Brand-building: What Brand Beliefs Can You Really Move?

I have been involved with a ton of brand tracking over the years on well-established brands and admit that current tools provide only half the answer to the following big question: In order to build my brand, what attributes should I try to improve upon?

• The half of the answer that brand trackers do provide: correlations of attributes to brand preference.
• The half they are missing: Which correlated attributes are actually movable and how much advertising spending will it take to move them? I have never seen a quantitative answer before but now, drawing from the work I have been involved with on the Movable Middle, I have discovered a way to address this.

The math to understand attribute movability comes from the same math as OBM2 (white paper here), where we proved that responsiveness to advertising is predictable as a function of someone’s likelihood to choose a given brand. Responsiveness to advertising is maximized when someone’s probability of buying a brand is 50% and it trails off down either side. Your sweet spot for targeting advertising is the Movable Middle, those with a probability of buying your brand between 20-80%. This finding is derived from math and verified by empirical evidence.

The same math implies that, in the aggregate, brands have maximum advertising elasticity if their market share is 50% because this is when the Movable Middle is largest (solving via the Beta distribution); elasticity falls off symmetrically as you approach either a 0 share (better change the offer) or a 100% share. (If interested in the calculus regarding logistic equations and beta probability distributions, you are welcome to email me.)

Now let’s talk attribute ratings. Because we now know that mathematically, percentages closer to 0 are harder to move with advertising spending, we can solve for which attribute ratings will be most responsive to ad spending with aligned creative. We can even envision a “ROAS for attribute ratings”, where points of attribute movement are in the numerator and calculate this for each attribute. This is all based on previously undiscovered math that the movability of an attribute rating in response to advertising is maximized at 50% (falling off towards zero symmetrically as you approach 100% or 0% ratings).

And because the math is unified, we can even solve for additional ad spending required to move the chosen attributes by a desired amount. Now we have the second part of the answer!

Beyond the math, consider the logic: imagine Costco (not a client, so entirely hypothetical), has attribute agreement scores as follows:

• Sells all the brands I prefer: 10%
• Provides a great shopping experience: 50%
• Offers great value for the money: 90%

Our movability math says that the first and third attributes will be hard to move with advertising, while “great shopping experience” – with the right creative – is a highly movable attribute. Also thinking about this logically proves the point: Trying to convince shoppers through advertising that Costco sells the brands they prefer when the company’s retail model is based on limited national brand offerings is certainly silly.

Another way to think about this is in terms of brand perceptual maps. Marketing researchers opine intuitively that it is harder to move a brand from one quadrant to another, but we never had the math to back up this intuition. Now we do. It is better to try to move towards the outer edges within the same quadrant, perhaps at a somewhat different angle, but stay within that quadrant.

Following the math of attribute movability, what attributes should you try to move? Those that are…

• in a mid-range of association with your brand overall and for your Movable Middle.
• highly correlated to brand favorability.
• strategically on target.

If you focus creative on attributes that meet these criteria, and focus ad placement on audiences that are rich in the Movable Middle, you should produce superior performing campaigns that drive sales and build brands time and time again.

Wouldn’t it be cool to turn trackers into a predictive tool by having a war games simulator where you could input ad spending increases, alternative targeted attributes, and explore share and brand profile changes?

With unified math, this is now achievable.

A version of the preceding article was originally published on the author’s blog, Joel Rubinson on Marketing Research.