Thoughts On Sales and Marketing Effectiveness and Return on Capital

The best companies in the world are those that can invest large amounts at high returns on capital.

ROIC and RONIC (Return on Invested Capital, and Return on New Invested Capital) both use measures of "invested capital", where "capital" refers to capital expenditure and to a lesser extent working capital. (footnote 1)

In new economy industries though, and software in particular, “investments” take not the form of “capex”, but more often sales and marketing expenses to build up a customer base that will then bring recurring revenues. Salesforce.com, for example, persistently “invests” 40%+ of revenue on sales and marketing. Okta, another cloud native, spends (or “invest”?) 50-70% of revenue on sales and marketing. In contrast, traditional "capex" requirements for these companies can be relatively little.

So thinking about returns on capital for these companies is essentially thinking about sales and marketing ROI.

One Way to Measure

This is the unspoken premise behind Theta Equity’s post on Slack. A key component of that paper is measuring returns on sales and marketing with the following procedures (a very high level simplification here):

Sales and marketing ROI = Post Acquisition Value (PAV) divided by Customer Acquisition Cost (CAC)

• PAV is a function of 1) retention, 2) revenue dynamics, and 3) variable margin. First you work out the revenue curve for each acquired customer, then assume some 1) cost of service, R&D, and G&A to get a variable margins curve per customer. Then you discount that at the weighted average cost of capital (WACC) to arrive at PAV.
• For CAC, we can simply take sales/marketing expense and divide that by number of customers.

This is really great work. However, data about customer counts and cohorts are often not available, so here’s my even simpler way to not so much quantify, but think about sales/marketing effectiveness.

Simple Way to Just Think about Sales & Marketing ROI

I  basically compare sales and marketing expense as a percentage of revenue against the revenue growth rate achieved. Then I subjectively judge the recurring nature of that new revenue gained, and conjure up a ball park incremental margin.

I’ll go straight to a hypothetical example. If you spend 50% of this year’s revenue on sales & marketing, and revenue next year only grows 30%, is this a good “investment”?

Of course, the answer depends on lifetime value of that customer. In this example, you start with $100mm of revenue in year 0, and spending $50mm of that on sales and marketing gets you $130mm in revenue next year. This is $30 of incremental revenue for the $50mm sales and marketing “investment”. Now consider the two scenarios:

• A: Assume this new customer give you 3 years of revenue stream, which has 40% incremental EBIT margin (excluding sales and marketing cost).
• Then you would have $30 * 3 * 40%= $36mm of EBIT contribution (again ex sales/marketing cost)
• So you're spending $50mm to get $36mm back. Even without time value discounting, we can say this is no good.

• B: How if that customer life run 10 years, instead of 3 years?
• Then you’re getting $30 * 10 * 40% = $120mm. You spend $50mm to get $120mm (albeit over 10 years). This is much better! 

The income statements give us sales/marketing cost in year 0, as well as the incremental revenue in year 1. So what we have left to figure out is 1) average duration of customer lifetime, 2) margin contribution (ex sales/marketing costs).

There are ways to get ballpark estimates. If the company says customer attrition rate is 10%, you may say customer life time is roughly 1/10% = 10 years. To get margin contribution, you can use data from comparable companies that are more mature.

But still, it would be unwise to simply take empirical data and assume they stay the same. If in the past customer lifetime is 10 years, would you really be comfortable that the new product launched this year will also have 10 years life? The competitive landscape would surely have completely changed during a decade, particularly in the fast changing world of technology.

This is where quantitative data hits their limitations. Your analysis would now shift to focus on qualitative considerations like competitive landscape, market positioning, switching cost, network effects, and so on.

Footnote 1:
Traditional ROIC calculation implies a clean delineation of “capex” and “opex”. These are all just accounting identities though.

In the real world, what matters is cash in and cash out. You incur cash outflows for say 3 years, and expect to get cash inflow for say the next 7 years or even perpetuity. The cash outflow part we call “investment”, the cash inflows parts we call “returns on capital”. In the old ways “investment” is mostly some manufacturing facilities (what counts as capex), but in the broader sense this “investment” could also be a software intellectual properties, brands, customer relationships, or whatever that generates the cash inflows later (“returns" on capital).