine learning takes things from a shotgun approach to a sniper — personalizing marketing and predicting behaviors.
  • Machine learning identifies patterns, allowing marketers to strategically take actions.
  • Machine learning turns data into something actionable.

Q3: What about consumers? How can machine learning be valuable to them?

For years, marketers have been focused on one main priority: how to establish, sustain and expand relationships with their consumers. Here’s the reality: creating an amazing customer experience is what drives sales. The consumer is in the driver's seat. They are the ones researching, discovering, browsing and purchasing products.

If marketers want to reach those individuals, they need to understand them. Machine learning can be the backbone to delivering the most relevant and engaging experiences to consumers. In your opinion, how is machine learning valuable to consumers?

  • Machine learning provides consumers with relevant recommendations, saving time when browsing products they may be interested in.
  • Machine learning can help customize the experience, tailor it where it suggests things the consumer didn't know they wanted.
  • Machine learning collects data on a consumer's past behavior, delivering suggestions pertinent to them as individuals.
  • Machine learning gathers insights across new mediums — opening up information that wasn't previously available.

One can only dream, Adel!

Imagine wanting a cup of coffee and suddenly finding it before you, freshly prepared to your exact standards.

Q4: With machine learning present and available, should humans be afraid for their jobs? Why or why not?

We see it every day — drones, robots, virtual assistants, smart machines and much more. You can’t deny it. Technology is becoming more and more intelligent — and it has people rightfully worried about job safety. But can machines and technology really replace human judgement?

Machines can’t empathize. We can. Machines aren’t creative. We are. And when looking at how to market to consumers, these innate human characteristics are critical. With machine learning present and available, do you think humans should be afraid for their jobs? Why or why not?

The vast majority were in agreement that machines will never replace the human need within marketing. 

  • Just like machines evolve, humans will need to evolve in tandem.
  • It's an "adapt or die" world. Machine learning will take over mechanical tasks. To survive, marketers need to gain new skills to work as a team with technology.
  • It's true that many jobs will be automated away with machine learning. But if you can learn on the job and get new skills, no need to fear.
  • Humans that can't adapt should be afraid for their jobs.

Q5: Is machine learning more prominent than we even realize? What delightful marketing experiences have you had where you think machine learning was involved?

Spend time with any marketing team and you’ll most likely hear "machine learning" come up. Industry visionaries are already talking about how this technology is going to be an influential part of a strong digital marketing strategy. All of this talk makes it sound like machine learning is coming up, but the reality is that machine learning is already here.

Machine learning is ready to work for marketers right now. The best way to understand the value of machine learning is to look at how brands are using it today. So, what do you think? Is machine learning more prominent than we even realize? What delightful marketing experiences have you had where you think machine learning was involved?

  • Spotify: provides users with tailored music recommendations.
  • Amazon and Dunkin' Donuts: personalize content for each individual.
  • Facebook and Instagram: create a unique experience with relevant content based on previous user behavior.

Great point here, Eric: 

I honestly can't think of a particular one. And maybe that's the point. It should be seamless and feel like magic.

Q6: How should marketers responsibly leverage machine learning (without crossing the creepy line)?

We already know that machine learning is all about data, specifically consumer data. We also know that collecting and implementing consumer data can be a sensitive subject. It makes sense. Accessing specific, individualized data can feel like a breach of personal privacy. It’s a tricky topic and it’s definitely not black and white.

There’s no rule book outlining the right way to navigate the waters. While marketers want to put this invaluable data to work, they need to ensure they’re not crossing the line — the thin line between being delightful or creepy. How do you think marketers should responsibly leverage machine learning (without crossing the creepy line)?

Interestingly enough, a large portion of the audience had a positive reaction to brands using their data. 

Being relevant is never creepy. Just do what's right for the customer.

  • If you use machine learning to provide value, you're not going to cross the creepy line.

I think marketers need to be mindful of re-marketing and personalization. A lot of people still find too much of it invasive.

  • Marketers can use the technology to identify groups to market to.

Su makes a smart recommendation: 

Anticipating consumer needs is less creepy when coupled with an opt-in subscription or loyalty program. The user has to see value.

Q7: How can marketers tell if their machine learning is working? Are there metrics that matter?

As with pretty much every digital marketing strategy today, it’s extremely difficult to determine whether or not a program is successful. Now, marketers are dealing with vague metrics, some of which we can’t even define. The term “engagement” gets thrown around all the time. What does engagement mean? How do you measure that?

We’re also talking about slightly more tangible metrics like time spent on-site, conversions, click-through rates, likes, comments, etc. All of these metrics bombard marketers, especially when it comes to implementing a technology such as machine learning. How do you think marketers can tell if their machine learning is working? Are there metrics that matter?


Simple and straightforward.

If your machine is learning, your business will be changing.

  • Retention and upsell.

Engagement, of course, hit the list: 

The new metric that matters is the amount of engagement. Not just clicks, but conversions.

  • Increasing ROI, NPI and profit.
  • Look beyond click-through rates.

Q8: What is the realistic potential for machine learning going forward? How will it differ in the future?

We all know the story. To stay ahead of the competition, you need to be one step ahead of them. With technology evolving and growing stronger every second, what works today might not work tomorrow. That’s why marketers are always looking to the future.

Marketers are already working with machine learning — trying to understand it, utilize it and implement it in a way that strengthens their strategies. Like we said above, technology is advancing. This means machine learning is also progressing. What do you think the realistic potential is for machine learning going forward? How do you think it will differ in the future?

  • Machine learning will remove guesswork when targeting consumers.
  • Machine learning will predict future consumer behavior and foresee trends.
  • Machine learning will infiltrate all of digital life — making new apps more sophisticated.

Every successful new application built today will be an intelligent application.

And finally: 

The potential of machine learning and AI is beyond what we can currently comprehend. Most wouldn't believe it.

That's a wrap!