Over the past decade, the prevailing narrative about the cloud was that it offers low costs, no maintenance, and limitless scalability. The pandemic only seemed to underscore the promise of leveraging the cloud in a suddenly remote world: Cloud adoption was expected to accelerate, but now it’s growing at an unforeseen pace. According to Gartner, 70% of organizations planned to increase their cloud spending this year in response to the impact of COVID-19.
At the same time, we’re beginning to realize the limits of “limitless” cloud computing. While the cloud does offer affordability, little maintenance, and flexible scalability early on in a company’s journey, the long-term costs eventually overwhelm the savings as a company grows, leading to lower profit margins. A16Z recently published a report on the long-term cost implications of cloud infrastructure that found that across the top 50 public software companies currently utilizing the cloud, the costs of cloud (versus running the infrastructure themselves) lead to an estimated $100B of lost market value.
A16Z’s findings emphasize a long-held belief at Elsewhere Partners when it comes to building software companies: There’s no one right path for every software startup. It’s why we don’t subscribe to a billion-or-bust, growth-at-all-costs mentality and why we don’t solely invest in cloud-based software companies. With a pragmatic approach to growth, we see potential in on-premise and hybrid software business models that many traditional investors ignore.
What does this mean for founders? As a CTO, I’m here to tell you that you don’t necessarily need to launch your software startup in the cloud. Let’s explore the benefits and limits of the cloud, and outline what you should consider when it comes to the right delivery model for your initial product.
The vast majority of early-stage founders that I work with are building software-as-a-service (SaaS) applications and, by default, putting them in the cloud. For the most part, this makes sense; I do believe the cloud is often the best option. But, I challenge the assumption that it’s your only option.
To quote one of my favorite product leaders and Elsewhere Operating Advisors Denny LeCompte, “You can have it fast, good, or cheap. Pick two.” Leveraging the public cloud is all about “fast” and “good,” because as you scale, there’s nothing “cheap” about it.
When you’re early in your startup journey, it’s true the public cloud can save you money and time in several ways. These include:
Though the cost and time savings these three elements can offer you shouldn’t be overlooked, consider these through the lense of scalability. If you’re scaling quickly in the cloud, you will need to spend more and more on computing expenses over time. Think of it this way: It’s always cheaper to buy than rent. As Sarah Wang and Martin Casado wrote in A16Z, “If you’re operating at scale, the cost of public cloud can at least double your infrastructure bill.”
So, how do you know if launching your software startup in the cloud is the right strategy for you?
For the most part, the type of software product you sell will tell you whether you should be in the cloud or not. While many people often use the terms “SaaS” and “cloud” interchangeably, the truth is that not all SaaS products need to utilize the public cloud. There are three main ways to offer a software solution:
Many enterprises prefer to host critical infrastructure software, such as software used for network automation and cyber security, on-premise or in their own private data centers which operate like a private cloud. This allows companies like Itential, which provides network automation, and AppNovi, which offers cybersecurity attack surface discovery and visualization, to avoid public cloud expenses during the early days while their focus is on large organizations as customers. As companies like these scale and expand their target addressable market (TAM) by expanding down market, the addition of public cloud-hosted SaaS offerings often emerge. By then, revenue growth is well underway and capital is available via venture funding.
If you go this route, you can still deliver “software-as-a-service.” By bundling the licensing fee and ongoing maintenance costs into a subscription model, you can sell software “as a service” without the cloud. This provides the consistent revenue benefit of cloud-based SaaS with the core product offerings of traditional software. Plus, if you’re installing the software on-premise for the client—rather than paying for your own private data centers—you’re also saving on costs.
Building and offering your software product from the public cloud as a SaaS offering is by far the most common method of building and launching a software product today. This method can be lightning fast if you have an experienced product team, an agile approach, and you can get an MVP into the market quickly.
While the long-term infrastructure costs are high, one of the key advantages of offering your product as Software as a Service (SaaS) is that you can collect and analyze data on user behavior. This is especially important for companies leveraging a product-led growth strategy. Building these new product offerings in the public cloud is fast, secure, and easy relative to building out and maintaining your own cloud infrastructure.
Product improvements gained by studying user behavior typically benefit both types of customers: Those leveraging private data center deployments and those consuming your product via SaaS.
Contrary to popular belief, there are more than three cloud providers in the world. Amazon (AWS), Google (GCP), and Microsoft (Azure) are certainly the 800 lb. gorillas of the public cloud market, but there are hundreds of other companies providing these services. Some of these are traditional technology companies, like IBM and Oracle, as well as service providers like AT&T and Orange. Additionally, many of the large collocation and data center providers now offer cloud-like services. Oftentimes these companies offer significant advantages over first-tier public cloud providers including lower prices, friendlier data export policies, and better portability.
The question of kicking off your startup in the cloud really goes back to your specific product and growth goals. As your startup scales, your decision to utilize the public cloud or not will likely shift. In many cases, we’re witnessing software companies eventually land on a hybrid model, whether that means launching a complementary cloud offering or repatriating workloads from the public cloud to private data centers (of course, the latter is trickier to do than the former).
Ultimately, trust this: The public cloud isn’t your only option when it comes to starting a software business, despite what many traditional venture capitalists will tell you. You can still build a company with great potential on-premise, and if you’re looking for an investor that’s willing to hear your pitch, our team is all ears.
Check out the first post in our cloud series, “Why We Look for Software Investments Beyond the Cloud” and look for our next post on how to think about cloud usage as you scale your business.