I was more than a little surprised to find that n8n Cloud did not follow the very wise trend of providing a Free tier in their pricing structure that supports the “let me get invested in this platform and grow my product to where I need more throughput and can sustain a paid subscription” user-converted-to-customer model.
Let’s elevate this surprise to “gobsmacked!” when the “Trial” signup process presented a requirement to submit Credit Card details. Who does this today, especially for a Beta product in a competitive marketplace?
A $5 indie hacker pricing Tier would be another good option for users who want to scale their Free trial to give them a little more throughput as they mature their product with limited user trails.
The Self Hosted option for n8n is nice but is not going to work for the no-code and low-code market.
There is a reason why no-code startups like Bubble and Figma grew rapidly and did well onboarding users to become paid subscribers. Asking users to cold commit to a 20 Euro up front monthly billing schedule is the fastest way to put friction on growth.
Make it easy for users to get committed and climb the paid Tiers as their product need dictates. Users “want” to use a product and “need” drives them to level up their commitment.
If n8n has over committed to venture capital interests that demand unrealistic financial growth to earn back their millions prematurely then what can be done? Why else would a fledgling startup in a competitive market ask new users to commit their credit card before a trial period?
Maybe n8n want to increase the friction to push users to Self Host? If so then ok, for me personally I have a Digital Ocean Droplet with LXD containers and can do that if nothing changes. DO’s pricing Tier is extensive and there is a price for every level of customer need.
Totally agree, we would love nothing more than being able to give everybody, who wants one, a free cloud account. The reason why we do not offer one has however mainly technical reasons (which has big financial implications). n8n is not a multi-tenant system. Meaning we can not start one server and run very many people on it at the same time, we rather have to start a server (in this case a virtual one) for each user and each of them needs a certain amount of RAM reserved. This sadly results in each cloud account not having costs of a few cents per month and the cost grows with usage, rather multiple Euros and that, no matter if it the instants gets actually used or not. So you can probably do the math and see how easy it would be to burn multiple millions very fast with such “free” accounts.
Not saying that would be totally impossible, many startups are doing that, it is just something we decided against as we want to grow sustainably instead. Esp, would we right now any time trade 1 customer who really loves and pays for n8n because he/she/they get real value out of the product with 100 that use us only because we are “free”.
The credit card requirement for the trial is there sadly no different. The reason for that has to do with the payment provider we use for our cloud offering. That one does sadly not allow to create a user without payment information.
Both above-mentioned technical reasons are also not impossible to overcome and as you can imagine we discussed them internally already multiple times. You can overcome almost every technical problem but any company (no matter if very large or as small as we are) has only a certain amount of resources and you have to think about how to use them best. Making n8n for example multi-tenant is no easy task and would take many months to complete, months in which not much other progress would be made. For now, we did, however, come to the conclusion that these resources are better used in other ways, for example generally improving the product.
Thanks Jan. Sorry to hear about the technicalities limiting scalable multi-tenancy and being locked in that payment provider.
I was curious about your multi tenant hurdle and wondered if the limitation comes from not being able to securely isolate lightweight Docker instances per VM?
Have you heard about Google’s attempt to make it easy to securely orchestrate sandboxed Docker containers to build multi-tenant clouds using a product called gVisor?
If gVisor seems too cutting edge perhaps Canonical’s LXD, which runs both full system containers and VMs, might do the trick:
LXD containers are secure by default, like a VM.
Canonical also has Juju to make it easier to manage Kubernetes and multi-tenant SaaS:
I found this interesting thread discussing features and plans and it added some background for me:
I discovered today that the Dgraph founder and team left the VC funded project about six months ago due to running out of funding and not being able to secure more. Dgraph is a GraphQL native graph database much loved by its users. The founder writes about the project and sustainable open source:
Thanks a lot @nick-walt that is all super helpful! As we had planned to update some things on cloud, all of those things come at a great time. Is very appreciated!
Also super interesting with Dgraph. Esp. that Manish did take a very similar route as we did with an own non-OSI license around sustainability.