David Heinemeier Hansson from 37signals recently announced on the podcast “This Week in Startups” the launch of a new project by his company. The project is called once.com and is the first product in over 20 years of the company’s history that does not rely on a SaaS business model. That project is significant for several reasons and could have far-reaching consequences for software vendors. To understand why, we need to revisit the early days of the SaaS model and outline some of its traditionally perceived advantages over the traditional on-premise model. Check our SaaS vs On-Premise Software comparison.
SaaS vs On-Premise Software: advantages of the SaaS distribution model over the on-premise model
- Compatibility and Security: In ancient times, before the mobile era, on-premise software was designed with a single platform in mind (e.g., desktop PCs). With the rise of MacBooks and smartphones, people wanted to use their software everywhere, especially when working remotely. SaaS is accessible everywhere and compatible with everything.
- IT team focus and operational complexity reduction: Engaging IT to build a CRM system from scratch for your company is, in most cases, a monumental mistake that distracts the team from more critical tasks. Because SaaS is hosted on the servers of the company selling it, users don’t have to worry about server space.
- Time to Value: On-premise software often requires installation and customization to meet a company’scompany’s needs. SaaS generally works—although sometimes a few clicks are needed to get the most value.
- Value grows over time: SaaS providers tend to add more features and integrations over time. On-premise software, aside from occasional patches, is like a stone carving—unchanging.
For these reasons, SaaS’s march towards increasing market share seemed unstoppable. However, cracks have appeared in this otherwise beautiful picture in recent years.
What are the problems with the SaaS model?
- Inflation has put pressure on software developers, and they have passed on their increased costs to customers through higher prices. As a result, the value-to-cost equation has become less favorable for clients than before.
- The subscription model in many companies has shifted to a user-based pricing model—pioneered by Salesforce 20 years ago—and later to usage-based pricing, pioneered by Amazon Web Services in 2008. In these models, the operational leverage from buying software with a fixed subscription has vanished. The simplest example is if we have 200 users and are looking for project management software, the difference between Asana (which charges per user) and Basecamp (which offers a classic subscription model) would be significant. In this scenario, Asana could be 22 times more expensive than Basecamp. Usage-based pricing can lead to high inefficiencies and is hard to budget. Entire companies have emerged specializing in optimizing bills for Amazon Web Services, Google Cloud, or Microsoft Azure, and the endless debate between CTOs and CFOs over cloud bill forecasting is nearing adulthood.
- In many SaaS products, the growth in value for users has essentially stopped. It’s no surprise—often, the software is good enough, and delivering new features and integrations in a way that most users will adopt is impossible. It’s an open secret in customer success teams that most users are not utilizing their potential, let alone what is yet to be developed for them. There are categories where there is never enough integration (e.g., e-commerce platforms or Zapier) and others where not much is happening (e.g., Slack, Gmail) because, really, why should it?
These cracks are causing some categories, which made perfect sense as B2B SaaS, to make more sense as B2B on-premise now. 37signals is starting with a Slack competitor via once.com. This is a strategic move that showcases their foresight and understanding of the market. It’s a brilliant choice for several reasons:
- A long time ago, 37signals had very efficient software in this category (Campfire), so they knew how to play this game.
- Slack has many huge clients. At the time of its acquisition by Salesforce, out of 100,000 Slack customers, over 1,200 were paying more than $100,000 annually, which accounted for 51% of the company’s revenue. I wouldn’t be surprised if, post-acquisition, thanks to boosted distribution through partners and salespeople, the number of Salesforce customers increased 2-3 times. If you employ 10,000 people and everyone is on Slack’s most minor plan, your bill is $600,000 annually. Companies in this situation would be more than happy to reduce this bill to $1,000 one-time and switch to 37signals, especially in today’s cost-cutting and efficiency-focused climate. This potential for significant cost savings is a hopeful prospect for many companies.
- This move will have a natural distribution advantage. Currently, many entrepreneur groups use the accessible version of Slack. Since the beginning, the community has asked administrators to switch to the paid version because the free version does not provide message history for conversations older than the first 10,000 messages. Administrators don’t want to buy the paid Slack version because if access to the community costs 49 PLN and Slack 5$, the fee for Slack will represent half of the administrator’s revenue. Therefore, these communities will quickly switch to 37signals’ product, allowing their users—decision-makers in their own companies—to adopt this product soon and want to implement it in their organizations, saving even more. This potential for increased efficiency is an optimistic outlook for the future.
Currently, 100% of my startup portfolio consists of B2B SaaS companies. I plan to continue investing in this category. However, this move by 37signals is very clever. I am surprised if part of the market concludes that switching to an on-premise model makes sense in the search for savings and operational leverage for medium and large companies. I would be terrified if I were the CEO of a simple application that last generated new value for clients 3 years ago. After all, why can’t software operate under a dynamic similar to that seen in the pharmaceutical industry, where cheaper generics appear after some time?
