Insights from the BrandMetrics Database - Newsletter, March 2008

Database is not a term I would have used with any degree of comfort a decade ago. In fact, it’s quite possible I had never heard of it. It is a very specific computer term which the Oxford English Dictionary defines as a “structured set of data held in a computer”. Now, it looms large in my life.

Soon after launching BrandMetrics I commissioned some very clever computer people to acquire a website for me and to convert the mathematical equations that are the backbone of the BrandMetrics model to computer codes. Andrew Myers has a master’s degree in mathematics from Wits University and runs Redi Internet Services with his associate, Gary Malich. Working closely with my partner, Johann de Villiers (then a professor of corporate finance and himself a mathematician, now Dean of Commerce at Stellenbosch University), they pulled this off. I use that phrase deliberately because it was hard mental work. At one stage as they battled their way through test after test, someone asked why I wanted this so badly. “Because” I said, “one day we will have a very valuable database of normative brand measurements.”

And so we do.

Since 2000 we have conducted 360 valuations the total value of which is R670, 9 billion . That massive number includes the unique and novel valuation we conducted of our country so, correctly, we should deduct R517 billion, which was the 2006 value we placed on the nation brand (the original value of R380 billion in 2003 was conducted manually and was not captured in the database). Nonetheless R154 billion is a lot of value. And that is the structured set of data held in our server which lives (securely) with and is maintained by our friends at Redi.

Because the database only stores complete data not all 360 valuations are analysed. There are duplications, tests and examples. Once these are stripped out the number reduces to a still healthy 311. These are categorised by sector definitions under which are any number of sub-description. So, for example, I could specify fast moving consumer goods (fmcg) of which we have over 160, and break that down into analgesics, coffee, wine and many others.

It might interest you to know that the average ratio of brand value to net profit after tax for the entire fmcg category is 600%. So if your brand earns R1 million after tax, it is probably worth in the region of R6 million. On the other hand our resource category, which comprised Eskom (South Africa’s power utility), which we have valued three times, London Electricity, Virgin Energy and South West Electricity Board (SWEB) - all in the UK - have a ratio of 24%. So the total value is severely negative. Interestingly, when we valued Eskom it had quite considerable value. Now we know why.

Other ratios are: financial services, 309%; Industrial companies, 238%; retail 147% and services, 587%

If you are familiar with the BrandMetrics process you will know the terms “franchise run” and “economic expected life”. The first refers to the up-front number of years in the discounted cash flow projection and the latter to the total number of years for which we project the earnings. We have found that between 75% and 85% of the brand value is captured within the franchise run period, which on average lasts for 6,8 years (the standard deviation is 4,7 which means that two thirds of the brands we have valued fall within a range of 2,1 to 11,5 years). The highest ever was Coca Cola which was close to 20. Potentially our total projection (franchise run and decay) could run for 60 years and in the case of Coca Cola came close to this. But the average number of years of franchise run and decay for all the brand we have valued is 22, 9 years.

I should explain here to readers who are not familiar with BrandMetrics that expected economic life is not a life cycle prediction. It is not saying that the average brand will only live for 22, 9 years. The BrandMetrics algorithm is designed to project the economic life of the brand in a curve that is initially, mostly, upward curving to the extent of the franchise run and then sloping down during the decay phase. Thus we capture the total value, even if the balance of, say, 12 years only encompasses 15% of the value.

Services brands such as MTN and Vodacom record the longest average franchise runs of 9, 4 years and the shortest is fmcgs with 6 years. Two inputs to the model influence the number of years: the category and brand strength. The category variables evaluate the stability or volatility of the category brought about by external forces such as regulation, legislation, competition, the economy and the availability of raw materials, among others. Brand strength is a function of market research. We manipulate the survey data so each brand is expressed as a percentage and is therefore comparable. The strongest brand in the category sets the upper limit to the number of projected years and the weakest brand the lower limit. The proximity of the target brand to either extreme determines its expected life and franchise run.

It is interesting that retail, at 7, 7 years, is more stable than fmcgs. This will have a lot to do with the highly competitive character of most fmcg markets and the buoyant demand led economy South Africa has experienced since the turn of the century.


BrandMetrics uses economic profit as a proxy for a non branded version of the brand being valuing. A brand must add economic value and economic profit – the after tax profit earned that exceeds the firm’s cost of capital - is an almost perfect surrogate for the generic version.

Once we have worked out the economic profit we need to know what proportion of this is attributable to the brand. We devised Resource Recognition Procedure (RRP) to derive this number. All of our clients over the past eight years have been through this fascinating process. It starts by developing a list of candidate drivers of economic profit (the reason why the firm is able to generate this very desirable class of profit). The list is reduced typically to between five and seven and these are then quantified in two ways. One: to show the relative importance of each driver; two: to show the influence of brand equity on each resource. When these sets of numbers are subjected to BrandMetrics arithmetic they produce the desired percentage. In the process information is developed that is very useful to the company as a strategic tool.

The table above is our latest analysis of drivers for all the 311 brands in the database. Interestingly, the order has changed somewhat since we last pulled out these numbers from the database in 2004. Then, supply chain management was number one and cost control was three. What is significant is that “leadership” has moved significantly up the order as has “innovation”. The others are more or less as they were.

The percentages themselves have changed very little. Overall the average for all brands is 66% with a standard deviation of 11 (66% of our valuations fall within a range of 55% to 76%). Media at 85% is the highest and resources at 48, 2% the lowest. Fmcgs are 70, 5% and financials are 57% interestingly, banks are higher than insurance companies within this category, because of the influence on the latter of intermediaries such as brokers.

It’s worth breaking our rule of a three page newsletter to provide a final insight.

Over the years we have valued a number of brands which comprise the portfolio of companies on the stock exchange. It has always been interesting to see what proportion of the market premium (the difference between the net asset value and the market capitalisation) the brand (or brands) represents. While it used to be an academic exercise the new International Financial Reporting Standard (IFRS 3) that deals with mergers and acquisitions makes this mandatory. Here are three illustrative case studies.

Distell

The name Distell combines Distillers Corporation and Stellenbosch Farmers Winery. These were the two giants of the industry and they merged in the early 2000s. In 2003 we were commissioned to value the entire portfolio for Capital Gains Tax (CGT) purposes . The total value of the portfolio was R2,8 billion. At the time the market capitalisation was about R3, 45bn while the NAV was R3,2 bn. This was very uncomfortable because it was showing huge negative goodwill. I questioned it and was told this was probably a true refection. The market was waiting to see if the merger would bed down and perhaps more importantly, if the Competition Board (CB) would make any demands for unbundling when it issued its directive.

Sure enough, following a virtual clean bill of health by the CB, the market value rose to R7 billion (today it is R11 billion). Thus the brand portfolio accounted for almost two third of the premium, as you would expect from a company that relies so heavily on its brands.

Vodacom

Our request to value the Vodacom brand came about due to a spat the cellular telephone market leader was having with its closest rival. We have valued the brand twice and in 2005 placed a value on it of R32 billion. Shortly after competing the valuation, British Vodafone - already a 35% shareholder in the firm along with the government telecoms company Telkom and a specialist investment company, Venfin – made an offer to buy out the Venfin 15% minority.

Vodacom is not listed on the stock exchange but the price offered placed a value on the company of R109 billion. Since the NAV at the time was of the order of R12 billion, this offer (which was accepted) indicated a premium over the net asset value of R97 billion.

With a brand value of R32 billion, the brand was clearly a major factor in the creation of this value.

Freeworld

Feeworld Coatings listed on the Johannesburg stock exchange in December 2007. Previously it had been the coatings division of industrial conglomerate, Barloworld. The listing was the culmination of the unbundling process that saw a few Barloworld companies being spun off. Soon after its listing the new company had a market capitalisation of about R2 billion. We valued the Plascon brand portfolio for the listing (having been involved with the brand on two previous occasions). If the goodwill, shown on the balance sheet at R1, 7 billion is removed, our value of R686 million, would be one third of the market premium.

These three cases simply attest to the value that brands add to tangible assets; a component of shareholder wealth that is at long last recognised.

It is almost certain that BrandMetrics is unique in the world in being a web based valuation methodology. Which could mean that the information in this newsletter is not only unique as well, but might be described as ground breaking.