Chandal Nolasco Da SilvaWith the practice of retargeting on the tip of every marketer’s tongue we’ve put together a three-part post dedicated to the different angles on this topic. Read Understanding Retargeting Part 2: Dynamic Search here.
Personalized ads that respond and change based on what is being searched can get even more automated and optimized when machine learning is brought into the picture. Machine learning is one of the building blocks of modern AI and certainly a tool marketers are already solving problems with, simply by having a better understanding of the increasing data available to help make smarter marketing decisions.
One of the latest Search Engine Nerds podcasts discussed the topic of machine learning to help marketers leverage all the data available to us now. Our CEO, Marc Poirier, was the guest on the show and discussed how despite the fact that machine learning algorithms have been around since the 80s, that we’re only now labelling them as just that and applying them to marketing. So while marketers may feel disconnected from the AI discussion, they’re already a part of it. Yet just saying that mysterious algorithms are working silently all around you right now does draw an abstract picture. Part of the problem is that each version of AI and machine learning will look and act differently depending on the problem it’s intended to solve.
For this article, thankfully the focus is on what machine learning can do for paid marketers looking to retarget their audience.
Machine Learning and Retargeting Basics
Bid and budget management (BBM) is a big part of the campaign manager’s job. Managing budget spend across the month while getting the most clicks for the lowest CPC is an ongoing challenge for every PPC campaign. Our BBM optimization feature is powered by machine learning and well-known in the industry for taking this task off the campaign manager’s shoulders while delivering unprecedented results. BBM optimizes campaign performance for PPC and retargeting campaigns alike.
In a normal search or display campaign, BBM would:
- Maximize clicks and conversions in the campaign period
- Deliver on-target CPC/CPA based on budget constraints, without overspend
- Allow campaign managers to scale and optimize the number of campaigns they can manage
Machine learning would also help to manage the seasonality of PPC campaigns for example. The way that BBM helps to optimize regular PPC campaigns can be applied to retargeting campaigns with some unique considerations.
The bid and budget machine would learn from slightly different data in a retargeting campaign and therefore could perform slightly differently. One of the best examples is with the bid amount for retargeting campaigns. Because retargeting ads are shown to people who have already interacted with your brand, BBM would learn to increase bids because these people are more likely to convert. However if the “retargetable group” is too small, the campaign risks low traffic volume and BBM may start by bidding high on new campaigns. Machine learning does best with a lot of data, so the more people who are in the retargeting campaign, the better. Similarly, machine learning campaigns do best with some time to learn, so if the retargeting period is too short, BBM may not have enough time to learn and optimize the campaign performance.
The machine is always improving its understanding of each campaign’s sweet spot where it can compete with others in terms of bids while ensuring perfect budget management. For greater clarity, let’s look at a concrete example unique to the automotive sector.
Artificially Intelligent Retargeting Campaigns for Automotive
The idea of dynamic retargeting for automotive is to show people the same vehicle they viewed on your site in an ad that appears and reappears as they browse the internet. If for example a person looked at many vehicles in the same category on a dealership’s site, the dealer could retarget them with ads featuring cars from the same category (like SUVs).
Using dynamic campaigns for retargeting automotive campaigns is easiest. The ads will show the inventory that the dealer has on their lot today and the pool from which ads could be created would be updated the next day with the new inventory of vehicles in stock. Dynamic remarketing is particularly important for automotive marketers, pulling from a new and used vehicle inventory, where in the case of used, there is truly only 1 available. When coupled with machine learning, automotive marketers have a recipe for success with dynamic and somewhat personalized ad copy and the most optimized bids in the industry.
Dave Meindl is the PPC Specialist at Mudd Advertising. He recently discussed machine learning and automotive marketing including the benefits it’s had on his campaigns as well as new types of retargeting that had never been done before. He pointed out that while dynamic display and search remarketing ads have been available through AdWords to eCommerce clients for a long time now, that in the automotive space, vehicles are a prohibited item in the merchant center. He approached Acquisio to work with him to develop smart dynamic remarketing campaigns for automotive on search and display. Dave is now the first marketer in the world who can run dynamic display remarketing campaigns on the GDN, with the bid and budget powered by machine learning for the automotive sector. You can read a full interview with Dave about how he creates his dynamic display remarketing campaigns for automotive here.
The Golden Age of Remarketing
The evolution of remarketing has come so far, from a website visit to dynamic campaigns for cars run by other machines! Successful remarketers will follow Google’s best practices regardless of the type of remarketing campaign they’d like to pursue or the sector they’re representing. Not doing so could lead to consequences that no one will be happy about. Anyone who misuses retargeting and puts a person’s identity at risk could have their ads or even worse their campaigns suspended. Google draws a line when it comes to personally identifiable information (PII) and violators will be prosecuted. It’s not always rainbows and candy drops in the land of retargeting.
And whether we as marketers get it right or not, sometimes things are out of our control. Safari just announced that they will soon block retargeters from following anyone around the net, which.. uhh kinda makes our jobs a little harder. Regardless of whether we get hyper-intelligent, automated, creative and optimized with our ads, if the future of remarketing is uncertain it won’t matter.
Rest your little marketing heart though, dynamic remarketing is currently alive and winning on browsers other than Safari. And exciting times are here with machine learning marketing tech! Google has just announced a host of changes to AdWords powered by machine learning. In the near future we should be able to dynamically target people who came and left our store with the most intelligent bids imaginable, powered by an unthinkable amount of data! We’ve come this far and I’d like to think that with dynamic, personalized, machine-powered search and display campaigns we’re now living in the golden age of remarketing.
That’s a wrap in our three-part retargeting series! We hope you learned and got the resources you needed, but if you’d like to learn more about retargeting or have any questions, please leave them below – we’d love to hear from you! Now go get some of that traffic back 🙂
Image Credits
Feature Image: Unsplash/Carl Heyerdahl
Pam Neely
