Business Strategy For Data Scientists: Brand Valuation
Learn How To Value Brands And Other Intangible Assets
In the business world, there are tangible and intangible assets. Tangible assets like vehicles and factories are easy to value – we can see and touch them, and they are used (by businesses) to produce measurable cashflows with which we can estimate a reasonable price for it.
Intangible assets like brands, the focus of today’s post, are much harder to value. We can’t see or touch them. And they don’t directly throw off any cashflows or tangible benefits with which we can value them. But in many cases, the brand is the business – the business is only able to produce the level of profits that it does because of its brand.
So given that brands can be and often are the primary driver of a business’ value, it becomes quite important to be able to estimate its value. For example:
- Investors need to know how much a brand is worth so that they can figure out whether a stock is over or under valued.
- Data scientists and analysts should know the key levers that make a brand more valuable and powerful and how that impacts their company.
How Do We Measure Something We Cannot Directly Observe?
Given that brands are intangible, how can we measure how much they’re worth? Let’s brainstorm a few potential solutions:
- We can estimate how much it might take a new company to create a brand of equivalent quality from scratch.
- We can compare the financial and stock performance of companies with strong brands and those with weak brands within an industry.
- We can do a first principles breakdown of how the brand improves key metrics like customer lifetime value (CLTV) and customer acquisition cost (CAC).
Notice how each entails a lot of uncertainty. Unfortunately with business valuation in general and brand valuation especially, there is no quick way to arrive at a reasonable answer. Instead, we need to investigate the business deeply so that we understand the key drivers of the brand’s value, which then allows us to triangulate towards an estimate of the brand’s value.
Why Have a Brand at All?
Before we look into the 3 methods mentioned above, let’s first quickly go over the business goals that firms try to achieve with their brands.
Premium Image
Of all the reasons for a business to invest in a brand, this is probably the most well known and well researched. A luxury or premium brand allows a firm to sell its product for a higher price due to the perceived scarcity of the product as well as the "status" that it confers (Apple, Tiffany & Co., Louis Vuitton).
Note that premium doesn’t have to mean luxury. It could also signal a good tradeoff between quality and price (Costco) or reliability and safety (Volvo) or consistency (In-N-Out Burger).
Familiarity and Emotional Connection
Another reason companies invest in brands is to separate themselves from competitors in commodity markets (where there are no real differences between each firm’s product). This is generally tougher to do as each firm is attempting to use its brand and marketing to forge an emotional connection with potential customers. And its ability to make this connection has almost nothing to do with the actual product (gas from Shell is the same as gas from Chevron) and everything to do with its marketing savvy (plus luck)). But if it can successfully do so like McDonald’s or Starbucks, it can greatly increase the demand for its products. That’s because consumers faced with very similar alternatives will opt for whatever is most familiar or makes them feel the most warm and fuzzy (assuming that they have been effectively marketed to).
Recap
So to recap, companies attempt to achieve one of two things with brands (or both if they execute incredibly well) – higher prices thanks to a premium image or higher quantity sold thanks to familiarity and emotional connection.
Cost of Building From Scratch
This is probably the easiest to estimate, but also the least helpful. We can look at how much companies like Apple, Pepsi, BMW, Louis Vuitton, etc. spend on marketing each year. We would probably want to scale up the annual dollar amount to reflect the fact that we are starting from zero (whereas those companies that I just named are spending to maintain and strengthen already established brands). But there are two wrinkles.
Sidenote: if the company you are analyzing is a discounter, where its brand indicates low prices, then you should consider including expenses associated with maintaining those low prices (discounts, promotions, buy 1 get 1 free, etc.). Similarly, if a company’s brand is built around technological innovation, then you would want to include at least some of its annual R&D spending. Oftentimes, the cost of maintaining a brand does not reside purely on the marketing line of the company’s income statement.
Wrinkle 1
Just like how spending a ton of money promoting and marketing a movie doesn’t guarantee a hit, ramping up your firm’s marketing budget doesn’t mean that you will definitely have a strong brand.
Even with a huge marketing budget, there is still an element of chance and luck (and execution). Despite all our spending, customers might not identify with our brand or message. If spending $20 billion guaranteed you Apple’s brand, then half the companies in the S&P 500 would do so.
Thus, we need to account for this uncertainty if we decide to use cost to estimate brand value. The most obvious way is to estimate a probability of success and adjust the value based on that probability. For example, if you decide that it costs $5 billion to build the brand from scratch with a 50% chance that you will be successful, then your estimate of the brand’s value is $5B/0.5 = $10 billion.
Wrinkle 2
The second wrinkle is that we are focusing only on the cost side of things, which can be problematic. For example, if all the players in the industry overestimate the advantages of having a strong brand, they would all end up overspending. This would cause our cost based valuation to be overstated. A brand is only as valuable as the advantages that it confers. Thus, any valuation attempt that ignores the benefits will most likely be off the mark.
Strong vs. Weak Brand Comparison
Ideally, we would identify companies (in a single industry) that have strong brands and ones that do not. Then we could compare their financial performance and market capitalizations (the amount in aggregate that investors are willing to pay for their shares) against those of their weaker counterparts.
For example in an industry like luxury goods, we can rank order, albeit with some subjectivity, the players in terms of brand strength. We can then compare the revenue growths, profit margins, returns on invested capital, etc. of the strong and weak brands.
It helps to look across industry as well. Some of the more interesting analyses will probably come from comparing a commodity industry like energy (where there are effectively either no brands or very weak ones) with a brand-driven one such as luxury goods.
If we want to value a brand this way, we might consider using a statistical model. Linear regression is the one that comes to mind for me, where our Y variable is the market capitalization of each of our firms and our X variables are all the factors that we want to control for plus a categorical variable for whether the company owns a strong brand or not. Some care should be given to which factors we control for. The parameter of interest is the beta coefficient on the categorical variable that denotes whether the company has a strong brand or not. That specific regression beta tells, us all other things equal, how much more investors value a company with a strong brand over one without one.
But here is where we must be careful. The value of the beta is accurate only if all other factors are held constant. But if we included X variables that are driven by the presence of a strong brand, then we can no longer take the beta at face value.
For example, let’s say we found that the beta coefficient is $5 billion. That means, all other things equal, an investor is willing to pay $5 billion more for a company with a strong brand. But then we notice that one of our X variables is annual sales. It’s pretty reasonable to assume that a strong brand can drive higher annual sales (that is after all one of the points of having a brand). But that means that we can no longer say "all other things are equal". Rather, some of the value attributed to a strong brand (the beta on the strong brand categorical variable) has most likely leaked to the beta on the annual sales variable and any other variables that are impacted by brand. This would cause the regression to understate the value of our brand. One way to address this is to explicitly consider interaction effects by including interaction terms in our regression.
There’s a lot we could learn by building such a model. So we may as well build it in a future post. I’ve been meaning to write an "Understanding Linear Regression" post anyways. So we will save the details for another day.
First Principles Valuation
This is my preferred method. It’s less error prone than just focusing on the cost and more intuitive and direct than the linear regression I suggested earlier.
Still, it requires us to estimate many things. But the benefits of using a first principles approach is that it forces us to understand how brand and business interact.
In an earlier post, we introduced the two key metrics: customer lifetime value (CLTV) and customer acquisition cost (CAC). This part should be reasonably obvious – a strong brand should:
- Increase CLTV by some combination of higher customer spending or stronger customer loyalty. A strong brand should lead to some combination of higher transaction frequency, higher prices for the company’s products, and increased customer loyalty (longer average customer life). Any of these would lead to an increase in CLTV.
- Increase CAC – a strong brand makes it much easier to attract new customers. When your company’s brand is strong, that is an advertisement in itself and it massively increases the marketplace’s awareness of your goods and services.
As your brand becomes stronger and stronger, we hope to see CLTV and CAC evolve like in the following picture:
The tricky part as always is figuring out how much of the improvement in these two metrics that we can attribute to having a strong brand. The amount varies from business to business and from industry to industry.
Ideally, we want to compare for the same company how CLTV and CAC appear in a world where it has a strong brand and an alternate one where it doesn’t (and everything else is more or less the same). But a brand either is or isn’t (so philosophical right?), so there’s no way to actually make that comparison. Thus, we need to resort to a survey approach again:
- Make a list of companies with low to zero levels of brand power. This list is our control group – our best estimate of what the business would look like if it did not possess a brand.
- Then look at each company’s CLTV and CAC (if you are purely relying on publicly available financial statements, then you will need to make some estimates as some companies don’t directly publish these numbers).
- Compare our company’s CLTV and CAC to the average CLTV and CAC of the companies in our control group. The difference in the metrics between our company and the control group average can be primarily attributed to our brand. We might want to scale down the calculated difference a bit as some of it could be attributed to things like better execution or more discipline regarding expenses.
- Finally, once we have an estimate of brand benefit on a per customer basis, we can multiply it by the number of customers that our company has.
Until Next Time
There’s no 100% correct answer in brand valuation. It’s definitely an inexact science. And let’s face it, there’s no benefit from trying to be overly precise – the difference between estimating our brand to be worth $18 billion versus $20 billion is not a significant one (and not worth losing sleep over). It’s more important to understand the advantages that the brand confers on our business.
Hope this was helpful, cheers!
More Data Science and Analytics Related Posts By Me:
Business Strategy For Data Scientists
Business Simulations With Python
Share This Article
Towards Data Science is a community publication. Submit your insights to reach our global audience and earn through the TDS Author Payment Program.
Write for TDS
