Feedback: self-hosted pricing

Honestly have never seen pricing that everybody liked.

In the end, you can ask ten different people and receive ten different answers. The only thing they usually all have in common is that they will propose something that makes it cheap for their use cases (and I am not different in that regard, it is human nature). In our case will users with a lot of executions argue for charging by workflows. Users with many workflows but few executions will argue for executions, and so on.

So if we would change it today, the only thing that would happen is that tomorrow somebody else would be unhappy. Unless we find a price that ensures that everybody pays very little, for unlimited everything. Obviously, like that, we will not be able to build a sustainable business.

Honestly not sure what you mean with “Based on this conversation it feels like the community edition will have this ability stripped in the future for the purposes of profit”. That is not what I am saying or have ever said. The community version is and will stay unlimited regarding the amount of workflows and executions. And we are incredibly far away from anything which can be called “profit”. The goal, for now, is to become self sustaining, which should be in the interest of everybody.

And you made a great point with “Why use Teams at all when community allows us to run unlimited workflows? There goes your business model”.

  1. That is exactly one of the problems we are facing. We give incredibly much value away for free, but we have to charge for something. And is again the same as mentioned above what we charge by. Here people will always argue that the features they need are free or cheap, and the ones they do not require, to be expensive. But we very purposefully choose features people do not require to use n8n successfully.
  2. It shows that n8n is still a viable product for you and others, even without it. You get thousands of dollars worth of value even without paying anything. But there are also companies that can not simply operate without the sharing functionality. Especially the larger ones and that ones we want to monetize.

Regarding being able to host unlimited workflows because that’s the purpose of self hosting. That can be one advantage and apparently the main one for you. But for most of our users, especially the larger ones, is the main advantage, for example, that it runs on their infrastructure which is important for security and data privacy reasons.
That also brings me again to what I linked above. n8n was never supposed to be the free/cheap solution. It is rather the most powerful solution out there. That is maybe also where this disconnect comes from. Anybody should be able to run n8n for free if they are on a budget. But it is not given that n8n (the pro features) will be cheaper for power users or big companies. It is possibly even more expensive, and that is fine for them. Again because they do not use n8n over Zapier/Make/Workate … because it is cheaper, they use n8n because what they want to do can not be done in others.

Regarding the $500/month you mentioned. Why should you, for example, not pay that? In my eyes, is that not a crazy outrageous price. Especially if I think about what we are paying for some apps. Many one-person startups or stores seem to pay that and more for Zapier every month. Even if each workflow just saves you 15 min a week, do we talk about 100h/month.

For me, it feels like, people have a tough time paying reasonable amounts for open-source/fair-code projects. They are OK paying a lot for proprietary ones because they are used to paying for them. But if they start with free, most of them are okay with paying something, but that something is often way lower than what the proprietary solution would cost.
A possible reason is that for the proprietary product, they consider the value of the whole product, and price X seems fine.
For open-source/fair-code, they get so much value for free, which they take for granted. And then see that the paid plan adds only feature Y, and that seems crazy. How can all of that other stuff be free and that one small, simple feature so expensive? What they forget is, that even though companies like ours may charge just for that one feature, they still have to build and maintain the rest. So they do NOT just ask for money for that one feature, they ask for money to build the whole product, maintain it, create documentation, give support, and so on. Just because they give the rest away for free, does not mean it is cheaper or free for the company. The only difference is that they only charge for that one feature and give away the rest for free.

So the assumption that using an open-source/fair-code project has to be always cheaper than their proprietary alternatives is a misconception that has to change! 100% of the same work goes into it, and a lot of that work goes into supporting users that will probably never pay, and that is fine. But that should not be a reason for trying to shortchange open-source-fair-code projects.

And to make it clear, I do not speak at all about any particular person. That is a problem I generally see.

9 Likes

Why should you, for example not pay that?

I think you are either misunderstanding me again or taking what I said out of context.

The issue is not paying $500 per month - it is paying $500 per month and being limited as to how many workflows we are allowed to run on our own compute (which we also pay for).

Even Zapier allows you unlimited workflows for far less than that - although to be fair, they are gouging people based on the number of tasks.

To me, both of these practices are egregrious - it puts you in the same camp as your competitors where now the product becomes severely limited based on an arbitrary number and quickly scales into the thousands of dollars per month if you are even a somewhat-technical user.

Thanks a lot for clarifying!

We have charge and limit by something. We had the choice and thought a lot about what that should be. The obvious candidates were executions, tasks/steps, workflows, or users. Why we went with active workflows, in the end, gets also explained here, and I still think it is the best and fairest choice. If we had optimized for revenue, the choice would surely have been executions/tasks instead. Allowing an unlimited amount of workflows and then limiting by executions was for sure also discussed but I am sure, that would have not just been much more expensive for most, also has other disadvanages as explained in the original post. Also very certain people (including you as it is even more limiting) would have liked that even less.

Regarding paying $500 and still being limited. There is, for sure, a price for which we can allow an unlimited amount of workflows. But what should it be? $10k, $1k, $500, $100, or $10? Generally still think it is fair to pay more the more people use it, and the more value gets generated. It would also sound wrong if you pay X for unlimited active workflows and have 100, and a big Enterprise company has 5000 and pays the same.

And again, I am aware that it is not perfect, but it is not that bad either. And considering the number of users we have and the number of views the above post got where we communicated it (as of today 1.1k) there were few complaints about it, as you can see. For me, that is actually a great sign and shows we did not get it totally wrong.

Regarding putting in the same cap as our competitors. That is very wrong; they do not even live on the same planet. Most of them charge not by active workflows, or even executions; they charge by tasks! Many of our users have workflows where a single execution would literally eat up 10k and more tasks. If such a workflow executes once an hour, we talk about 7.2m tasks a month. Even going now with the cheaper alternative Make, that would be $8.2k/month, and is literally only active workflow in n8n, so $5 a month (to be 100% correct would be $25 as that is our smallest plan starts with 5). It is an extreme example, but even if you go with 100 tasks every hour, it is 72k tasks/month and would bring you up to $55/month, so 11x.

And you are right, it runs on your own hardware. But that is honestly not where the magic happens. So does not make a big difference in my eyes.
On top, is that also not something that makes it cheaper in most cases, it is even something people are willing to pay a lot of money for (it is a feature, not a bug). No idea what our competitors charge for their self-hosted option (and how many have one), but I am pretty sure they do not even take a call for $500/month.

Regarding the arbitrary number and the fee quickly scaling into the thousands of dollars per month. Literally, all pricing for digital products is in big parts arbitrary. For most SaaS solutions, if you access it with 1 or with 100 users does not make a huge difference in cost but it gets limited by it. That is how it works, and to get into the thousands (the first number that would qualify there is $2k) you would need 400 active workflows. If, for that huge amount of workflows $2k is not a fair price, then I would say there are other things wrong. We at n8n are 35 people right now, and not even we have probably reached that number yet (and we obviously automate a lot lot with n8n).

1 Like

Please consider the case of using a closed network environment where it is difficult to pay for license authentication.

@sackoh as far as I know did we consider that and a license would also work in that case. We probably would just have to change a few settings to make sure the license is valid a longer time as the automatic refresh would not work.

2 Likes

@jan i hope you consider changing the number of active user on n8n.cloud paid version, im n8n affiliate, im not gonna lie, i run selfhosted my own n8n too for my own development or making future example that i can offer to my client.
I hope you can change your mind about adding at least 1 user to the cheapest n8n.cloud plan, why?
Mostly my client dont know how to code, dont know how to selfhosted, but i always suggest it’s better go with the managed n8n.cloud since this will be support n8n to grow too, but the problem rise when my client want to create some new workflow because they need to share their account for me to get access to their current workflow, and that’s really annoy me because even the client trust me 100% as the man who really care privacy i really annoyed by this, at least kindly please at least add 1 more user.
So my client dont need always ask me for what they need to do if there is critical update or something they need to from n8n like example:
There is some bug that need user to upgrade their n8n.cloud instant to another version, please something like this is easy for me but for my client their so panic amd dont know what to do.
Please consider adding 1 more user at least as managing dashboard only, so this user can manage the dashboard and the instance it self, no need to have permission to create workflows or accessing the workflow, but at least for managing the n8n.cloud instance and can reach support because this additional user included in the plan.
Please tell me how i can keep supporting and offer my client n8n.cloud if my client doesn’t know anything about automation, but want their business grow amd they want to start with the smallest one aka the cheapest one but they force to share their credential with me for something like updating their instance or check how many workflow left.
My request here just please considering adding 1 more user for accessing dashboard, doesn’t need to ne accessing workflow at least for managing their instance in n8n cloud.

Yoy may limit this user for not accessing workflow, again my purposes is to help client if something need to be done at portal level, you may closed workflows etc to this new additional user, but please give this new user access for managing instance like updating, restarting, and have the same level for reach the support.

I like offering n8n cloud but this kind of issue is really limit me to offer it to small business or someone who want getting started.

Thank you. And i hope you consider this idea

Hi @Cipherdale

I like the idea of a service user of some kind to allow consultants to help their clients better.
But I think it would be better to open a feature request for this. As it has nothing to do with the topic discussed in this thread. :slight_smile:

1 Like

Thank you very much @Cipherdale for your input!

Can for sure totally understand where you are coming from, and it is more than appreciated that you want to help us to generate revenues. Thanks a lot for that. But honestly, even if we want to enable that functionality, it is sadly nothing simple as it does not exist yet. It would require a lot of work to add multi-user support to our cloud dashboard.
As we have just limited resources and it is nothing that we did hear much about in the past, is it currently very unlikely that we would prioritize it any time soon.

So what @BramKn requested is actually the best way to move forward. So please open a feature request for that, and depending on the interest (so upvotes) we can then prioritize it accordingly.

I hope that makes sense. We are a small team after all, and so have to prioritize features that have the biggest impact.

1 Like

Yeap, i will open in request feature, sorry for my bad English.
Yeap it’s okay for me if this user work as service user aka cannot access the workflow, so only for managing instance and reach support, because the primary client just dont want to play with it but want everything work behind the layer, so this new service user is really needed.
Thank you for responding :beers:

3 Likes

I would love to see n8n grow to a product that can really compete with Zapier. In fact, I wouldn’t mind contributing with a small monthly “subscription” like $5/month, and I think many other developers would be willing to do the same, as long as that would give them access to all the exciting new features of the self-hosting option, leaving the door open to contribute more if they wish to do so. I don’t think putting exciting new features like “environments” in the more expensive tiers is a good idea, but I would obviously leave actual enterprise features like LDAP integration or team collaboration out of the “fans tier”.
In summary, let the people cheering for your success support your development with a minimum small monthly subscription (or larger if they wish), while sharing with them all the intriguing new features.

Thanks a lot for taking the time to share this.

Is always great to hear that people want to support us, it is very appreciated. Thanks a lot for that. So I want to provide some background. We currently give away a lot for free, meaning there is already not a lot of reason for most users to pay. Every additional feature we add to the community version makes it less likely that users pay in the future. And each feature we would make available for very cheap makes it less likely that people will start to pay for the higher plans in the future. So we have to create some differentiation.

There are for sure more people out there in the world that would pay us $5/month vs. for example companies that pay us $2.000/month. But obviously, does one of the 2.000 ones make up for 400 of the $5/month. From what we have learned in the last years and some basic math, did we figure out that we can we build a sustainable company by focusing on these large users but not the other way around. Because even though there are more people out there that are willing to pay 5 than 2k, as mentioned before, is multiple, not 400, rather incredibly much smaller.
Or, to say it differently: To produce, let’s pick a round number, $1M in annual revenue, is it easier to find ~42 companies that pay 2k/month than 16.666 users that pay $5/month and it keeps on getting harder as we scale.
Just think about how many people are in the world that are able to self-host n8n + have the need + would pay vs. how many companies are out there where the same is true.

It is not that we would not like to give everybody access to those features, but we also have to think about building a sustainable business. That is not just in our interest, but also in the interest of the whole community.

I hope that makes sense and makes it clearer.

2 Likes

Thanks Jan for your response and your great work!

I don’t believe a business large enough to spend $2k/month would go for the $5/month tier, not just because of the missing enterprise features, but because they need someone to yell at when they are losing money because something does not work.
My calculations are that having 20k+ (less than 1% of current Zapier’s user base) $5/month enthusiastic supporters would definitely help you attract more of those high-valued $20k enterprise level customers away from Zapier, not the other way around.

You are welcome and thanks!

About the rest. All I can say is, that you can probably ask any open-source (in our case fair-code) company out there, but there are a lot of companies out there that could spend $2m+ and they are totally fine using the free version. Ask the most databases for example.
Is also understandable and nothing you can blame anybody for, why pay more or go through the hassle to pay (which is not fun in a large organization) when you do not have to.

For what it’s worth, we went from being raving fans of this product (I have contributed time to supporting others in this forum, posting solutions etc) and being willing to pay upwards of $125 a month for a team of 2 to moving our most critical infrastructure into custom Python containers running in our cluster.

The idea of spending $200+ USD / month on hosting and then having to pay more and more for the “privilege” of running workflows on our own infra is - to put it bluntly, a punch in the gut.

The cost is simply not worth it when you combine the cost of hosting on our own infrastructure with having to pay for workflows. Add in the founder’s insulting and dismissive “we know best” replies to myself and several others in this thread, and I think we can make a pretty good guess as to why the pricing structure is set up in such a way as to make running on your own infrastructure “not worth the cost” vs the cloud plans.

Now, it is very generous of N8N to provide a Community Edition, but for how long? Until the company can not find enough enterprises who want to switch from Zapier at the “$2000+ a month” price range and decides it is too expensive to offer a free product?

Now with all of the above risk (and what constitutes “risk” in the business world is subjective), add in the fact that Zapier also offers enterprise packages and has been doing this for a lot longer - not to mention, much faster updating + expansion of their library (it has been more than 2 months and the BitWarden node for N8N is still not updated, to which the response has been “you can contribute to the project yourself if you want it badly enough”). I am not advocating for Zapier btw, we switched away from them because this project seemed to hold a lot of promise. But to suggest that N8N is a for-profit company, while still continuing to act like an open-source project “build it yourself”, seems hypocritcal at best.

1 Like

Very sorry it was not meant as a simple “we know best” response at all. I tried to explain the reasons behind the decision but at the same time is that a discussion that would probably never find an end. Trying to finish every debate and talk about every single point is not possible, no matter how much I would like it to be. Next to a company to run I also have a family. So sorry if you think my answers are not sufficient.

I always try to be very transparent and honest, but you also have to understand that we have more data and more information as we are running the company. So we see the usage, how many people pay, how much they are willing to pay, and which people pay. We have a lot of conversations with users, other projects, and also investors who did support successful open-source companies in the past. I am not saying we know best and that we make the perfect decisions but we also do not make these decisions lightly and a lot of thought, work, and research goes into them.

About contributions. Right now, probably more than 99% of the code is written by us. So it maybe not a bad suggestion by you to say we do not accept any contributions moving forward. Not because we do not value them, just to avoid that problem moving forward. We will think about it. Thanks a lot for that input.

1 Like

Do you mean why pay $5/mo when you can get it for free and/or why pay $2k/mo when you can get it for free?

I first encountered n8n due to constraint on project budget. My research led from the big paid ETL players such as Zapier, panoply, fivetran, talend etc to n8n. I want to work on a POC but don’t want the idea get killed or stucked just for requesting software budget. I was an employee for a big engineering corporate, IT budget was tight as big bucks had spent on software maintenance, and expectation was high for customer service and support, on top of SLA.

I agreed with @dp1791 that n8n users are willing to pay at a lower subscription coming from zero cost of ownership. As the market is crowded with ELT/ETL players, you need to aware who is your user base and leverage on it. But if you have a different target audience, such as a $200 subscription, then how do you distinguish n8n against the big players? I think to be a paid company, you want to do a case study or see how others gone thru the transformation journey. I found an interesting link monetizing-open-source-business-models-that-generate-billions.

2 Likes

In my case, Zapier covers 100% of my modest needs with just the starter package, but I was attracted to the do-it-yourself approach of n8n. Unfortunately it looks like n8n
would rather have a handful of multi-million dollar customers than a million of small developers, so while they decide, I will stick with Zapier’s cheap plans :slight_smile:

1 Like

You are spot-on @engowen. I think the problem is n8n cannot fight the “free riders” without killing their supporters too, and to be successful I believe you need lean towards love rather than hate, but that is a tough pill to swallow.

1 Like

@dp1791 to make it clear. We are not fighting anybody! We value every user/community-member very much, no matter if they pay us or not. If you look through our forum, you will see that. Everybody receives top support and help with all of their problems.

About your “you need lean towards love rather than hate”. The world is, most time, not black & white, things are rather grey and exist on a spectrum. That is the same here. Imagine a graph. Each side has very extremes. On the left side is “most love possible” where we would give away all features that exist and will ever exist for free, offer everybody free 24/7 personal live support, and probably best pay users on top to use n8n (again, is the extreme). On the other side, we have no free version + no support at all anymore, and it is priced in a way to maximize the profit for n8n, totally ignoring the community.
Now imagine parallel a second graph for how sustainable it is. On the left, running out of money within a week, and on the right is where the maximum amount of profit gets generated.
The trick and the hard part is not just to maximize “love”, the trick is to find the spot where both are as high as possible. But it has to be past the sustainable point, because the “sustainable part” is black & white, it either is or is not.
I assume the above should not be very controversial. What is more complicated and not as straightforward is for sure where that point is. To do that, we use as mentioned above data (which also includes conversations and input from the community/users) + experience (internal & external).

Is sad to hear that you move to Zapier, especially considering that you now pay the same as for n8n cloud, where you would have much more executions/tasks. Zapier is a great product, but also one with a very different focus. As mentioned in other posts, we do not want to be the cheap/free Zapier alternative, we rather want to be the most powerful tool out there. So it is possible that there was also a mismatch from the beginning. If you have one day more complex use cases where Zapier is not a good fit anymore, I hope you consider n8n again. For now, have a good time with Zapier, and happy automating!

@engowen Thanks a lot. Yes, that are all things we are considering working on. Generally, do larger companies value that they can self-host for security and data-privacy reasons, the power and the focus on more technical users, and how customizable and extensible n8n is.

Generally, I will stop to try to answer every comment, even if it is hard for me. Not because I do not care (I hope at least that is clear by now), it is instead because it is very time-consuming (after all, is there just one me, but many people in the community), and it seems like it does not have the impact I would like it to have. Like that, I have more time to focus on other things which have a bigger, more prolonged impact, and in the end, actions are more important than words.
And to be clear, I will keep on reading every comment as I always did, and each of them will be considered when we make decisions. So I encourage and ask everybody to keep on sharing their thoughts as they help us a lot and show us where we are doing good and where bad. Through that, we can learn and make better and more informed decisions. Thanks!

5 Likes