Balancing long-term and short-term
Welcome to the Wheelhouse, a popular series of blogs from Ebiquity’s Marketing Effectiveness team.
In this latest edition of The Wheelhouse, Director Gavin Desir makes the case for building econometric models that balance both the long-term and short-term impact of marketing investment. He puts a particular emphasis on taking proper account of the long-term impact which many – particularly digital-first, attribution-based – models tend to overlook.
Powerful measurement techniques
As marketing mix modelling (MMM) grows in popularity and modelling techniques become more powerful, it can be tempting to focus on short-term results. Brand marketers and their partners often find it attractive to look just at the impact that comes in the 13 weeks (or quarter year) post-campaign, putting the results into an optimisation engine and using that as a basis for future campaign planning.
And yet those fully grounded in marketing science know that campaigns have the potential to change thoughts and feelings, attitudes and behaviours over a much longer period. That’s why short-term results are only part of the story and it’s crucial that long-term measurement feeds into the decision-making process to drive the best outcomes for the brand.
Why does long-term measurement matter?
Traditional MMM is a gold standard measurement technique, but it does have its limitations. One of the key shortcomings is that it skews towards the bottom of the purchase funnel, zeroing in on people who are already “in-market”. As a result, it will typically show the best return on investment for channels and campaigns which help to drive acquisitions or subscriptions – frankly, whichever short-term KPI the analyst is modelling. While MMM can also measure the impact of brand campaigns, these campaigns have a much lower impact on converting those in-market and instead work over a much longer time horizon.
A case in point
This is problematic because many industries do not work on a short-term horizon. Let’s take car insurance as an example. A campaign today will help to convert those whose policy is currently up for renewal. But as car insurance contracts typically run for 12 months and are cumbersome to get out of, this means that most people can’t switch products or providers even if they really want to.
This problem applies to all subscription-based services, but even beyond this, there are categories where the purchase cycle is just as long – if not more so – and this means that traditional MMM will underplay the full value and impact of the long-term brand-building activity. In our studies at Ebiquity, we’ve found that this additional long-term impact could be anywhere between an additional 20% to 400% on top of the ROI that’s been measured through the MMM.
This has led us to conclude that, while MMM is a useful guide to how marketing has worked in the past, without any form of measurement of long-term impact, any recommendations or future-facing optimisations need to be treated with caution. This is because focusing on the short-to-medium term only captures the impact of just four in every ten pounds, euros, or dollars invested in media.
How do we go about measuring it?
Having established how important long-term measurement is to a marketing effectiveness programme, the next logical step is to discuss how we go about measuring that impact. In practice, this is often much more difficult to do than measuring the short-term impact because it demands having much longer and deeper historical records of brand investment and performance.
Brand tracking studies are a critical dimension in the marketing effectiveness toolkit. They enable us to understand which audiences campaigns are reaching, whether these audiences understand the key messages, and whether these messages drive the audiences to take the desired action. Where these studies can fall short is in the final step of understanding how uplifts across the purchase funnel eventually manifest themselves as long-term sales.
Enter structural equation modelling
At Ebiquity, our primary methodology in this space involves understanding how brand momentum evolves over time. We then use this intelligence to build a structural equation model to understand how media impacts the various brand metrics to which we have access. Then we measure how those brand metrics directly and indirectly impact the brand’s momentum.
Because of the slow-moving nature of the metrics that brand campaigns are typically aiming to shift – dimensions such as Familiarity, Consideration, and Preference – measuring long-term effects is not something to do often. This is where knowing your media multipliers can help. A media multiplier is a ‘scaling-up’ factor to get you from short- to long-term ROI. Once you know the most relevant scaling-up factor for a brand or a category, you or your MMM provider can apply these to the short-term results to show the long-term view of effectiveness and the corresponding adjusted future facing scenarios.
Summing up
At a time when marketing budgets are under increasing pressure, understanding the full effect of each campaign and media channel on the plan is critically important. Putting a programme in place to track long-term ROI will enable the whole organisation to appreciate the impact of marketing activity in the round, leading to more informed, strategic decisions. Music to the ears of the CFO.