[discussion] Why RethinkDB (an open source database) failed

Since we are all in the business of open source, this is a very important read.

Are we optimizing (erpnext) for the right market, users, features?

Your thoughts?

Edit: its an old article (from Jan 2017)


I am a long time RethinkDB user, the real time push is nice. Frappe the Framework is doing well targeting developers. This market is crowded with good alternatives and I think it is good not to target directly developers. My thoughts are to keep focusing on the ERP as main market, especially on easy of use and vertical markets (eg: schools), at the same time building a great framework under the hood to attract developers.


Developers bring users and users can be clients over time, direct or indirect.

I think most of the decisions made lately in the frappe foundation have been successful and in many cases excellent, and that is why we are contributing what we can here, some with money, others with knowledge, others reporting bugs, and hopefully we will contribute in more translations principally on the manual.

I have used since the beginning mongodb and one of the things that have done very well, apart from developing excellent software, have been giving free training from the company itself and collaborations very striking as with fourquare. But specifically those who use mongodb here, all without exception, we have made the courses.

Another point is that behind Mogodb there is a company with large contracts, and this also weighs. If it works for foursquare because it will not work for your company? This as an developer/installer gives you confidence.

I believe the evolutionary process at ERPNext is approaching what should be the “knee of the curve”. This refers to the point in a logarithmic curve where the amount of growth (evolution) increases faster over time to the point of almost vertical indicating rapid growth. That graph sort of looks like this:

ERPNext is approaching the point where it should beginning to notice they are moving up (y-axis) faster than they are moving forward in time (x-axis).

However, if you are struggling with finding that “knee of the curve” point, then it is time to go back to your project and think about the parts of that project that will need to change in order to promote that logarithmic growth pattern. The really sad thing to do would be to attempt to control that growth or slow it down so you can think everything through as thoroughly as you would like. Believe me, I have been there, done that, and didn’t survive long enough to get noticed.

You see, those of us with the engineering mindset are prone to wanting to control EVERYTHING!! What we fail to realize is the business (or consumers interests) do are not always logical and tend to follow more of an emotional path. They have to like the product even if it is not the best product for their needs. That is how Quickbooks managed to steal the market from so many of the original software makers in the 80’s and 90’s. They put out a product that was easy to use and promoted some good accounting practices. Their product was not originally the best at getting the job done, but it was easy to use.

Ok, so then why did that matter? Well, because at some point it split the customer base into 2 distinct groups. One group that were accounting majors in college and could be seen carrying around the old HP style calculators designed exclusively for accounting. The other group were the people that did NOT have a degree in accounting but desperately needed a way to manager their business accounting without having to go back to college!

In this case Quickbooks (Intuit, Inc.) went about making their software easy enough for the average well rounded business person to use. The others kept their college level terminology and rules making im place too long to realize they were being overtaken. They were the accounting purists and never saw the Quickbooks train coming. Quickbooks became the default standard for almost all business in less than 5 years because almost ANYONE could use it and master it in a very short period of time. Was it the best accounting available? Certainly not. It lacked (and still does lack) many account tracking methods that are needed in business, but their users find other ways to handle them offline. Over time, Quickbooks eventually filled in all of those missing tracking mechanisms with their own solutions. There are still some things they cannot do, but for the most part, they are a complete system now.
Ok, but how does this compare to ERPNExt and the ERP market? Finally the right question and it all boils down to being able to identify when your user/customer base is about to split into “purists” and “consumers”. EPRNext has been moving along for a long time as a developer’s hobby kind of thing that they seem to have been able to partially monetize with their hosting. However, the fact that ERPNext is finding itself on near equal ground as SAP, means that it has attracted the attention of the largere players in the matket. Now don’t take the “hobby” word incorrectly here. Open source developers are primarily doing what they do to either sharpen their skills with practice, or to satisfy an urge to make something better just because they can (which is pretty much a hobby).

The key now is to find a way to keep the “consumers” coming back to the product and keep the “developers” interested in adding to the core packages with improvements and applications. This requires a different mindset than the engineering type. In my past I developed a very good product that did not exist before and then I tried to control the growth of the company so I wouldn’t crash and burn. Well, I controlled the growth so well, and my product was so good, that well funded competition was able to blow past me with something that worked and was able to meet the demands of growth before I could figure out I was being trampled. I was too focused on improving and not enough on getting it out to the public.

In the case of ERPnext, the primary users almost exclusively need to be other developers just because of the complexity of setting it up and keeping it going. Even just making a new kind of printed document template requires extensive knowledge of HTML. Yet, the consumer following is constantly trying very hard to adopt ERPNext as a viable business program. This is where the ERPNext team is approaching the “knee of the curve”. This is where the users/customer are beginning to split and the team needs to be able to recognize the splinter groups and find the one it wants to keep for its own. If the team keeps on the same path as before, they will always be the “developer” favorite. The real question becomes “Is there enough skilled developers out there to sustain product growth?” Or would a path to something a bit more consumer friendly be a better path to financing the growth of the product?

No longer can the team keep their engineering mindset and still sustain the project. It is time for the team to adopt some business minded individuals in their leadership in order to start exploring ways to make ERPNext the default Enterprise Resource Planning instrument that it really could become. The ERPNext did rightfully pick the small to medium business market to be it’s battleground, but the big players are taking notice. I have been contacted by the SAP sales reps 6 or 7 times over the past 2 months just because I spent time online looking for such a system for a small business. This means they are looking for their footing and they want to outpace you.

All of this points to the truth behind the very first sentence of this post. If the ERPNext does nothing outside of their comfort zone to identify the coming split in their users/customers, they will be run over by other well financed players. The key is to figure out the differences in the users and pick a group that you believe will best sustain the project.

If you choose to stay with the “developer” type customer as your primary target, then you will need to find small ways to keep “consumers” interested until your product is mature. If you choose the “consumer” type customer as your primary target, then you need to develop an new way to keep the developers coming back and working on the “consumers” concerns until the project reaches maturinty.

The ERPNExt team needs to look inward and decide if it has all of the right members to continue to grow. Does it include business minded leaders? Does it include strategic project managers? Do they all talk to each other and really put the project ahead of their own interests? How many of the team members are only wanting to sharpen their coding skills? How many want to be part of business that could one day launch a for profit corporation with this? Has anyone aver looked at how RedHat managed to keep its developer community and still sell a great operating system?

I may be wrong, but I have seen this path many time before and it is not easy to accept change. However, change is coming and if the team is not prepared for it, they will get left behind. The team does not get to control the change. That control was lost when the product took on a life of its own. The team now only has the chance (very much like a parent) to help guide their creation to success. As a parent you watch for changing environments in order to guide your kids to the right paths. It is no different in the life of a product. It is just time to start identifying the changing environment and figure out how to keep their creation relevant. That begins with knowing how your users will divide themselves into tribes and then deciding which tribe will become dominate. Which path will the project team see as the best for their future.

Open source and successful projects are NOT mutually exclusive. It just requires balancing the team with different perspectives that are not normally comfortable to engineers. Business, Sales, Accounting, and Project Management are all uncomfortable things for the average engineer. Finding that balance of knowing how to bring in money while still keeping developers interested are the keys to success for this project. The path to the money is the most difficult to grasp and that all comes down to your customers. The ERPNext cloud platform is already a great start toward financing the future. Even packaging stable releases for download for a small fee would be a benefit as long as it was sure to work for the user.

If you decide to target “Consumers,” user experience needs to change. Frustrated consumers tend to spread the bad news far more than they do good news. It must be easy for them to adopt the program into their businesses or else all of your work will be for nothing. There needs to be a backup method that inspires confidence. There needs to be a way to edit document templates without being a HTML
programmer. There needs to be single reference or document containing the logical path of data so that users can easily figure out how things are connected. These are some of the things that need to be in place if you target “consumers” as a way to finance the project.

This quote was worth repeating, because it gets at the heart of how a project will survive. The result really depends on how well rounded your team can become, and if you can identify the market environment changes that matter to your project. Are you going to be a developer project or a consumer project?




Thank you for that link. A few things are stated that applies to almost all companies: 1. Who is the customer, 2. What does the customer want? 3. Where are the customers? 4. How do we reach them?

I think it’s important for Frappe/ErpNext to also answer these questions very clearly. I think it’s important for the core team to ensure a sustainable revenue model because both the core team and the foundation need to be strong for this project to realize it’s potential.

Purely, from a commercial stand point, the core team needs to target businesses to sign on. GST, Digital India and the increase in internet penetration are great drivers. Lack of another cloud-based ERP solution means there is a void but the question to ask is does the average company/customer care about this? Perhaps a GST-centered marketing push with easy invoicing and purchasing could drive adoption in a big way. Just thinking off of the top of my head.

Would love to hear your perspective.


1 Like

The key word in the above quote is"customer." We already know who they are. They are the small to medium sized businesses that want to leverage a software package to help them better understand their business and make better business decisions. Where they are and how you reach them is almost completely already in place. When searching for open source ERP systems, there are not very many around. Most of them were abandoned a long time ago. Those that are left are easily searched and only ERPNext shows continuous growth and improvement.

If you want to do more advertising to reach “customers” (notice the quote was not “developers”), then you need to focus on what kind of user experience will bring "customers’ back every time and have them singing your praise. Only when you have made the “product” worthy of average business consumer praise, should you begin spending time and money on advertising.

The story at the link is one of humble beginnings and unfortunate lack of perspective. They knew they had the BEST possible product, but got lost in the purist mindset and never realized how to get it to the public in time. They missed the customer signals at “the knee of the curve” and went the wrong direction.

I think @rmehta is noble and asking the right questions to see his project make the next step in it’s evolution without seeing it fail. You have to respect the fact that he wants to make the right moves as early as now. I wish the project team well in their next brainstorming session. Thinking this part through is critical to their eventual success.



One thing that I forgot to comment on in the previous message, and that seems to be normal but it is not so, is that Frappe listens to the community as in the case of account numbers. This things are debated and if you see that you are not asking for anything out of place and that does not break the workflow of the application, the job is done. This, as a user for 20 years of free software, does not happen as often as is desirable and is a big plus. This makes us, ‘small fish’, feel part of the community and that we will want to collaborate as much as possible.


Thank you everyone for your comments and feedback. I guess this is going to take a long time to parse :slight_smile:

MongoDB also had large funding so they could have a sales team that goes after these big projects. I think we need one “famous” company to start using ERPNext and become an evangelist, but I have no idea how!

Yes growth is good, but no where near phenomenal. Compared to Odoo, ERPNext is still very small (but growing!)

I also agree with this. Ease of use should be the #1 focus and it has always been a focus for us. And we still have a long way to go identifying where users can get stuck. At this point, I think ERPNext is pretty easy to get started. The problems in ERPNext are of the nature of the process of shifting your accounting from one platform to another, there are bound to be some things you have to adapt to in a new system. We are working smoothing that process too.

Would love to find a better way to do that :slight_smile:

Like I said before, the problems related to switching are hard. A company needs to close its books in one system and open in another, for a company with an existing business this is a non trivial process. We have identified some places to make it faster, but yes this seems to be the biggest bottleneck.

GST has been miked by every vendor out there so I don’t think that opportunity exists (or is big enough). The point you make about the void is real. Again I think the bottleneck is the cost of switching. How can we make it lower…

Thank you for your kind words :slight_smile:

:blush: Thanks! Almost all of the product has been built listening to users. We are not religious about features but about consistency in user interface, naming, workflows etc.

This is one of the Base ideas, to have a balance between the system out of the box - with its all functionality - and the customisation.
Since a period i want to use erpnext completely - but every day i find thinks i did not understand, or it is hard to customize. at the end, sometimes i am not happy about that, but on other hand i have the hope, that all processes become better and many small things can be realized. Most of the functions are islands.No clear workflow.
But ok, i have to learn much to understand the philosphy of the system.

what about the idea, to have a market of customisations? Also for most websystems, the base system is free (open source) - but i can get extensions and so on for a fair fee.
Is there nobody, who hase one or more sales-recipts (other, better design) layout, which i can buy und customize only minimal.
Is evrerybody starting at level sero und starts to setup all new?
That is sometimes my question - and a little part of the answer, to have a look to the customer.

useability, better processes, functional customisations (widgets) would be things, customers will make happy.
At the end a base of business - for erpnext - also the system-integrators / it-companys which are setting up a system for the small and medium business.
As i workted on it, i get requests of some companys, whic are searching for a software like this. But actually it is a long period of customizing, until anybody con really work on it.

Thanks a lot for the great system - and think about the customer :slight_smile:

I am sure you can get your own Invoices made at a reasonable cost. but yes, having pre-designed invoice formats is a good idea.

We have worked on this for GST and can contribute a good HTML format. Where should we do it?


Can you share screen shot of GST invoice which you made?
If you think it is useful, you can put code on Home · frappe/erpnext Wiki · GitHub

I am also working on GST Print format.

Maybe a broader approach towards better usability for prospect ERPNext users but I think it might help in terms of attracting bigger organisations in the healthcare industries.

I think the fear of a big efford of switching to a new system can be smothened by presenting more advantages in regard to time consuming or complicated workflows. If the target organisation has an incentive to switch by getting a so to say quick and big ROI then they will do it no matter how steep the hill seems at the beginning. The fact that you guys were approached by a big big Dog (starting point for the whole "lets improve the HR Module story) proves that ERPNext has the potential to target the right clients.


So very true!! The client that wanted a better HR module before they switch is really one outside the original scope of the ERPNext target users. It sure is nice though to see that this project we have all been working to improve is getting traction outside it’s target audience.

Kudos to the foundation for helping to keep this project organized enough to show up of the radar of much larger clients.

I look forward to hearing more good news about the path to success the foundation and developer community has us on.