mrb : home
After more than three years helping to build a business around a developer facing SaaS product, I've left to start my own consulting company called Computer Modern. We do sales and Marketing software and process consulting for developer-facing, business to business SaaS companies. Get in touch if you make a product for developers and need help!

How to Sell SaaS

When I decided to start trying to sell my services as someone with unique insight into marketing and selling software to developers, a thought occurred to me: I should probably back my experiences up with some research. I had read some papers relating to understanding customer behavior and revenue patterns, and delved a bit into the statistics and theory behind the KPI measurements that startups tend to rely on, but I hadn't done much searching for research about sales and marketing. I wondered what was out there.

After some searching, I came across a nice paper titled "How to Sell SaaS: A Model for Main Factors of Marketing and Selling Software-as-a-Service" by Pasi Tyrväinen and Joona Selin. In it, the authors develop and refine a model for understanding the important components of marketing and sales in a Software-as-a-Service (SaaS) business model. This paper interests me because helping companies develop go to market strategies is on the top of my list of projects I'm conquering with Computer Modern.

After some reflection, I realized that the reason this paper appealed to me is because it helps to do one of my favorite things in an elegant way: it breaks down a large, messy subject into constituent parts, backed by research to bolster its claims. Papers which provide overviews of existing disciplines are always interesting to me, and I look forward to reading more deeply in this area.

The authors are out to achieve a few basic goals in this paper:

"Our purpose is to initiate academic research on marketing and selling SaaS by presenting a model of the key factors for marketing and selling SaaS and related key performance indicators (KPI)."

Now that's what I'm talking about. It's great to see work aimed at taking situations where most people go on gut, and attempting to transform that into a more scientific, data-driven approach. All too often in my career I have made the mistake of assuming too much without bothering to check details, and I've seen other leadership do the same. This paper is definitely a step in the right direction.

In this post I'll pull out some of the more pertinent aspects of the paper to the work I'm doing with Computer Modern. Since I can't cover it all, I hope you'll read the paper in full and share your thoughts with me.

Definition of SaaS

One of the more useful components of academic work is its ability to provide definitions for complex ideas. Though the definition of SaaS as provided by the authors may not be complete or accurate for every case, it's a great starting point, and a useful one if you're explaining the concept to someone who is not familiar with it. The authors define the "criteria for compliance with the SaaS model" as:

It's interesting to see how this definition contrasts with one you might put together - this can definitely be an enlightening exercise. The result of a business which exhibits the above criteria is typically that a customer can easily adopt and buy the software, turning the standard enterprise sales model on its head.

It's worth noting that the cloud represents a major shift in how SaaS can be delivered - "on premise," single-tenant products (often known as "enterprise versions") are often sold alongside the multitenant ones. Offering dual product lines is so pervasive that companies like Replicated have sprung up to serve SaaS providers looking to provide single-tenant versions.

An initial model

The heart of this paper is the development of a model representing "The main factors of marketing and selling SaaS collected from the literature." The authors surveyed a number of papers and online sources in order to determine a set of variables that determine strategies for marketing and selling SaaS products.

Each dimension has a scale, and the model is meant to be understood this way: "the categories in the middle of the diagram are likely to co-occur and the categories in the outer rings are likely to co-occur." For example, under the "Service and implementation model," you would expect Self-Service products to co-occur with small customers paying a small amount of money (from the Customer Size and Entry Transaction Size categories), and Consultation based services to be sold for more money to larger customers.

It was rewarding for me to see this diagram because it reinforces something that I bring up a lot when discussing go to market strategies with founders, CMOs, etc. - no decision in the realm of the "business side of business" can be made in isolation. Customer Lifetime Value, initial deal size, customer persona, sales channels, how to market, customer size, etc. all influence each other, and how you approach one of these aspects will influence your ability to successfully approach others.

An evidence-based approach

Once the authors have described an initial model based on their synthesis of existing research, they document their multi-case study, where they survey a number of Finnish startups in order to see how the companies actually function with respect to their initial model. The resulting model is a very interesting set of tweaks on the first:

The basic refinements here are based on recognizing that of the original eight dimensions, there are groups which are so tightly coupled as to necessitate being collapsed into one. This has the benefit of turning something from being eight-dimensional to four-dimensional, which is considerably easier to reason about.

Additionally, these four categories are augmented with KPIs for short and long term success in each given area. This is one of the most useful aspects of the refined model for me - seeing KPIs paired with the dimension of the model they correlate with makes it easy to have conversations around the numbers, not just the ideas. Churn, Customer Lifetime Value, Customer Acquisition Cost, and Entry Transaction Size are all crucial numbers that are useful for measuring performance, but they're also incredibly important when discussing strategies and tactics for growth. Determining how the decisions you're making will impact these numbers isn't always possible to predict, but being in the mindset of keeping these numbers in your head is inarguably effective.

The revised "clover" model that the authors present is a fabulous resource. Starting from an idea and trying to turn it into a business, with a developed go to market strategy, is extremely difficult. Structured ways of thinking about how your idea will actually turn into a product in the market are crucial to measurable, sustainable, repeatable growth, and can make these exercises considerably more efficient and less prone to biased thinking.

The big picture

So how does all of this add up? What kinds of things should startups looking to develop a go-to-market strategy and grow be paying attention to? There are a few quotes in the conclusion of the paper which speak to these questions - I'll pull out a few them here.

"The dominant factor to determine the target group of customers was the number of potential users, which is directly related to the headcount of the customer organization."

Whenever I have the opportunity to discuss business strategies with technical founders, I always ask them this question first: "How many customers will you need to be successful?" This is a good way to put a real face on what is kind of an abstract question. The idea is that if you need $1MM in ARR to reach your first milestone, how many companies will that consist of? 10? 100? 1,000? Do you really think you'll want to support 10,000 customers? If so, how will you grow their revenue, etc.?

See? A good starting point. I love that question because of all of the other questions it forces you to answer.

"The main sales channel was direct personal sales supported with Internet-based marketing communication. Internet as such was not much used as a sales channel, and advertising was not used in marketing communication."

Developers who make products to sell to other developers often get trapped by attempting to apply to many tools too early. Online ads are an easy thing to pay for, but are they the right thing for your company right now? I'm not sure if the conclusions the authors make here regarding the relative dearth of online advertising applies to everyone's strategies, but it's definitely something worth being critical about. The next quote implies why this is the case.

"The most important performance indicator for marketing and sales was customer acquisition cost."

This is a big one. What I think startups should do is invest earlier on in making sure they're nailing their go to market strategies. What I DON'T think they should do is overspend on frivolous audience building activities or even hire someone full-time unless they're fully capable of understanding the domain of the product until they're clear about their goals with respect to the four "leaves" of the clover. Messing up and overspending to acquire customers who will cancel quickly and not provide obvious value is a recipe for the dead zone. Avoid it at all costs!

"Finding a marketing and sales approach ... will be critical to success of any SaaS firm in the near future."

This quote really spoke to me, because it's the whole reason I've been ranting about this stuff lately: how you package our product matters. Go to market is the thing that will separate your popular Open Source project from a sustainable business that can keep you going for a long time. Don't wait until its too late to get invested and excited about the positioning, packaging, and pricing of your product.

Wrapping up

Overall, the authors do an admirable job of collecting data and impressions from a wide variety of sources in an attempt to codify something very complex: how your SaaS business should approach Sales and Marketing. Whether you agree or disagree with their categorizations or choice KPIs, the "clover" is a great conversation starter, and I'll definitely be breaking it out when chatting with folks through Computer Modern.

Works Cited

Tyrväinen and Selin, "How to Sell SaaS: A Model for Main Factors of Marketing and Selling Software-as-a-Service" [PDF Link]