A copy of this memo was first published here by the CFA Institute. April 2021.
GAAP sometimes misrepresents business reality. Let’s use that fact to generate some alpha.
Continuing from the first memo, we’ll start by examining revenue recognition, the cash conversion cycle, and free cash flow.
“Revenue” Isn’t revenue (it’s contract timing)
Revenue is recognized when a contract between a business and a customer has been performed.
Here’s how it’s done according to FASB:
Fig 1. The Revenue Recognition Process
There are several areas where GAAP revenue recognition can hit a snag and you can find an opportunity.
1. Multiparty Transactions
In multiparty transactions, “revenue” can mean gross revenue dollars in a transaction or a subset that is recognized as one company’s net revenue. Your last $20 Uber ride probably generated $16 in net revenue for the driver and $4 in net revenue for Uber.
Net revenue can get distorted when multiple parties transact before an end customer receives a product. Imagine that a drug manufacturer controls a distributor and the distributor increases its orders in anticipation of end customer demand. These new orders puff up the manufacturer’s net revenue numbers. But what if the end customer demand doesn’t materialize? The manufacturer’s reported organic revenue growth might just be pulling forward future revenue and stuffing it into the distribution channel. These category definition games can present traps for growth investors and potential alpha for shorts.
2. Changes in Performance Criteria
When performance criteria change, reported revenue can become an unstable metric. For example, the same software sale can result in different GAAP revenue numbers depending on whether it is structured as a license or a subscription. Subscriptions show less GAAP revenue early on, whereas licenses may bundle a multi-year contract into a single quarter of revenue. That’s why the perpetual-license-to-SaaS transition is a popular private equity play: You can take a company private to change its accounting standard outside of the spotlight, then bring the company public with freshly cleaned books and a new story. Companies that do make this kind of transition while public, like Adobe, can present meaningful alpha opportunities for investors who understand how the future accounting will turn out.
3. Multiyear Contracts
Should it matter if a transaction is recognized on December 31st or January 1st?
Companies want to report strong year-over-year growth for each period. Savvy customers wait until the end of a quarter and then ask for a discount to book a transaction before the period ends. It’s similar to buying a used car after Christmas from a salesman who is desperate to make his year-end quota. In bad scenarios, a company can get caught pulling forward discounted demand every quarter just to chase last year’s numbers. In the worst case, that company will run out of future demand to pull and their sales pipeline will fall flat.
But GAAP doesn’t make it easy to distinguish between temporarily pulled forward contracts (noise) and increasing customer demand (signal). This is also true in reverse — GAAP revenue doesn’t differentiate between slowing customer demand (signal) and temporary sales delays (noise).
Private investors can look at the sales pipeline (which will never be publicly disclosed) and what I’ll call “the contract term structure.”
Fig 2. The Contract Term Structure
What you’d really like to see in GAAP is annual contract value (ACV) and total contract value (TCV). ACV is the amount of business currently under contract for that year — whether it’s already recognized as revenue, invoiced but not performed, or contracted but not yet invoiced. TCV includes contracts and invoices for future years. With ACV and TCV, you could see revenue recognition within the context of the full sales picture.
But any FASB proposal to add the contract term structure to GAAP would meet with stiff resistance. What CFO would want to let the competition know how they are doing? So the contract term structure will likely stay hidden and, thus, be an excellent spot to hunt for opportunities.
The cash conversion cycle should be measured as a percentage and include deferred revenue
The cash conversion cycle (CCC) measures how long each dollar of working capital is invested in the production and sales process of an average transaction.
The idea is to track working capital efficiency from the cash paid to suppliers to the cash collected from customers.
Fig 3. The Cash Conversion Cycle (Current Formula)
The CCC is like a mini return on equity (ROE). Each driver can be improved in order to increase the return on working capital. But unfortunately, there are three flaws with the current CCC metric.
The first problem is that the CCC is calculated in days. What we’re really measuring is capital efficiency over a period of time, usually a year. That’s a ratio. Nobody calculates ratios in days. We should measure the CCC as a percentage.
The second problem is that its terms are under-inclusive. The CCC currently includes accounts receivable (cash owed by customers), accounts payable (cash owed to suppliers), and inventory (cash paid in advance to suppliers). For the purposes of the CCC, "inventory" should be renamed "prepaid opex" to include all capitalized operating expenses directly tied to the transaction (e.g. deferred compensation to salespeople).
The third and most critical problem is that a term is missing -- current deferred revenue (cash collected in advance from customers). It’s easy to see the CCC’s oversight when we look at the other working capital line items related to customers and suppliers:
Fig 4. The Cash Conversion Cycle Should Include Deferred Revenue
Updating the CCC makes it easier to identify capital-light businesses.
Businesses that collect cash from their customers ahead of contract performance (deferred revenue) can be highly cash-efficient. But if the CCC excludes deferred revenue, then investors might overlook that these businesses can grow revenue at a GAAP net income loss without dilutive equity raises. This omission may explain why SaaS and consumer subscription businesses were misvalued five years ago (along with being misunderstood durable customer relationships). If you can find the parallel today, you’d be like the public SaaS investors of 2016, well ahead of the curve.
The updated CCC also makes it easier to flag the dreaded SaaS death spiral. Quickly growing companies can also be fragile when they depend on deferred revenue to meet ongoing cash needs. If their GAAP revenue growth peters out, they may rapidly find themselves in a cash shortfall. Bizarrely, these companies can show excellent GAAP revenue growth numbers while teetering on the edge of bankruptcy. If the CCC doesn’t include deferred revenue, you won’t be able to see the canary in the coal mine.
“Free cash flow” isn’t free cash flow (it’s an accrual metric)
“Free cash flow” doesn’t always equal the actual cash generated by a business.
This raises a problem for academic finance because the keystone model for stock valuation is John Burr Williams’ discounted cash flow (DCF) analysis. You might ask, "if investors can’t reliably measure free cash flow (FCF), how can they reliably discount and value those cash flows?" Good question.
Here’s the standard definition for free cash flow:
Fig 5. The Standard Free Cash Flow Equation
This all seems straightforward until you look at how much discretion goes into the accrual numbers for a given period and how much those accrual numbers impact FCF.
Fig 6. Why “Free Cash Flow” Might Not Be Free Cash Flow
Internally-developed intangible assets are the danger area in today’s market. Most investors agree that we should capitalize some portion of R&D and SG&A expenses, but no one is sure how long these intangible assets will last. Google’s search engine should endure in some form for decades to come; AskJeeves, not as likely. At the end of their useful lives, those internally-developed intangibles are also harder to sell tangible-capex assets -- try to value AskJeeve's old code base vs AskJeeve's old office space. So how can we come up with a consistent rule to amortize the Google and AskJeeves engineering efforts ex-ante?
To make matters worse, intangible capex may be hidden in cashflow line items that aren’t included in FCF calculations. If you look closely, a company’s acquired intangibles and financed leases might just be capex in disguise.
Properly accounting for internally-developed intangibles may be the most significant unsolved problem in GAAP.
Lastly, companies have a strong incentive to pump up their perceived equity coupon. Investors who focus on free cash flow yield often analogize equity dividends, and thus FCF, to bond coupons. But because current FCF is chock full of these accrual assumptions, we can’t naively project current FCF to estimate normalized FCF. A juiced FCF yield is akin to a shaky bond with a high yield, also known as a fool’s yield.
The alpha opportunity is identifying when normalized FCF will differ significantly from current FCF. Stocks where the company needs to cut the equity yield — be it dividends, stock buybacks, or debt payments — can be good shorts. Long opportunities can arise when a major portion of current capex, R&D, or sales spend flips to an amortizable fixed cost. The real difficulty is ensuring that the fixed asset you are betting on isn’t about to become stranded — lest you end up backing AskJeeves instead of Google.
Moving to the Balance Sheet
Here’s how the puzzle pieces begin to fit together for longs, shorts, and entrepreneurs:
We can recharacterize the balance sheet too. From there, we can revisit the weighted average cost of capital as well as the market value of equity and share-based compensation.