What interesting client project are you working on currently?
The economic uncertainty we’re in means a lot of brands are asking us questions about what the future impact of a range of factors might be on their business.
One retailer I’m working with is particularly sensitive to changes in the housing market. We’re modelling how different interest rate paths over the next 12 months could affect their sales trajectory – from rates holding steady to further hikes. To do this we use a combination of business driver modelling (BDM), our unique approach to marketing mix modeling (MMM), and an approach we call war-gaming.
BDM factors in macroeconomic variables as standard, so we have a baseline of how things like interest rates have impacted the company’s sales in the past. From there, we use war-gaming to determine what sales will look like under each different interest rate scenario and how their marketing investments could shift the outcome.
I trained as an economist, so I enjoy combining economic theory with the marketing science I’ve learned at Gain Theory. What’s great is that this type of work goes beyond marketing teams and helps them have more strategic conversations with CFOs.
Agentic AI is another topic that is creating a lot of debate at the moment. What’s your take on its impact on marketing analytics?
We’re exploring how agentic – or more likely semi-agentic – workflows could improve the work we do: speeding up parts of the analysis and freeing up time to focus on where we add the most value. This has pushed me to map out every step we go through in MMM, many of which have become second nature over almost a decade of doing this work.
Take exploratory data analysis (EDA), which we do before we start building any analytical model for our clients, as an example. In theory, you could lay out every step for an agent to automate. But I don’t think that’s the right approach.
EDA is where we develop the hypotheses that we test in the modeling and is often a deeply creative process because raw data only tells you so much. It’s the conversations with clients – where you tease out context like a recent change in distribution strategy, or a competitor price war that won’t show up in any dataset – combined with the experience of our analysts, that shapes the most useful hypotheses. If you automate the whole thing, you’d get a technically competent analysis that misses the point.
Where AI already adds genuine value is in the more mechanical parts of the workflow – things like data validation, testing media response parameters or model robustness checks.
Are there any expert humans outside of Gain Theory you follow?
I like to know what people who challenge conventional thinking across marketing, business, and economics have to say. Scott Galloway is always thought-provoking. I like that he makes bold predictions and admits when he’s wrong because that’s rare. But I do disagree with a lot of his views about the end of the era of brand building as I see evidence of brand effects consistently in the MMMs we build.
The David McWilliams podcast is one I highly recommend. He’s an Irish economist who wrote a great book called The History of Money. What I like about him is that he covers topics you wouldn’t necessarily think of as economics – singledom, cartography, religion – and makes the connections back in ways that are genuinely surprising.
What do you like to do when you’re not thinking about marketing, analytics, and economics?
I really enjoy outdoor swimming. I go with my brother and cousins – it’s a fun, social activity and something we’ve done together since we were kids. We swim in a lake and we’re looking to build up to a 5K race this year. It’s great to get away from a computer screen and there’s something about how shockingly cold it is that makes me feel really good!
Contact Michael to discuss any of the subjects raised in this Q&A.
