Contribution margin per unit sold, clear separation into fixed and variable costs so that we can identify platform building costs and changes in customer lifetime value broken down into vintages of acquisition will go a long way in helping investors separate the wheat from the chaff.

The dot-com boom spawned several noteworthy network businesses. Those that emerged stronger from the bursted bubble did so by investing in gaining market share as quickly as possible. Examples include: Amazon, Facebook, ebay, Etsy, Alibaba, Twitter, Instagram, Linked In, Pinterest and Airbnb, Uber and Lyft. 

When these companies and their less successful peers saw a number of users reach a tipping point, an older customer would implicitly get a new customer—leading to a virtuous cycle resulting in runaway winners and left behind losers. Let’s look at this in practice. If I am searching for an alternative to a hotel room in a new city, I head to Airbnb, not to VRBO or because owners of apartment rentals in the specific area I am interested in that suit my preferences (i.e. close to a Whole Foods store and organic restaurants) are more likely to list their places on Airbnb. Because people like me look for such places, apartment owners who have units that fit my profile are more likely to list them on Airbnb. The greater the number of riders on Uber, the more attractive Uber becomes to drivers. Facebook started as a free online service in 2004, launched advertising in 2007, went public in 2012 and entered a hyper-growth phase in 2014.

However, the current financial reporting model obfuscates, rather than informs, capital providers about the economics of these businesses. Let’s consider Facebook as our positive outlier. In 2008, Facebook reported revenues of $272 million and a loss of $56 million. In 2014, revenue came in at $12.4 billion and income at $2.94 billion. Those numbers had grown to $86 billion in revenues and $29 billion of net income by 2020. The company has become successful enough to attract antitrust interest. 

Facebook was either lucky or skilled enough to hit the tipping point early in its life. There are many businesses which keep investing and hope for that magical tipping point moment to eventually arrive. Sometimes, they buy unsustainable growth by giving away the store despite recurring evidence that their customers expect steep discounts. Meanwhile, they urge capital providers to stay patient and blame the absence of long-term capital as a barrier to innovation. The current reporting model is particularly unhelpful at enabling investors identify such losers in a timely manner. 


Consider Uber before the confounding influence of Covid. For end-of-year 2019, Uber reported revenues of $13 billion but incurred losses of $8.5 billion. How much of these losses reflect “investment” losses, as opposed to incentives to attract fickle customers, or mounting operating costs whose benefits expire within the year? Would an alternate reporting model help an investor isolate platform building costs and separate business models that are genuinely making progress towards the tipping point from others that are mere pretenders?

Anup Srivastava, my co-author, suggests a focus on unit economics and sensitivity of that metric to projected scale in the future. Building on these thoughts and my prior work outlining a more ideal reporting model, I propose a model that asks CEOs to report the following line items:

·      Revenue (price per unit * number of units sold * foreign currency fluctuations)

o  Break down by new customer and repeat customers. Unless repeat buying increases, reaching the tipping point in a reasonable time frame becomes difficult.

·      Direct variable costs associated with sale (variable cost per unit sold * number of units sold* foreign currency fluctuations). In Uber’s case, drivers’ compensation per ride, credit card fees per ride.

o  Break down by new customer and repeat customer. If the firm keeps providing incentives for customers to come back, the tipping point becomes elusive.

·      Contribution margin per unit, defined as the difference between revenue per unit and direct variable costs per unit.

·      Indirect variable costs such as server time with Amazon Web Services.

·      Period specific costs that are not central to the creation of the potential network or platform (example: utilities, rent, administrative expenses). Provide detailed breakdowns of these costs by category.

·      Capitalize the costs that the manager believes creates the network or platform to the balance sheet.  Provide detailed breakdowns of these costs by categories in terms of R&D, selling or brand building expenses. 

·      Customer reporting: Concurrently and perhaps most important, report the total number of customers serviced that year, the number of new customers, the number of repeat customers and the frequency distribution of their repeat buying by customer vintage (i.e., how many purchases and the dollar amount of purchases this year by customers acquired in 2017, 2018, 2019 and 2020), the number of customers lost this year.

·      Disclose and clarify: Disclose the projected lifetime dollar value of the customer and associated acquisition costs per customer. Clarify the actual and projected cross-selling of goods and services across customer vintages. For instance, an Amazon customer who came in 1999 first to buy books eventually bought music, video, groceries and so on. An Amazon customer who came in 2009 perhaps never bought a book. This analysis will also shed light on the portion of customer portfolio that could not be reactivated into buying again.

A somewhat crude (no pun intended) template of this reporting scheme is the lifetime value of oil reserves that exploration and production companies are expected to disclose. For instance, Exxon reports the net proved developed and undeveloped reserves of oil at the beginning of the year, revisions in those estimates due to improved or declined recovery of oil from those reserves, purchases and sales of those reserves, extensions and discoveries, and the fall in these reserves due to production of oil this year. These revenue-based numbers, in fair value dollars and number of barrels of oil, are supplemented by data on future costs. Exxon discloses the present value of future cash inflows from sales of oil, future production costs, future development costs and future tax expenses. The discount rate in oil is set to a standard 10%.

For a company like Uber, the analogy would work out as the fair value of customers relationships at the beginning of the year, upward or downward revisions in those estimates, purchases of customers via acquisitions, new investments made this year, impairments (if any) in the potential value of such relationships, expected future production and development costs and future tax expenses. 

Another wrinkle. Consider disclosing these estimates during the best, worst and expected scenarios in the judgement of the CEO with adequate disclosure of the underlying parameters of each scenario. Companies will pushback that this is unnecessarily burdensome. I would counter that the board of directors and the C-suite should have access to this information to manage the company well in any case. In essence, I am calling for disclosure of more management accounting or control and planning type data produced inside the company to outside capital providers.

How will this help an investor kick the tires and learn about Uber’s progress or lack thereof towards the magical tipping point? The investor can assess whether the contribution margin per unit is rising over time. One can get insight into the potential break even point or the number of units that need to be sold for contribution margin per unit to cover indirect variable costs and period specific fixed costs. Once the tipping point is reached, the rate of growth year over year on platform costs and other capex will fall. Net income per unit will grow and will more than pay for the platform costs capitalized on the balance sheet.

Without a revised reporting model, businesses will continue to spin narratives that losses on the income statement merely reflect investment losses and peddle the falsehood of the elusive tipping point being within striking distance. In addition to my attempts to create an early design of a future reporting model, I am somewhat surprised that sell-side analysts don’t push management to provide such granular data. Suggestions are welcomed on alternate reporting structures that will help investors finally separate the wheat from the chaff of network businesses.