Staff Reporters
Jul 18, 2016

Machine learning: Threat or benefit?

As marketing transitions from intuition-based to data-oriented, will machine learning handle much of marketing down the road?

L-R: Florian Pihs, Ohad Hecht
L-R: Florian Pihs, Ohad Hecht


  • Florian Pihs: Senior planning director, SapientNitro
  • Ohad Hecht: Chief operating officer, Emarsys

Which area of marketing is investing the most in machine learning?

  • Pihs: Pay-for-performance ads, brand ads, DSPs, DMPs, personalisation algorithms on content management systems and ecommerce platforms, A/B testing and multivariate testing for landing experiences and banners.
  • Hecht: Brands with a focus on ecommerce or e-tailing. Those who started online early have the advantage over traditional types.

Who are the biggest players in machine learning in marketing?

  • Pihs: Search engines, social and ecommerce. Ad tech and digital marketing platforms (Abobe, SiteCore, IBM) claim to use it. Big ad networks also invest, largely on the media side.
  • Hecht: All the internet and ecommerce firms. With instant messaging (IM) and mobile-first culture, personal communication and ecommerce are converging — a big area for machine learning.

How do marketers feel about machine learning and why?

  • Pihs: Some media are scared that parts of their inventory are devaluing and feel safer selling in bulk. This creates uncertainty for the industry.
  • Hecht: It depends on whom you ask in which market. If I were a website manager four years ago, I would’ve been relying on my preference in choosing content to post or product to sell. Now marketers are more aware that things are data-driven.

What are the obstacles to machine learning in marketing?

  • Pihs: Data integration/aggregation. Big Chinese platforms’ siloed ecosystems prevent integrated third-party-data aggregation.
  • Hecht: Having enough data to test and run control groups to optimise their campaign.

What can marketers do to make a start on it?

  • Pihs: As soon as you start using search ads, you’ve made a start. You can then utilise machine learning on DSPs and website personalisation to harvest the benefits of higher performance.
  • Hecht: Thinking about data collection and how to connect and unify the data.

How is it changing marketers’ role?

  • Pihs: The role hasn’t changed: the mindset has. Marketers need to think beyond campaigns and into always-on communications. 
  • Hecht: With so many data points, human mind simply can’t keep up. Machine learning won’t replace the role of marketers but will allow them to scale decisions by letting machines do the work.  

Related Articles

Just Published

2 hours ago

Campaign Crash Course: Leadership lessons for ...

How do leaders create effective teams? Dentsu Aegis Network Greater North chief Cheuk Chiang explains how to foster better talent by applying leadership principles, tips and learnings from past mistakes. Watch his video and test your knowledge with a quiz.

2 hours ago

The most beautiful pizzas you will ever see

INSPIRATION STATION: How a pizzeria in Vietnam teamed up with boutique agency Ki Saigon to create beautifully inspiring (and edible) tributes to world peace.

2 hours ago

Lenovo eschews high-voltage celeb endorsements for ...

Consumer electronics brand embraces functionality over bells and whistles as it seeks to stand out in competitive market in a campaign by SuperHeroes.

4 hours ago

How remote pitching is showing up the agency showmen

With the pandemic stalling the traditional pitch process and the rise of the remote or virtual pitch, what is known is showmanship is no longer winning the day.