This is a problem that has been around for a long time. At the start of the 20th Century, the US merchant John Wanamaker famously said ‘half the money I spend on advertising is wasted; the trouble is I don’t know which half.’
Today, one hundred years on, we are still asking the ourselves the question “How can I tell which parts of my marketing are working, and where should I focus my money?” In normal times, this is a difficult question, and these are not normal times. Attribution studies and MMM (Market Mix Modelling) have been the standard answer for several years, but they tend to require large amounts of historical data, intrusive techniques (e.g. tags, beacons, and cookies), and cost more than $100K for an initial model. Econometric modelling has long been a staple for measuring media effectiveness, primarily TV, but media schedules are more diverse & complex than ever before.
Those of you lucky enough to have been at the Potentiate event in February in Sydney will have seen an alternative, one that Kimberly-Clark Australia has been using with its Huggies brand. This alternative is a light and rapid approach to solving multiple market measurement challenges, provided by a North American company called Kvantum. When I say rapid, the initial models take 2-4 weeks instead of 6-9 months.
What problems does it solve?
Examples of the sorts of problems that have been solved include:
• How should we change our marketing spend to increase its ROI?
• Which channels work best for my brand?
• Should we invest in TV outside of major metro cities?
• If a new product has been launched, was its growth new growth or cannibalisation?
• What was the role of influencers?
• What are the key emerging channels for the brand?
How does it work?
The process is linear, rapid, and works with as much data as you have. The more data the better, but it can manage with less data than has traditionally been considered necessary.
1. Isolate the impact of marketing on sales (remember there are lots of other drivers too).
2. Feed in marketing spend, marketing activity, and sales data.
3. Use machine learning to identify the roles played by each input in generating sales
4. Optimize spend levels, track the outcomes, fine tune and repeat
The Role of Prescriptive Analytics
We at Potentiate are convinced that Prescriptive Analytics will be a game-changer for marketing. The journey for analytics has moved from Descriptive Analytics (showing what is happening), to Predictive Analytics (what happens to Y if I change X1, X2, and X3?), to Prescriptive Analytics (what is the optimal change I should make?) Kvantum offers this Prescriptive approach, it suggests the optimal allocation of resources, and then learns from the ensuing data how to improve its model and recommendations.
The Case Study?
As Harpreet Singh showed at the event in Sydney, the Australian experience with Huggies was able to show the specific ROI values for key channels and deliver a 10% increase in marketing ROI. We were also told by Kvantum and Kimberly-Clark that the system had answered key questions about the campaign, its effect (by channel) and how to optimise things going forward?
Alla Nock from Kimberly-Clark said “The Rapid Read tool allows us to evaluate the effectiveness of our media in an efficient way. Having this data has allowed us to make sure we put our money into media that not only drives brand equity, but also delivers short-term sales.”
Alla Nock, Head of Consumer Insights, Marketing Analytics & Capability, Kimberly-Clark Australia & NZ
Want to know more?
Contact us today and reference 'Prescriptive Analytics' in the comments. We will connect you with the Kvantum team and provide more background on how you can bring marketing optimisation and Prescriptive Analytics into your business.