Contact Us

Brand Halo: The Elephant in the Room

The good news is that it isn’t that hard to get him to move

We’re all at a wine tasting, eating cheese, making contacts, posting pictures on Facebook. We’re having such a good time that no one has noticed the rather large elephant standing quietly off to one side. Perhaps, I should say, some at the party may have noticed, but no one is willing to spoil the fun and point out there is an elephant standing directly in front of the restroom door. Sooner or later, though, after enough wine is consumed, it will become painfully obvious that there is an elephant in the room and his presence has become more than a little inconvenient.

That elephant is called brand halo.

Brand halo can be defined as an overall brand impression that biases all individual brand ratings. For example, if you have a deep affection for the Apple iPhone, you might tend to rate the iPhone very high on all brand imagery statements, even those where it doesn’t deserve high ratings (e.g., GPS functionality). Like a mother doting over her only child, all you can see is perfection.  Perhaps a more mundane but pervasive example is in consumer packaged goods. When comparing a CPG national brand to private label, you’ll often see a dramatic example of brand halo. The national brand, benefitting, presumably, from millions of dollars of advertising year after year, will be rated higher than the private label on all imagery statements, including “value.” You will typically see the national brand being rated in roughly the same range on all attributes except those on which it truly does not have strength, such as value. But even there, with the steep dip downward, the ratings may still be higher than the private label. It is arguably not the case that the national brand is a better value than the private label, particularly in commodity-like categories such as spices or sandwich bread, but this pattern remains.

One could visualize brand halo as in the figures presented here. Figure 1 shows that the national brand has scores higher than the private label but also shows a big dip for value. Figure 2 shows the national brand scores with a constant removed and/or a constant added to the private label brand. The brand profiles, the shape of the red and blue lines, remain the same in the two charts. What has changed is the “gap” between them. In this overly simplistic illustration, that gap is the brand halo.

Figure 1
Figure 2

“So what?” you say. Who cares if brand imagery ratings are inflated (or deflated) because of an overall “feel” for the brand? Well, anybody who wants to understand what the brand really stands for and, more importantly, what really drives brand purchase, that’s who.

That elephant really starts to get annoying when identifying brand drivers, that is, those perceptions that influence a consumer to buy (or not buy) a specific brand. Brand halo can be viewed as a latent factor that influences all brand perceptions. As a result, brand ratings can be highly correlated with one another.

The most common way to identify brand perceptions that drive purchase is to build an OLS regression. But collinearity can make that model completely misleading (even to the point of reversing the signs on coefficients). The standard “fix” is to factor analyze (with an orthogonal rotation) the brand ratings. Even though this creates factors that are not collinear, this approach can fall short as well. It’s a brute force approach to handling the collinearity that sometimes fails to capture the underlying structure.

I ran a brand driver analysis for a large retailer a few years ago. I took an imagery battery and store visitation behavior and plugged them into a simple stepwise regression model. If you ignore item collinearity, you conclude that service, price and distance to store are what drive the decision to visit the client store. If you run the regression with orthogonal factors, you get a similar conclusion.

But these models ignore the elephant. The simplest model that I’m aware of that can accommodate brand halo is a structural equation model (SEM) which allows us to define a latent factor that all brand statements load on (the elephant) and to include that factor in the regression.

If you run a SEM on my data here, with a brand halo latent factor that loads on all brand rating statements, you see that the brand halo, not any specific store traits, far and away is the most important driver of store choice.

It’s interesting to note that, with a brand halo latent, I still found the same latents uncovered in the earlier exploratory factor analysis (service, price and others). But in the SEM, service and price had no effect on store selection. Brand halo ate up all the explained variance availabe to service and price. Further, in the SEM, easy access was shown to be more important than that simple distance to store. By including brand halo in the model, the conclusions are dramatically altered.

There are other, more elaborate approaches to accounting for brand halo. William Dillon, for example, has written several wonderful papers describing an analytic technique he has developed. I know of other, equally elaborate approaches. But at a minimum, SEM allows you to account for the elephant. It is relatively simple to use, and the software is widely available.

So no excuses. It’s time to stop pretending the elephant isn’t there. He is standing in front of the restroom and you’ve had a bit too much wine. The good news is it isn’t that hard to get him to move.