Brand Marketing Matters

“What’s in your wallet?” I remember an immigration officer pleasantly asking on my way back to the US from Toronto on my way back from a business trip years ago. This interaction, early in my career at Capital One, made quite an impact on me about the power of brand marketing.

As I reflect more, many such interactions I’ve had with famous brands over the years stand out to me: from “thums up-taste the thunder” growing up in India (Thums up was a Coke competitor in India; now acquired by Coke) to Nike’s “just do it” to Airbnb’s “belong anywhere.” A brand campaign that resonates with people gives a brand an iconic status that helps people think about that brand even when it’s a long cycle purchase: something I would argue matters a lot even in an AI driven world.

As a firm believer that every consumer facing company of a material size needs to make an effort investing in their brand, here are some common questions/objections that I hear from quantitatively oriented people, particularly in the financial services world, where I have spent a bulk of my career:

1. How do you measure the ROI of brand marketing? How accurate is it?

2. What percentage of your overall marketing dollars should be spent on performance marketing vs. brand marketing?

3. Is it worth spending money on brand marketing if my product purchase cycle is a long cycle?

As someone with a strong quantitative, data & credit background, but also as someone who thoroughly appreciates the importance of brand marketing this series presents my take on these key questions. This article, the first of three, will focus on the first question.

Keep reading if this sounds interesting to you! 😊

ROI of Brand Marketing

One thing that is important to recognize is that brand marketing ROI measurement can never be 100% precise. Unlike in credit modeling, where an analyst can develop a precise estimation of a customer’s chances of being delinquent for 60 days having been delinquent 30 days, marketing, and particularly brand marketing, ROI relies on a triangulation of methods to build a reasonable estimateI. A few methods that I have found to be useful are:

Media Mix Models

Many people have at least heard of Media mix models. If you are not familiar with them, these are non linear models that estimate every marketing channel’s contribution to the overall impact of marketing. These models leverage key marketing attributes like channel spend, duration & timing of spend in addition to important non marketing attributes like price of a product to come up with these estimates. Media mix models are a somewhat blemished tool because of their complicated math and dependency on key assumptions about ad stock & half-life, which are jargon for how long an ad’s impact lasts on a customer. Ad stock shape & half-life vary by channel. Even though these assumptions are grounded using actual data for a brand, there is a fair amount of judgment involved, which therefore makes it hard to rely on the outputs as facts. Building & setting up a media mix model also requires a significant amount of testing data. However, these models do a reasonable job of estimating brand marketing’s short-term sales impact. I recommend that a customer facing brand with marketing spend over $75M - $100M should consider investing in a media mix model solution.

Incrementality Testing

In this method, consider developing an A/B testing strategy focused on separating test markets vs. control markets. Ramp up your controllable brand spend in the test markets while keeping your spend at a BAU level in the control markets. This type of a test is hard to pull off particularly if your company is a national TV advertiser because there aren’t easy ways to run A/B experiments with national TV buys. You can attempt to address this gap by purchasing local TV ads in the test markets over and above the national spend level. However, this type of test is easier to pull off if the brand spend is primarily focused on social media, out of home advertising and addressable TV. This test is only worthwhile if your organization commits a year to it. The key metrics measured are: increase in aided & unaided awareness, increase in leads or number of applications and increase in sales. Some important watchouts are: Test & control markets should be sizable enough & comparable enough, need to work with your agency partners to execute this test as intended and be patient for the test reads to bake!

Empirical Correlations

Lastly, for a brand committed to a multiyear brand journey, a strong empirical way to establish brand marketing ROI is to correlate increases in unaided & aided awareness to your sales outcomes. In my experience, my teams have been able to establish this type of relationship to say something like: a one point increase in unaided awareness leads to X points of sales lift. We have also been able to derive a relationship between brand marketing & performance marketing to say that performance marketing is A% more efficient in markets that have at least Z% unaided awareness. In my view, this is the place that an organization should aspire to evolve to be recognized as a top-notch marketer!

I will end this article by saying that brand marketing matters & yes, there are some tangible ways to measure it’s effectiveness provided you are patient, committed & comfortable with triangulation of estimates rather than a precise answer!

Have questions? Want to discuss? Gautam Puranik (Gautam@arica.co)