Fooled by SaaS Metrics: Gross Margin

Gross margin seems like a metric that is hardly prone to confusion. Gross margin is defined as revenue minus the cost of product and services (COGS) to deliver that service. A simple calculation, yet SaaS companies often end up making mistakes in defining and calculating this key metric that can have a wide-ranging impact on their business execution.

Software companies traditionally have enjoyed ultra-high gross margins. Originally, when software was shipped over CDs, COGS consisted of simply the cost of producing and shipping the CDs. SaaS, however, has completely changed the mechanics and delivery of software.

Software has gone from being deployed standalone within enterprise premises, to dedicated instances in the cloud, to multi-tenant architectures in the cloud where several customers share a single instance. Understanding and calculating COGS thus requires more nuance in SaaS.

What is COGS?

COGS should include all variable costs incurred in delivering the product or service. This typically does not include R&D costs.

Put simply, any cost that directly increases with more deployments and more revenue is part of COGS.

When your revenue doubles, COGS should also grow in proportion and similarly reduce when revenue declines. 

Let’s say a company provides cloud software to large and mid-sized enterprises. The enterprise deployment, in this case, also involves the use of professional services to help with configuring the software to cover specific workflows and also custom development to integrate with other software the customer already has, such as SAP, Slack, data visualisation tools etc. How should the company think of their gross margin?

The professional services cost of configuring the software is part of COGS. Typically, where the ACV is high, enterprises expect the product to be customised to their use case and this cost (and the associated revenue) should be included in GM calculations. Also third party licences for software that is embedded in the SaaS offering - like Twilio, Mongo DB or Zoom - should be included under COGS.

ALSO READ

Ask the Question: What do I have to lose?

What about the software development for things like integration plug-ins that are driven by initial customer deployments but are reusable and relevant to other customers down the road? This can justifiably be accounted for as R&D so long as it is something that will become part of the main SaaS offering.

If, however, the integrations end up being one-off and not applicable to other customers, they are variable costs and need to be part of COGS,

Startups don’t like to include professional services costs under COGS as frequently these are priced ‘at cost’ to land a deal. But doing that is simply inflating GM.

Another key component for SaaS companies is cloud costs. Storage and compute typically grow in direct proportion to new deployments. Companies with large data crunching, and complex machine learning/AI models will tend to have higher cloud costs. 

ML models need continuous refinement to increase accuracy and as a result data consumption and compute for data analytics do not reduce even at scale. As such these should be part of COGS. 

ALSO READ

The elusive third co-founder

The gross margin

So this is how a SaaS company GM would look like:

Gross margin = Revenue - (per-customer Cloud costs + professional services config and customer integrations + embedded licences)

Cloud costs can drive SaaS company GM to 70 percent or even 60 percent.

While moving from standalone software on CD to SaaS and cloud delivery has several advantages - higher GM is not one of them!

Low GM is not necessarily bad, it is just important to recognise what kind of GM is sustainable for the business in the long term.

As what looks like custom software work at initial deployments turns more into customer-driven product enhancement, GM can begin to increase at scale.

Also as a company grows it can offload custom work to SI partners which can further increase GM and ability to scale.

Does it matter that you measure GM correctly? It does.

A high GM means that the profitability of the business can come at an earlier stage, and also leave more room for experimenting with pricing. If you incorrectly assume a higher GM, you will find yourself requiring larger capital raises than planned in order to scale and ultimately break even.

The freemium conundrum

One area that leads to confusion is accounting for freemium customer acquisition. Freemium is where a part of the product is free for the customer but accessing certain premium features requires payment. Yet there is cost associated with servicing a customer who is on the ‘free’ product. Is this cost part of COGS?

A freemium acquisition strategy is designed to generate leads. It is a ‘loss leader’ to attract customers to try out the product and encourage them to upgrade to the paid version.

While more paid customers acquired this way will also mean a higher cost of servicing ‘free’ customers, it is best accounted for as a marketing cost. It is an acquisitions strategy and is more confusing if it is included as part of COGS.

However, pricing discounts are different and should be accounted for in GM calculations through reduced revenue.

In a nutshell, consider these aspects when calculating gross margins:

1. Include all variable costs associated with getting the product or service up and running at the customer including cloud costs, customer-specific professional services and embedded third-party licenses
2. Gross margins can increase with scale when what was initially accounted for as custom software development (eg. integration with other enterprise systems) become customer-driven product enhancements and part of the base product - when it is no more a part of COGS.
3. In financial planning, be conservative in projecting gross margin. Overestimating gross margin can lead to vastly incorrect projections on capital required to scale and break even.
4. Acquisition tactics such as freemium pricing is best accounted as marketing costs and not part of gross margin.

[This article first appeared on LinkedIn]