Untangling SaaS gross margin and OPEX
Like anything that has to do with startups, I’ve found that the best way to approach an idea is by asking “why does it matter” and revert back to first principles. What started as a narrow discussion on gross margin (GM) ended up a bit wider, but stick with me and we’ll get to the tactical stuff — how to build a GM business model and optimize your OPEX presentation.
Value, product-market fit (PMF) and capital efficiency (CE)
Foundational work has been done over the last few years on how to measure PMF and turn that into a science — Jonathan Hsu‘s work on this is one of the best out there. These efforts have led to the emergence of a new set of generally accepted methodologies to quantify PMF, which I believe is the main value driver for any startup.
If we accept PMF as a primary value driver and recognize that GAAP accounting is a lagging indicator of value creation, then why should startups bother investing resources to prepare GAAP financial statements at all? Isn’t it simply a cost of doing business?
Yes financial reporting *is* a cost of doing business, but that doesn’t mean that you can’t get value out of it. To see how this can happen, let’s express the value a startup creates as function of two main drivers — PMF and CE:
Value = PMF x CE
Where PMF measures the extent to which users engage and value your product and CE measure how profitably that product can be delivered.
While PMF will always overshadow CE in terms of value creation, at some point, typically when you add your first layer of VPs sub $3M ARR, measuring CE becomes increasingly important. The intuition here is that if you spend any resource on something (say sales or marketing), you want to make sure that this spend is as efficient as possible and/or that it builds a real option for future monetization opportunities.
Once you reach that point, GAAP actually starts to become useful since it can help you benchmark your profitability drivers, specifically when it comes to analyzing your GM business model and sales & marketing (S&M) efficiency.
In short, don’t take GAAP too seriously until you need to start preparing for an audit near your Series B. In fact, I’ve found that a helpful way to look at GAAP is not as a rigid framework where precision is the end goal — but rather as a process that will always be in flux and where the quality of *your* underlying assumptions is what matters. In summary, make sure you don’t mess with your taxes, but other than that the floor is yours to experiment and refine your own accounting definitions.
So why does GM matter?
In a nutshell, GM reflects the proportion of revenues you have left after subtracting the direct costs tied to servicing those revenues — *but before any S&M costs (aka CAC)*. Stated differently, in order for an additional customer to enjoy your application, you will incur costs that are directly related to said customer enjoying the application.
GM can also be seen as an indication of how much revenues you have left to invest in the three main operating expense(OPEX) categories, namely S&M (aka CAC), R&D (aka product & engineering) and G&A (aka management and admin). One last spin: GM is an important input to CAC Payback which means that it impacts unit economics and is thus positively related to valuation.
All that to say that GM is reflection of your application delivery model’s profitability. It is expressed as a % and defined as:
GM = (Total revenues(t) - Total cost of revenues(t))/Total revenues(t), where t can be a month, quarter or year.
The key thing to remember here is that any customer acquisition costs (CAC) should fall below GM.
To get to a GM business model, it is best practice to breakdown subscription revenues (ie recurring in nature, tech-enabled) from professional services (ie not recurring in nature, labour-enabled). These two buckets of revenues have wildly different GM profiles — subscription GM is closer to 80% while professional services GM is in the 0%-20% range. The charts below from the excellent 2016 Pacific Crest survey highlight the difference in GM across these two revenue components:
What should you include in cost of revenues?
Here’s a good place to start (note that not all items may apply to your business):
- Hosting and monitoring
- Licenses and royalties for 3rd party products used in your app
- Payment processing
- Channel partner royalties
- Customer onboarding, support and account management
- Professional services
You’re probably thinking, “wait a minute, what about our customer success reps who both support customers and also upsell them?”. Spoiler: this is where tracing the line in the sand between cost of revenues (aka direct costs) and S&M (aka CAC) requires a judgement call. Typical for startups is to allocate a chunk of customer success costs to cost of revenues and allocate the remainder to CAC. For eg. if your CS reps spend roughly half their time onboarding and supporting customer, then split costs 50/50 between cost of rev/S&M. All other costs incurred through the customer acquisition process (SDR, AE, online marketing, promo events, reps travel, name it…) should be accounted for below GM in S&M.
Wrapping it all together — GM and OPEX
Now let’s look at how we can present this information into a P&L format that makes intuitive sense:
A couple of observations from the above presentation:
First off, pooling OPEX into these three buckets helps connect elements of the value equation we discussed above:
The basic idea is that your R&D spend impacts PMF while GM and S&M spend will drive future profitability. G&A can be thought of as the infrastructure costs required to guide and balance those two activities.
Secondly, measuring these cost drivers as a % of revenues can help you benchmark your CE relative to other startups. The following spend ratios are good rule of thumbs when your ARR is in the $3M-$30M range:
- S&M 60–180% of revenues
- R&D 30–70%
- G&A 15–30%
Note that your S&M spend as a % of revenues should be analyzed in conjunction with its efficiency — ie the amount of new ARR it generates. You can check out my last blog for more colour on the Magic Number which measures new ARR vs. S&M spend. Note also that it is normal to see R&D representing roughly 1.5x G&A.
Best practice for managing these spend ratios can be summarized as follows:
- Always focus on finding PMF and measure it. Consider a growth team architecture to do that.
- Once PMF indicators are trending in the right direction run experiments on your S&M processes in order to develop a viewpoint on what the efficiency of that spend is and how it can evolve.
The interaction between PMF and CE can be simplified as followed:
Best PMF indicators
- Depth of user engagement
- Cohort analysis for logo retention and revenue LTV
- ARR growth
- Quick Ratio (or Net New ARR as a % of ARR)
- NPS
Best CE indicators
- Magic Number: if <0.7, scale down S&M as % of revenues — your spend is not efficient, if >0.7 consider ramping up your S&M as % of revenues.
- BVP efficiency score, measures OPEX efficiency, should be > 1.
- CAC Payback, measures how fast you recover your CAC. Best way to compute CAC Payback is to use your empirically observed revenue LTV retention profile and use that as a baseline for your CAC Payback model. Avoid 1/churn derivations, they are misguiding. Good model example here.
Finally, best practice for your *monthly* financial reporting is to include:
- Core set of PMF metrics that sit on top of your monthly P&L in order to keep track of value created relative to costs
- Then show P&L in the above format to see how costs are allocated as a % of revenues.
Going through this exercise dozen of times with early stage companies has lead us to the following template which we’ve deployed across our portfolio:
Like all frameworks, the objective here is to adapt this to the uniqueness of your business and iterate until you feel you get value out of it.
Don’t hesitate to reach out if you have questions or feedback as always.
Hugues
