Once upon a time, marketing scientists were focused on changing brand behavior as regards the use of price discounting, otherwise known as promotion. IRI (today Circana) presented BehaviorScan findings which showed that advertising heavyup trials increased brand profits 40% of the time, vs. only 20% of the time for promotion trials. Larry Light and Mike Donahue showed evidence that brand loyalists bought the brand regardless of price and constituted more than half of the brand’s profits even though they were more like 10% of the brand’s buyers.
These things didn’t have the desired effect of reducing the use of promotion. Brand managers knew that promotion caused spikes in sales whereas advertising had harder to read effects. The more obvious effects of promotion had psychological benefits within advertiser organizations.
Byron Sharp’s (and Peter Drucker’s, and Mike Von Gonten’s) exhortation to get new customers to a brand also worked against Mike and Larry’s idea of focusing on existing brand loyalists.
Today Mike Donahue is still extolling the value of brand loyalists and is also trying to get advertisers to measure the ROI effects of Influencers, Experiential Marketing, Purpose Marketing, and eCommerce, all worthy causes. He groups the latter four categories of stimuli under the rubric of Brand Activation, and points out that Brand Activation investments are 2.5X greater than advertising investments globally. Indeed, there continues to be a trend toward short term performance results over long term effects, and my invention of big data singlesource (TRA, now widely emulated) contributed to this trend. So did digital, which all existing data show to be better at activating current purchasers of a brand than bringing in new buyers to a brand, vs. television which all existing data show to be better than digital at bringing in new buyers to a brand. Digital and Mike’s four versions of Brand Activation all act more like promotion in that sense (except perhaps Purpose Marketing).
As I explained to Mike recently, he is thinking of a dichotomy when a trichotomy would be more realistic. Mike’s dichotomy is brand loyalists who are profitable vs. brand occasionalists who only buy on deal. This is not what TRA found to be the case in reality. TRA found three types of brand buyers, loyalists, occasionalists who buy only on deal, and occasionalists who without the need for discounting are the most sales responsive to advertising – we called the latter group “swing” purchasers after the political notion of “swing voters” – the influenceables where persuasive stimuli can make the biggest positive difference in outcomes.
Since TRA had frequent shopper card data going back for years in over fifty million US households, including what they paid for each UPC purchase they made, it was easy to identify the folks who bought on deal (and those who bought “green” etc.), and distinguish them from the swing purchasers who would pay full price for the brand, if the new ad reminded them of positive experiences with that brand from the past, and caused them to buy it again.
I frankly had been looking for a way to distinguish that group for decades as a result of an experience I had in my first few years in the business. Norton Garfinkle had invented the idea of syndicated studies which measured many product categories and also measured media, so that product and brand usage data could be used in media selection. His legacy today is seen in MRI/Simmons, TGI, Vividata, and other popular syndicated services around the globe. Norton’s questionnaire showed a list of brands within each product category and asked the respondent to put a “1” next to the brand they buy most often, and a “2” next to their second choice brand, and a “3” next to all the other brands they bought in the category.
One year, shortly before I became Executive Vice President of Brand Rating Index (BRI), the name of Norton’s company, BRI reinterviewed respondents who had been interviewed six months earlier. I became fascinated with analyzing the results of that reinterview study. What I discovered was that more than 4 out of 5 of the people who switched their #1 brand from their first to their second interview, had switched to the brand they had considered their #2 brand six months earlier.
This fired me up with a strong intuition that advertising probably was largely responsible for causing these switches from #2 brand to #1 brand. Why did I feel that way? Because I knew that advertising is a “light tap” according to Erwin Ephron, one of my idols along with Ed Papazian and my many other mentors, not as powerful as the recommendations of an unbiased friend or relative, so that other than during new brand launches, advertising had a hard time getting people to try a brand they had never bought before. But if someone had a positive experience with a brand in the past but had not been buying it recently, advertising could bring that person back as a buyer. It took me many years to come up with a way of measuring and making use of that guesswork on my part. TRA won 77 major brands on the power of swing purchasers and the ability to target them based on the TV programs they watched and the websites they visited. The average ROI increase produced by this method was +28%.
The TRA user interface enabled the planner/buyer to set the target based on heaviness of category usage, specific brand usage, percent share of requirements the client brand represented for the household, and price buyers could be filtered out.
Today a more modern approach would assign different weights to different targets, so that all prospects would receive some weight appropriate to the likelihood of their adding an incremental sale in response to the specific ads being used. RMT is the only service currently on the market (and validated by five separate disinterested research organizations) which enables a measurement of the resonance (alignment, congruence) between a specific creative execution and a specific individual or household, and also between that specific ad and each alternative media context. The true optimizer today would cover all marketing stimuli, would measure swing and price-buying characteristics, and would measure the resonance between each stimulus and each target, and the resonance between each stimulus and each context. In CPG, Cars, and other categories the distance between shopping trips could also be applied at the individual household level within such an optimizer. As taught by Howard Shimmel and Dan Aversano, the optimizer would be buying projected future audiences not using historical data per se.
Joel Rubinson has done the most advanced work on the shopping cycle approach, and also discovered what he calls the Moveable Middle, which is the same as what I had called “swing”. Leslie Wood called the same phenomenon “Category Heavies/Brand Lights”. They probably are aware of the need to filter out the price-buyers and the ways that can be done.
The distance between the scientific possibilities for marketing and the way marketers actually operate is still quite large. Hopefully in the years ahead we can significantly narrow that gap. Before we can cause such major behavior changes we shall need to prove that our measurements of incremental sales and profits are accurate enough to base decisions upon.
Bill Cook who was Executive Vice President of the Advertising Research Foundation (ARF) at the time, led an ARF Research Review of TRA, and while giving TRA high scores also agreed with my thought that the big data singlesource method (as well as marketing mix modeling) needed to be validated by random control trial experiments. Alas, more than ten years later this quest for hard scientific validation of our industry’s ROI estimation models is still in abeyance. Marketing Science Institute (MSI), now part of ARF, is engaged in a study of MMM which could be where this important change comes from, and I know that CIMM is looking at it as well. The explosion of interest in measuring attention values of ads and media contexts is also generating a more inclusive reconsideration of impression quality measures including resonance (alignment/congruence between ad and person and/or between ad and context), brand attraction (EEG frontal right-left asymmetry), positive emotion, long term memory encoding, specific brand perceptions.
Things happen fast and slowly in our industry. The more scientific, the more slowly. The less scientific, the faster. It’s our industry, we can change it.
Posted at MediaVillage through the Thought Leadership self-publishing platform.
Click the social buttons to share this story with colleagues and friends.
The opinions expressed here are the author's views and do not necessarily represent the views of MediaVillage.org/MyersBizNet.