ActionNext℠ at the Midwest Digital Marketing Conference

An interview with Ward Thomas, Director of Customer Data Science at Havas Helia

Ward Thomas, Director of Customer Data Science at Havas Helia, was chosen to speak about ActionNext℠ at the 2019 Midwest Digital Marketing Conference. We sat down with him to chat about his presentation, experience, and takeaways.

First question, why the Midwest Digital Marketing Conference?

MDMC is a digital conference with a focus on data, so it’s a perfect fit for us. We joined other presenters from places like Facebook, Under Armour, Google, and Nielsen to share the exciting things we’re doing and data tech we’re working with.

We’re passionate about the work we do, especially when it’s something we put a lot of time and effort into getting right. So sharing with people in similar fields with similar challenges was a great opportunity.

Can you tell me a little about your presentation topic?

Sure. I presented a retail client case study on ActionNext℠, our proprietary automated predictive modeling process. We used a blind client reference to protect confidentiality, but the approach and outcomes were based on actual results. I gave a description of their business challenge — they wanted to test whether sending personalized offers would increase conversion and boost sales — and the approach we took with ActionNext℠ to see if we could beat the control, which consisted of a one-size-fits-all promotional piece featuring many of their popular products.

The results were really encouraging — the randomly selected customers who got the personalized offers chosen by ActionNext℠ were 29% more likely to redeem them. They also had 32% higher incremental net sales than the customers who got the one-size-fits-all promotion. The client was thrilled and started using ActionNext℠ to personalize offers for more campaigns and marketing channels.

So what makes ActionNext℠ important?

The real break-through is being able to build predictive models at scale. When our objective is to personalize content — whether in an email, display ad, or another marketing channel — if we have a variety of offers we can make, the question becomes which customers should receive which offers? There could be 10, 100, 1,000 or more to choose from. So how do you pick?

Traditional predictive models are built methodically one behavior at a time to answer questions like how likely someone is to respond, churn, etc. What we needed in this instance was to build roughly 100 models so we could figure out which offers were most likely to resonate with which customers. And with ActionNext℠ we can build accurate predictive models in minutes instead of what would otherwise take days or weeks.

What reactions did you see from attendees?

I got a couple of really good questions from the audience, and several attendees approached me afterward.

I remember one asked if our models were applicable to existing customer databases or to pure prospects? In this particular case, we used an existing customer database with past transaction information. This definitely makes the predictions more accurate since the models are based in part on prior purchases. But ActionNext℠ isn’t limited to existing customers — we can work with prospects too. As long as you have demographic information about prospects who made an initial purchase, you can apply ActionNext℠.

Someone else asked about whether the retail case study I presented could apply to other industries. ActionNext℠ is almost universal. You could use it for entertainment, travel, automotive, financial services, etc. In one example, we used ActionNext℠ for a hotel chain to send past guests personalized offers, like spa packages, room upgrades, and more. In another case, we helped a console gaming client choose the best games to advertise to each gamer. The possibilities are really endless.

Did you attend any events outside of your presentation?

Yes. I attended a speakers lunch and had the opportunity to dine with several university professors. It was interesting to hear about how they were striving to keep their content relevant in the face of emerging digital marketing platforms, programming languages, and analytic techniques. A lot of the same challenges and goals we see in our own field.

What are your key takeaways from the conference?

There were over 2,000 attendees at the conference with different backgrounds including academia, end-user organizations, vendors, and consultants. There’s a lot of great work being done across all of these groups in areas like data, digital, creative, and e-commerce. MDMC brings these different perspectives together to collaborate and share best practices. I’m sure everyone was able to glean insights from their peers that they could put into practice.

To learn more about ActionNext℠ and how it can help you improve your customer experience, email hello@havashelia.com.