Late stage technology companies severely underinvest in growth due to two reasons: (1) a P&L constraint: there is investor pressure to show a higher EBITDA even though that metric is not relevant for most technology businesses; and (2) a balance sheet constraint: they want to preserve cash near term, even though the growth spend has a good ROI long term. Both of these constraints severely hurt long-term enterprise value. We need a new framework to think about these businesses.
The origins of EBITDA
Why do companies use EBITDA as a metric, to begin with, anyway? EBITDA is a relatively new concept, created by John Malone in the 80s when running cable giant TCI. Cable businesses at the time were investing enormous amounts of money in CapEx to scale their infrastructure nationwide, which led TCI and others to show massive paper losses on their income statements. TCI’s long-term investment in cable showed up on the P&L as depreciation, severely reducing net earnings, even though this CapEx led to subscribers that would pay a handsome ROI on that investment long into the future.
This depreciation expense led to many investors misunderstanding the true earnings power of these businesses. Malone saw this mispricing, and created an alternative metric that captured the true earnings power by adding back that depreciation expense to net income: we all today know and accept EBITDA as the core metric that captures a business’ “ability to generate cash”. This key insight led Malone to become among the most revered investors and operators of our generation.
From EBITDA to EBITCAC
The irony of technology companies using EBITDA is that more than half of the letters in the acronym have zero effect for them: tech companies typically have almost no debt (so no interest), make ongoing losses (so no tax), and have no real assets (so no D&A). The point of EBITDA is not to blindly repeat the metric, but to strip back costs that show up in the income statement that do not reflect the core ability of the business to generate cash.
This is what using EBITCAC, not EBITDA, captures for any company that generates customer lifetime value. CAC represents the discretionary investment in Customer Acquisition Cost (often comprising sales headcount, marketing, and go-to-market initiatives) today for ROI in the future. This applies across different types of technology businesses: in SaaS or B2B companies, the predominant use of capital at scale is CAC, given the nature of the longer payback cycles. In B2C companies, the phenomenon of ‘negative EBITDA, positive EBITCAC’ is even more pronounced, since most B2C companies spend a much higher percentage of their total opex on CAC.
EBITCAC matters as a metric because the profitability yardstick companies use today – which is EBITDA – leads to poor decisions hurting the long-term value of the business.
When a company is managing to a target EBITDA, invariably the first place a company looks to cut is CAC/S&M. This is a poor and short-sighted decision. If the business is generating a good ROI on its CAC, then it should be the last place a company should cut. On the flip side, if a company does not have an appropriate ROI on CAC, it should minimize spend on CAC, regardless of how good its EBITCAC looks.
Breaking down technology companies
Most later-stage technology companies that have found product-market fit can be modeled as two ‘sub-companies’ combined into a single entity:
- The ‘CAC machine’: the part of the business that invests in sales and marketing and makes a clear ROI on that spend, in the form of new cohorts of customers that generate future gross profits over time, until the investment in CAC is recouped
- The ‘operating company’: the part of the business investing in the longer term product and operating platform that enables the company to keep fueling its CAC machine. This sub-business spends on product, engineering, finance, operations, etc. Its spend is funded by the residual customer lifetime value (LTV) left over from CAC, and by cash on the balance sheet.
When you remove the CAC machine, technology companies are highly cash generative in nature: gross profit from existing customers, minus ‘operating company’ expenses (i.e. not CAC), is often a positive number. This should not be surprising; the only reason tech companies look unprofitable is because they invest upfront in CAC today, for greater gross profit in the future.
As an example, let's look at the cost structure of a typical growth SaaS company:
- Revenue: 100%
- Gross Margin: 70%
- S&M (as a % of Revenue): 40%
- R&D (as a % of Revenue): 30%
- G&A (as a % of Revenue): 20%
- → EBITDA Margin: -20%
While not all of S&M is CAC – as it may include expenses on upsell, cross-sell, and the like – this assumption holds to a first-order approximation. If we add back CAC to EBITDA, the typical company above jumps to generating a 10-20% “EBITCAC” margin.
Maximizing enterprise value
The right profitability framework for growth businesses is to have a target EBITCAC, not a target EBITDA, and to keep spending on CAC, until marginal CAC = LTV or until a yield on CAC that makes sense for their business. This is how a company maximizes long term value. In practice today, we instead see companies back into a CAC payback target constrained by a target EBITDA or cash balance instead of letting the appropriate payback be an output of this value maximization function.
By constraining themselves in this way, we believe companies severely under-invest in growth over the long term. The two common arguments we hear from companies for cutting S&M are: a balance sheet constraint (“we don’t want to extend our paybacks as it burns cash today”) or a P&L constraint (“we don’t want our EBITDA to look too negative on our income statement”). The former is a real capital constraint argument that speaks to a sources-of-funds problem. The latter is just an accounting argument not based on first principles that makes little sense when examined – here the solution is to understand, report, and optimize for EBITCAC and not EBITDA.
The balance sheet constraint on spending to a marginal CAC = LTV or appropriate yield on CAC is a real challenge for companies in a capital-constrained environment. Even if it is right to spend more capital on CAC, companies want to preserve their balance sheet as much as possible.
Traditional CapEx vs CAC
In the traditional industrial world, to “grow”, you have to invest in more capacity (more machines, etc.), which requires more CapEx. But you wouldn’t expense that CapEx when you make that investment. You capitalize it by creating an asset on your balance sheet (the machine) and depreciate it over its lifetime; the hard asset on the balance sheet means you can get financing for it. The trade-off between the balance sheet constraint and growth does not exist in the same way or at least has an efficient frontier that is much further out, as CapEx is not typically financed with equity. In that context, CAC is the new CapEx and should be thought of in the same way.
We have convinced ourselves that technology businesses are “asset light”, but we did not emphasize that they are “expense heavy”. In other words, to grow we have to invest in CAC, which is expensed on the P&L. As a result, technology companies look unprofitable and asset light on paper, even though they are fundamentally very profitable and “asset rich” in the form of lifetime value. Imagine what the profitability or EBITDA of a traditional manufacturing business that is growing capacity fast would look like if it was expensing all of its CapEx instead of capitalizing it – dramatically worse.
CapEx is mostly financed using asset-backed financing, as the assets are underwritable and there is an unlimited supply of capital for assets that yield a decent ROI in a structured manner. It would be unthinkable for a manufacturer to rely solely and continuously on repeat equity raises to increase capacity in their manufacturing plant: the return on equity (RoE) on those businesses would take a massive hit. A version of this is what has played out in categories like gaming, DTC, fintech, and we are beginning to see this in SaaS too. These are often good businesses with a bad capital structure, which can lead to bad returns, due to the massive hit RoE has to take to continuously finance CAC.
Once a technology business has achieved scale and is generating a consistent ROI on CAC, their CAC machine effectively has the properties of an asset and is highly underwritable. Therefore, a company’s CAC machine should be financed using an entirely separate pool of capital: it is relatively de-risked and has a structured return profile, similar to asset financing.
Re-thinking the capital stack
Unfortunately, tech companies today, even large publicly-traded ones, still fund almost all of CAC using equity. Companies have tried to use corporate debt or variations on it as part of their capital stack, but they are not designed for growth, as they do not isolate the risk to CAC, and instead encumber the entire business with risk in the event of underperformance and have full recourse to other parts of the business. This causes companies to move to a conservative mindset of managing their loan vs a growth mindset of managing ROI on their CAC. As a result, companies just end up using their equity to fund CAC instead. It doesn’t mean corporate debt is bad; it just means that technology companies don’t get scaled leverage on the isolated part of their business that is relatively de-risked and, therefore, rarely manage to lower their cost of capital.
There are two bad outcomes as a result: (1) technology companies massively under-invest in growth compared to how much they should be investing; and (2) technology companies generate a lot less cash than they should be generating, as it is tied up in CAC. Both of these lead to a severe hit to long-term equity value that plays out years after the die is cast.
Just to illustrate point (1) further: it is almost universally accepted that a 24-36 month CAC payback is appropriate for SaaS companies. But why is that the right CAC payback target? If the company clearly has a long lifetime value, shouldn’t it have a much longer CAC payback? Why not let the marginal CAC = LTV equation determine the appropriate point to stop continuing to invest in CAC, versus prescriptively constraining investment to a target CAC payback? We’d argue that most companies are underinvesting in growth compared to what their real potential is based on beliefs that have no grounding in rational economic principles.
On (2): if the cash tied up in CAC was available to the company near term, founders would gain the ability to re-invest in building new products and business lines, potentially increasing long term value. A balance sheet constrained only by a company’s ability to continue generating ROI on CAC would also give founders the ability to invest even further in growth, and actually live up to their true enterprise value.
Using some of our own conservative estimates, the Cloud 100 companies on average spend $10M+ per month on S&M/CAC, or (in aggregate) $1B per month and $12B per year. Just 100 private companies could potentially generate that much more excess cash in just one year to be re-allocated in much more productive ways.
If a company doesn’t need this excess cash near term, it could start building its balance sheet to fund future M&A – or begin buying back shares from employees or shareholders who would love a path to liquidity over time. Just imagine the DPI that the venture community could create if this capital tied up in CAC were given back to shareholders. This would really ease the pressure to orchestrate events like being acquired or IPOs, which are more at the whims of market sentiment and random factors than most people like to admit.
This approach to capital planning is how we believe one should build an enduring business, but today, we find very few founders and CFOs operate with this capital allocation mindset.
Have founders become too content with benchmarking themselves to metrics that have little grounding in basic economics? Instead of obsessing over short-term profitability, it’s time to obsess over long-term equity value.
Can the new age John Malone please stand up?
