How to Value Stocks Like an Operator
Move beyond the usual ratios and avoid frauds in the process
In my courses, I preach to my students the importance I put on per metric units, whether they be related to valuation or operational performance. Everyone can easily understand unit economics and the lack of emphasis generally placed on them surprises me.
One reason perhaps is that they are not applicable to every sector. Retail and restaurants lend themselves quite naturally to such analysis; property companies similarly, where per square foot metrics often prove useful. But one basic unit of capacity is an employee.
Before getting to that, a word from our sponsor:
How can you get expert insights faster for 40% less?
Stream by AlphaSense offers access to our library of 26,000+ expert transcripts — plus the ability to engage with experts 1-on-1 and get your calls transcribed (free of charge) — all for 40% less than you would pay for 20 calls in a traditional expert network model. Get the ROI Guide
Employees as a Capacity Proxy
Not every employee is a producer, obviously, as organisations require staff for indirect activities, generally considered as overhead. But I have found that tracking revenue per employee is a useful tool, and I use per employee valuation as a different way of calibrating what a stock might be worth.
Per employee valuations are not always helpful – quite some time ago, I remember teasing my friends at RIT, Jacob Rothschild’s vehicle, that they were paying over $50m per employee for Dropbox. They still made a lot of money, but revenue per employee is a metric which I like to track closely. It’s especially helpful when looking at labour intensive companies like Amazon.
Amazon Revenue per Employee Trend
Source: Behind the Balance Sheet from Sentieo Data
The revenue per employee chart is a useful indicator and gives a quite different perspective from the Amazon share price and rating. Obviously, the stock has been extremely profitable to own and the stock narrative has varied significantly over time.
Labour is a key input cost for most companies and if labour intensity is increasing at this rate, it generally indicates that margin improvement will be harder to extract from customers. This isn’t intended to frame any view on Amazon – I picked it at random as a large employer and I was surprised to see that labour productivity had halved. And perhaps the takeaway here is that Amazon’s business is higher quality than its employee numbers would suggest.
Alphabet Revenue per Employee Trend
Source: Behind the Balance Sheet from Sentieo Data
I then drew the same chart for Alphabet and again I was mildly surprised that the output wasn’t higher per employee – it’s well ahead of Amazon obviously, but lower than I would have expected. And Apple is even higher than Alphabet which is not what might have been expected [1] . This perhaps illustrates the quality of Apple’s business and highlights why this metric is helpful – it forces these questions.
Apple Revenue per Employee Trend
Source: Behind the Balance Sheet from Sentieo Data
US GAAP is less helpful in the provision of employee data than say a UK or European set of accounts – we enjoy extensive disclosure of average employee numbers as well as salary and related labour cost data. Many large US companies disclose employee data in their 10-K filings, but not all in my experience. Even so, most large companies disclose the approximate number of employees on their website. While a history may not always be available, you can ask– a simple email to HR need not mention that it’s for investing purposes which might be more productive.
For companies that are less secretive about this, further disclosure of US employee data by ethnicity etc is sometimes available in the EEO-1 Employer Information Report. Not every company discloses this detail, but some do and it’s worth searching as you also get a split by grade – executives, middle managers, sales staff, administrative etc. Here is Amazon as an example, and the doubling of warehouse staff in the US during the pandemic is clear:
Amazon US Employee Breakdown
Source: Behind the Balance Sheet from EE0-1 Data
Employee numbers are a good way of setting aside conventional valuation and growth parameters. Checking revenue per employee trends and EV/employee is generally a useful exercise. Using revenue per employee measures often tells you a lot about the real trend in productivity and is a great cross-check against the margin progression – if the productivity is declining, why are margins increasing – is the company being more aggressive in its accounting, for example? Looking at asset-based measures of capacity is even better.
Units of Productive Capacity
Thinking in units of capacity can be incredibly helpful in a number of industries – for example, hotel rooms, cruise line berths, airline seats etc. I have even used number of trucks for a logistics company and GW of capacity for electric utilities. It’s better than using per employee data although the per employee data is helpful in framing productivity comparisons across companies or over time.
I find that thinking in metrics beyond P/E and EV/EBITDA helps to ground my thinking and allows me to evaluate a company more as an operator would. And in my experience, the people involved in an industry in my experience often have a better understanding of valuation than stock market participants.
My background as a transport analyst led me to focus on this – transport tends to be an asset-heavy sector and these assets are almost invariably expensive, planes and ships in particular. Thinking about what you pay per berth for a cruise stock or per seat for an airline is a good way of comparing stocks in a sub-sector. It also forces you to think more closely about how companies generate different revenues and profits per seat or berth. I find it useful and paying subscribers can read on for a great example of how you could have avoided a fraud using this technique.