Minority shareholders are typically outside observers of businesses, and our ownership of companies via Castelnau has increased our appreciation of this point. The transparency from owning an entire business brings you far closer to reality than a minority shareholder. Without ownership control, disclosure is often very high level, via aggregate financial data that is in effect a snapshot summary, an average of what is happening across the business in a given period.
By accepting what happens on average, investors are therefore susceptible to missing what might be happening on the margin – especially important for a heterogenous business. Granular insight can therefore be vitally important, not to amass detail just for the sake of it, but to determine what matters most, to understand marginal unit economics and identify hidden operating leverage. To figure out what matters most we have to go deep.
Michael Dell once put it well: "averages obscures a lot, and aggregate financial statements are pretty meaningless. Our approach is to take really big numbers and "de-average" them. Unless you look inside and understand what's going on by business, by customers, by geography, you don't know anything". Of course, Michael Dell as CEO was inside the operation and had all the information he needed to hand. Dell already had reality on his side.
The legendary Phil Fisher called the search for reality his 'scuttlebutt method'; some investors loosely call it channel checks; we call it monitoring. It is a way of sense checking on-the-ground reality versus what a company communicates is happening. Monitoring offers us the ability to verify how the business actually works, how deep its moat actually is and whether it is widening or narrowing. It also seeks to identify sources of hidden leverage.
The inevitable problem with most corporate disclosure is that it is a summary of average activity. In reality of course, there is no average, just a messy amalgamation of often very diverse activity. There is a great story of the US Airforce in 1950 employing a survey of 4,063 pilots' body sizes to help design the perfectly sized cockpit. The researchers ultimately realised they were effectively designing for a non-existent pilot of average size – in the end they just made the seat adjustable! Analogously, out of c200m Ryanair passengers last year, we might guess there are likely very few actual customers who paid precisely the group average €49.80 fare to take a one-way flight of 780 miles average sector length. Rather, as with many companies with heterogenous pricing structures, there was more like a wide spectrum of customers paying vastly varying fares from €9.99 loss-leaders right up to the €400 ultra profitable last-minute booking. Pareto distributions, though common, are not always obvious.
In a recent interview, Geico CEO (and Buffett protégé) Todd Combs, made an insightful point when discussing his efforts to improve Geico's returns. "Where people go wrong", he said "is where they think they know their CAC (customer acquisition cost) and LTV (lifetime value) but it's from too high a level. When you slice the data, there are really big pockets where the LTV is negative, and that is where the potential is to improve margins". In Combs' specific example, by digging deep into the business, he was able to identify underperforming cohorts, which if improved or eliminated could have a material positive skew on the average. Charlie Munger used to crudely call this 'cutting out the cancer'. The opposite can also be true of course where a small portion of customers or products contribute pools of disproportionately high profits, and their loss would cause a sharp fall in the group average profit.
This phenomenon – pools of hidden profits or losses – is one of those things that, as the saying goes "when you see it once, you see it everywhere". Two examples come to mind. Adidas shareholders found out painfully late that the infamous Yeezy products, whilst only 5% of group revenues, were actually more like one-third of group operating profit! This hidden leverage became all-too apparent when the products were abruptly taken off the market for reputational reasons. Back in its pre-iPhone heyday, Nokia's Mobile Computers segment (phones with a selling price >€400) accounted for just 4% of volumes but 28% of group gross profits in 2006, the year before the iPhone launch! Investors relying solely on both companies' average performance would likely have missed the risk inherent in the business that was about to unfold.
We return to Mr Combs for the final word. Combs wrote in his foreword to the latest (and excellent) 7th edition of Security Analysis: "the difference between a Good analyst and a great one lies in the ability to keep things simple and determine what matters most. People misinterpret this to mean that's investors should keep things at the surface level. In fact, paradoxically, it takes a great deal of depth to stay simple."
