We bet our company on open source
I thought I’d share thoughts on why open source works for us. I’ve written this as much for myself as for anyone else, so I expect I’ll make some changes over time.
Context on our company and our market
For now, we make just one product. We make an API that developers use to offer enterprise single sign-on (SSO) support to their customers.
We developed our software as a fully open source project on an MIT license. We offer a hosted version of the software at no cost to our customers.
We’ve entered an established and competitive market, which means we need to positition our product and our messaging carefully to compete.
De-risking early user acquisition – social proof
Many conversations about commercial open source drift into a few common objections. How will you monetize? Aren’t you worried about your moat? What’s stopping someone from selling a fork of your thing?
I won’t suggest that those objections lack merit altogether. In some versions of our company’s future, any one of these concerns might command our attention. They’re just not important for our current stage.
The modal startup gains exactly zero traction. And since we’ve just launched our product, we currently have … nearly zero users. For now, we’re only interested in distancing ourselves from zero.
Startups can fail to escape zero – or even nearly zero – for many reasons. Often they just aren’t selling a compelling product. Sometimes, though, they just can’t overcome the inertia of zero. Remarkably few people will ever adopt products early. Not many of us want to try new things – especially in sensitive categories like ours.
If we had pursued a closed source model, we’d have found early adopters much more difficult to convince. We’d have needed to close a handful of early reference customers by personal reputation. Then we’d have written case studies or introduced those reference customers to new prospects. This stuff is costly.
Open source helps early customers take us seriously. After all, we’ve put all of the code right in front of them. Some potential users actually do fish around in our source code looking for mistakes. Others simply appreciate having the opportunity to do so. And everyone who visits our Github repo can see our stargazers, our forks, any documented issues.
Open source helps us gain initial traction. Since P(big outcome | nonzero traction) is much, much bigger than P(big outcome | zero traction), open source has in turn massively increased the implied value of our company.
De-risking early user acquisition – community
We have found open source companies responsive to our outreach and quite eager to try our product. Open source companies, it turns out, prefer to use open source products. We already have a competitive advantage within a manageable niche.
Counter-positioning
We have entered a difficult market at an interesting moment. As I mentioned above, developers will doubtlessly evaluate us against our competition.
Nonetheless, we understand that some of our competitors have soured their relationships with customers. This gives us opportunities to distinguish our offering in the extreme. We can be the most developer-friendly company in our market.
Where our competitors feel expensive, we offer a free alternative. Where our competitors pursue vendor lock-in, we opt for a maximally permissive, open source license.
Over time, our positioning will evolve, but open source helps set us apart.
Reinforcing our value proposition
Our product competes primarily on ease of use and speed of implementation. When prospective customers need to implement enterprise SSO support, we want our product to feel like the path of least resistance.
Open source just reinforces our ease of use for developers. With a free, MIT-licensed service, our customers sign no contracts, send no emails to their corporate lawyers, and spend no time worrying about vendor lock-in or business continuity risk. They don’t even need to use our hosted service.
To earn our place as developers’ default choice, we’ve slashed all barriers to adoption.
Sales and marketing efficiency – low entry ACV inbound
Most SaaS companies still haven’t figured out how to make money in the post-ZIRP era. Customers’ budgets just won’t grow. Churn has ballooned. Outbound yields miserable results. No one seems to hit quota. The conventional revenue playbook of the 2010s seems not to work very well anymore; we probably need to do something different.
Conventional SaaS sales and marketing organizations obsess over immediately quantifiable outputs like annual contract value (ACV). Faced with intense ACV-driven quotas, sellers will ruthlessly pass over lukewarm nurture prospects in favor of accounts that evidence high buying intent. The entire revenue engine focuses solely on who will buy right now. This isn’t always a bad model. It’s just really, really unproductive right now, and we have the opportunity to pursue something different.
We don’t entirely need new ideas to do something different, either. We can copy some ideas from MongoDB’s early sales and marketing strategies.
MongoDB made its Community Server product available back in 2009. From there, the company focused its efforts mostly on free users. Seven years and twenty million downloads into Community Server, MongoDB finally brought its Atlas product onto the market. The sales team didn’t need to go find cold prospects with buying intent. They already had tons of them … on the free tier. Sales were easy.
We’re trying to do something similar with our open source offering.
Could we charge money for the product? Absolutely. It just doesn’t serve our long term plans. We prefer to win widespread adoption among developers with relatively simple needs. In time, some subset of these developers will diverge from the others, evidencing more complex needs and larger budgets that we can monetize.
This model spares us the enduring inefficiencies of outbound sales and frees us to focus on developer experience.
Sales and marketing efficiency – early enterprise adoption
Software companies that focus on small businesses have a hard time. Don’t get me wrong; SMB-focused companies can succeed. But software companies typically make their money from big companies, which usually aren’t very price sensitive and tend not to churn. We’ll likely make the bulk of our money from big companies as well.
Open source companies like ours can win enterprise deals even at remarkably early stages. CIOs of big companies seem to like open source projects’ transparency and value the opportunity to self-host software when possible.
Moreover, we’ve already heard from some individual contributor engineers at big companies. We weren’t targeting big companies – realistically, we can’t target our marketing very narrowly anyway. We market to all developers the same way. We put our free, open source offering out into the community, and they find us.
Many of the developers who explore our product will pay us nothing, and that’s fine, because in the long run, a small number of developers will introduce us to very large customer accounts.
It’s more fun
I think we’ve both found the open source project much more interesting and fun than our previous work on closed source projects. I can’t entirely explain why. It might just be a better fit for our personalities. In any event, we basically enjoy doing this work and would prefer not to do things differently.