SaaS vs. Open Core: Which Business Model Is Best?(gravitational.com)
gravitational.com
SaaS vs. Open Core: Which Business Model Is Best?
https://gravitational.com/blog/open-core-vs-saas-business-model/
11 comments
This is great. How did you collect this?
I assume we need to tack on an extra 3 0s to each of the figures there?
I agree, these numbers seem low.
It’s a convention. SEC filings typically list financial data in 000’s.
We buy support.
I dislike Open Core as it tends to be unclear where features go in the future, and the pricing is usually insane.
At my workplace we need to be on-prem so SaaS is out.
We often pay for support we don't really need, in order to fund open source projects.
I think in the article SaaS is used to refer to proprietary SaaS companies.
In my opinion these are two different dimensions. You have the delivery method: SaaS vs. self-managed/on-premises. And you have the license of the code: proprietary vs. open-core (which can be fully open source, open-core, or a non-compete license like the one MongoDB and CockroachDB switched to).
At GitLab we have an open-core license and offer the product both as SaaS and self-managed, so we do both things in the title. The opposite would be companies offering only proprietary software that you run yourself, like Sketch.
Another dimension would be billing, people frequently mention SaaS as an example of subscription billing. At Gitlab we also bill for our self-managed software on a subscription basis instead of a big license fee upfront and yearly maintenance.
In my opinion these are two different dimensions. You have the delivery method: SaaS vs. self-managed/on-premises. And you have the license of the code: proprietary vs. open-core (which can be fully open source, open-core, or a non-compete license like the one MongoDB and CockroachDB switched to).
At GitLab we have an open-core license and offer the product both as SaaS and self-managed, so we do both things in the title. The opposite would be companies offering only proprietary software that you run yourself, like Sketch.
Another dimension would be billing, people frequently mention SaaS as an example of subscription billing. At Gitlab we also bill for our self-managed software on a subscription basis instead of a big license fee upfront and yearly maintenance.
What was the driving factors behind not doing upfront license fees with long term renewals for your on-prem?
"I think in the article SaaS is used to refer to proprietary SaaS companies."
That's true and covered in a previous post [0]. I also recently saw this tweet from @asynchio that also agrees there are different dimensions [1] and will be covering that further.
[0] https://gravitational.com/blog/kubernetes_networking_on_bare...
[1] https://twitter.com/asynchio/status/1136139175437062146
That's true and covered in a previous post [0]. I also recently saw this tweet from @asynchio that also agrees there are different dimensions [1] and will be covering that further.
[0] https://gravitational.com/blog/kubernetes_networking_on_bare...
[1] https://twitter.com/asynchio/status/1136139175437062146
Cool, thanks!
> In the second post, we discussed how there is actually quite a bit of overlap between SaaS and Open Core and to draw a distinction, we would assume the Open Core software is primarily being delivered on-premises for comparison purposes.
Hmm, that might be convenient for your analysis, but it seems an unsafe assumption and artificial categorization to me. I think an increasing number of "open core" efforts, possibly most of them, are trying to monetize the non-open-source portions via SaaS. In fact, as we saw in the Amazon/ElasticSearch fight, one of the main motivations for not using a FLOSS license for some or all of your software is to prevent SaaS competitors from using the software you are producing.
I tried to look at your list of companies analyzed categorized "Open Core" to see how many of them offer a SaaS, but I realized that for these companies I was unfamiliar with it was hard to figure that out quickly from their web pages. Let alone then figure out if they were primarily delivering as SaaS if they offered an on-premises option (as even Github does). But that would seem to be neccessary to validate your assumptions.
Hmm, that might be convenient for your analysis, but it seems an unsafe assumption and artificial categorization to me. I think an increasing number of "open core" efforts, possibly most of them, are trying to monetize the non-open-source portions via SaaS. In fact, as we saw in the Amazon/ElasticSearch fight, one of the main motivations for not using a FLOSS license for some or all of your software is to prevent SaaS competitors from using the software you are producing.
I tried to look at your list of companies analyzed categorized "Open Core" to see how many of them offer a SaaS, but I realized that for these companies I was unfamiliar with it was hard to figure that out quickly from their web pages. Let alone then figure out if they were primarily delivering as SaaS if they offered an on-premises option (as even Github does). But that would seem to be neccessary to validate your assumptions.
I've been working on an open source search engine (https://github.com/typesense/typesense) and this exact deliberation has been on my mind.
Ideally, one could offer both on-premise and hosted versions (like Gitlab), but that's really tricky and needs a lot of resources to pull off successfully.
Ideally, one could offer both on-premise and hosted versions (like Gitlab), but that's really tricky and needs a lot of resources to pull off successfully.
This question (or the basis of the analysis) really doesn't make a ton of sense. The overriding deciding factor is really: who is the buyer and how do they want to buy. An open-core model for a strictly non-technical audience is a lot worse off than SaaS. On the other hand, if you have customers with security concerns SaaS may be off the table (still) for many of them.
I'm really curious how bootstrapping as opposed to VC money would affect this comparison. Not really much in the way of data I imagine though... I suppose it's hard to compare too. Big piece of a small pie vs small piece of a big pie kind of issues abound. Although HN has a very reasonable slant towards VC backed startups, I'm much more interested in bootstrapped startups.
I have some experience with both models. Bootstrapping is hard mostly because of sales, not R&D. I'd say a healthy model these days is to bootstrap the tech before you start investing in sales as much as you can. The danger with that is poor product market fit: most techies are not great at doing sales and marketing and you may end up building what you'd want instead of what your customers actually want.
For a no-tech company, getting techies to execute your vision is hard and expensive and generally impossible without funding or a very solid co-founder. You are basically talking about outsourcing an mvp to cheap consultants that aren't that great. The result is throwaway code. If you can sell that; great and you can then start fixing it. If not, that's a problem.
A few of the open core companies pre-date their VC rounds. I know one of the Elastic co-founders from before he got involved with that. The project was several years old by the time they incorporated and got proper funding. They were bootstrapping using consulting and support type stuff (this is how I met him). After getting organized, they went from a small group of contributors to closing several big rounds and hundreds of staff in a short time. Also from there to IPO was quite fast. I'd say this was a good example of bootstrapping followed by VC capital to grow followed by what looks like a successful IPO: they are trading at 82$ after an initial price of 70$, they have revenue, and growth, and what looks like a reasonable valuation to my (not an expert) eyes.
IMHO there's nothing wrong with VC capital provided you do it for the right reasons and on your own terms and not as a shotgun wedding. In my current company, we got close to a series A a couple of times but chose not to proceed for the reason of wanting to stay in control of our company and protecting the interests of our early investors. The flip side is that this has slowed down our pace and limited our ability to grow.
For a no-tech company, getting techies to execute your vision is hard and expensive and generally impossible without funding or a very solid co-founder. You are basically talking about outsourcing an mvp to cheap consultants that aren't that great. The result is throwaway code. If you can sell that; great and you can then start fixing it. If not, that's a problem.
A few of the open core companies pre-date their VC rounds. I know one of the Elastic co-founders from before he got involved with that. The project was several years old by the time they incorporated and got proper funding. They were bootstrapping using consulting and support type stuff (this is how I met him). After getting organized, they went from a small group of contributors to closing several big rounds and hundreds of staff in a short time. Also from there to IPO was quite fast. I'd say this was a good example of bootstrapping followed by VC capital to grow followed by what looks like a successful IPO: they are trading at 82$ after an initial price of 70$, they have revenue, and growth, and what looks like a reasonable valuation to my (not an expert) eyes.
IMHO there's nothing wrong with VC capital provided you do it for the right reasons and on your own terms and not as a shotgun wedding. In my current company, we got close to a series A a couple of times but chose not to proceed for the reason of wanting to stay in control of our company and protecting the interests of our early investors. The flip side is that this has slowed down our pace and limited our ability to grow.
Thank you for that! I seem to remember Gitlab also saying that they tried bootstrapping before accepting VC funds and had a hard time at it.
Isn’t there survivorship bias by looking at notable successful IPOs?
Wrong question as usually both are combined
Edit: as already in other posts highlighted they are two different dimensions
Edit: as already in other posts highlighted they are two different dimensions
off topic but Neo4j recently executed an open source for open core bait and switch. thoughts on that?
It's a Jeopardy! clue.
[buzzer]
What is .. Software as a Service?
[buzzer]
What is .. Software as a Service?
I had also looked at companies like Slack and Zoom which target lower average deal sizes and it was quite remarkable how much better their metrics were. But didn't feel like it was apples to apples. That data is the spreadsheet though.
[0] https://docs.google.com/spreadsheets/d/1BV1Kk542KHultHOB8Lkz...