ERPNext with Rushabh Mehta

Hello, and welcome to Podcast Dot In It, the podcast about Python and the people who make it great. I'd like to thank everyone who has donated to the show. Your contributions help us make the show sustainable. When you're ready to launch your next project, you'll need somewhere to deploy it, so you should check out linode@linode.com/podcastinit

and get a $20 credit to try out their fast and reliable Linux virtual servers for running your next app. You'll also want to make sure that your users don't have to put up with bugs, so you should use Rollbar for tracking and aggregating your application errors to find and fix the bugs in your application before your users notice they exist. Use the link rollbar.com/podcastinet

to get 90 days and 300, 000 hours tracked for free on their Bootstrap plan. You can also visit our site at www.podcastinnit.com

to subscribe to the show, sign up for the newsletter, read the show notes, and get in touch. And to help other people find the show, you can leave a review on iTunes or Google Play Music, and tell your friends and coworkers.

You can also join the community at discourse.pythonpodcast.com

to find out about upcoming guests, suggest questions, propose show ideas, and talk to other listeners of the show. Your host as usual is Tobias Macy, and today I'm interviewing Rishabh Mehta about ERP Next.

So, Rishabh, could you introduce yourself, please? Yeah. Hi, Tobias. This is Rishabh.

I'm the founder of ERP Next.

ERP Next is,

is an ERP system for small businesses. It helps

businesses do accounting,

inventory, sales, CRM, purchasing,

projects, HR and payroll, and, and and a lot of other stuff. Yeah. So we're we're based in Mumbai, India, and we're a team of 15 people now.

And how did you first get introduced to Python?

So I, I started up as a a basic developer,

in in high school. My first computer was an 8088,

PC XT with 6 40 k RAM, and and and, and the only thing it it will run is text based, software. So I started with GW Basic on MS DOS, and then I moved up the moved up the basic chain with, Quick Basic and then Visual Basic for DOS and and then Visual Basic. And so, basically, I was a Windows, developer for many years. And then when I was in grad school, I I I was working on a project that was based in Java, so I had to learn Java. And,

I did not really like it. And then I had another conference where people talked about Python, and it was really nice.

I checked out Python. It the the syntax looked just like basic, and and I really liked it. And the more I discovered, the the better it was. Lots of libraries,

a great community around Python. So basically moved, to Python and,

and loved using open source and happy to get away from the Microsoft ecosystem.

So the

project that we're talking about today is an ERP system. So I'm wondering if you can explain what that anagram stands for and the kinds of businesses that generally require that kind of software.

Yes, Shweta.

ERP is basically an enterprise resource planning system. It's basically a system that helps organizations

manage all their information. So basically a data ware data warehouse for an organization. It helps them do accounting,

helps them,

keeping, keep track of their orders,

the shipments,

invoices, payments.

That's the core. And then if if the organization has inventory, then, it helps them keep track of where the material is, how much, what items, what are the serial numbers, what are the lot numbers,

what's expiring when, and all of that. And then you also have HR and payroll, which you also kind of integrated in 1 platform. And then we have project management. You have purchasing,

when you're an organization, you have centralized purchasing. So any IP system helps you manage all of that. And, we also have an ERP next. Since it's a web system, we also have a website generator. So, I mean, you, you already maintain all your products,

in the ERP system. So your website is is just a view of those products. And that's that's how, so so basically, ERP next has lots of features.

And what was the problem that you were trying to solve when you first thought of creating ERP next?

And I'm also wondering what factors led you to the decision of writing it in Python as opposed to Java, which I know is the language that generally gets used for these kinds of systems?

So,

so after I graduated, I joined, my family business. My family,

at that point used to run a business in manufacturing of furniture, and they were reeling from a bad ERP implementation. They had a there was a local vendor, who was not able to deliver their purchases their their purchase licenses, but, the customizations were not coming through. And the whole project was just getting delayed. And in all my foolishness, I thought I should, you know, maybe write my own ERP system. And then there was a motivation.

Yeah. So this was again, in 2005. So in early 2000, there was a a a lot of hype around semantic web and

and using,

metadata to kind of model stuff. So I use a lot of those concepts and built, an ERP system that was based on metadata. So that's how I started using in, using LP 9. And and the ERP next was actually my first major project in Python. I used to write in Java or

or basic, mostly in basic before that. And, so so that's how, then that's how we started with ERP Next.

And what does the architecture of ERP Next look like? And what are some of the biggest challenges that you faced while you were creating it? So, with ERP Next, we also created a web frame. We we call it Frapi. So again, this is 2, 005. When I was building the system, the only framework available for Python developers was ZOOP. ZOOP. And, it was horrible to use ZOOP because it was not intuitive at all. And, I I did not like the the decisions that were made with Soap. So I started kind of writing my ad. At some point, I decided I should write my own framework. I think Django was also maybe started around that time. So I didn't know about Django and Flask and other stuff that came in much later. Probably started around the same time, but became popular much later. So so, basically, I started as a solo developer for, for the 1st 2 or 3 years. I was the only guy building the framework and the application on top of it. And and that's how we grew. And then my family sold the business, and I continued,

building the ERP solution for the companies. And that's how I kind of made it into a broader,

a broader project. In terms of architecting, I think, the complexity of building an ERP system is huge because an ERP system is actually

a a combination of many different kind of projects. Accounting itself is a project. Invoicing is a project. You have an ERP system. You have lots of you have a report builder. You have,

you you have a sales system. You have a CRM system. We have emailing, which is out of the box.

We have I mean, we have lots of stuff. So basically, just architecting all of this was the biggest issue. So

my whole my initial idea kind of where we have a metadata modeling system really paid off because

because of the metadata, we were able to standardize a lot of stuff. We were able to write really

targeted code for business logic and not recreate the UI and wiring for every object that we created. So that that was a big win, and that helped us,

build such a large system with a very small team for many years.

And when you started building Frappe to help support your development of ERP next,

did you initially set out to make it a framework from the get go? Or was it something that you extracted from the larger project as it started to take form? Because I know that in a number of cases, a framework is something that gets extracted from an existing project when you realize that there are parts that can be generalized.

So, so so when I started out,

on my own, with a friend, I mean, I I didn't understand about ERP and then, like, just setting it up as a product. That that came in a bit later. 1 of the first projects we got through a reference was to build a a management system for a facilities management company. So facility management companies

is a company that manages, like, everything from plumbing, security, housekeeping

for a large corporations, like an entire

shopping mall or an or a bank with, multiple branches. So so they want a web based tool where they could track,

all their,

sites. And so so, basically, we use the same metadata modeling for that. And I was actually building 2 projects at the same time, so it kind of made sense to start, to to build a common library as a separate framework. I don't think there was an explicit thought of,

set out to build a framework, but it just happened that way.

And for somebody who's used to frameworks like Django or Flask, how would you compare frappe to those systems?

So frappe comes with, metadata modeling that's built in. So the models that you create,

the models in FAPI are called doc types. And doc types are, also doc types in that sense. And you can create a new doc type from the UI. Also,

in Frapi, you get a view which is we call the desk, which is,

which is a single page JavaScript application that,

in in which you can navigate through all the different models as a user. So it's kind of a Django admin,

a very,

a very customizable

and a very,

extensible Django admin in a sense. And,

and and inside the desk, the forms are all generated from the metadata. So all all the so so so so the UI is actually never generated at the server side. The server only gives out so we use basically a rest API to, manage

with the server. And the server only throws out, JSONs,

in that sense. And the entire UI is built in JS. So that's 1 of the big differences. And also we've packed in a lot of stuff, in frappe. We have emails that are integrated. We have,

we have real time integrated. We have background jobs with Python r q. We have, I mean, we have lots of stuff. I mean, we have a form builder. Then we have a portal UI, which is the more traditional,

framework UI where you where you can use, Jinja templates to build any custom web interface. And we can't kind of call that a portal. So I would say it's basically, it's a it's a much larger set of, libraries that come with Frapin, and that makes it opinionated too.

And given that the front end is so tightly integrated with the rest of the framework, is it possible to integrate other front end frameworks such as React or Angular as well?

Yeah. So,

so so so like I said, we come with our own,

view, which is the desk view, which has where where you see icons like you would see on an iPad, and then you click an icon and that's a module. And then you have a list view, a very standardized list view, a very standardized form view. And all that is basically generated from the metadata. So you don't have to write any UI for that. So you have assignments out of the box. So you so if it's if if your form is, customer form and you would assign a customer to somebody,

I mean and it could be any form. So all the assignments and emails and all of that, users and permissions also comes out of the box. That's another big, big piece. You could still

start your own templates right from scratch, like you would do in Flask. I mean, you could, set up your own base template and then, you know, include your own, JS,

libraries, whether you build it in React or any of that stuff. And then you can use the Flappy REST API to kind of

manage the, application, the data.

And given that the forms are automatically generated from the metadata that's contained in the models, does it also have automatic form validation as a feature? Yeah. We have a lot of form validation, like mandatory stuff, and and and then we have foreign keys which automatically become kind of,

auto select, in in the form. So, there is lot of built in validation, but not everything is configuration. I mean, there is there is a fine balance you have to hit between configuration and writing code. So,

a lot of this stuff is very configurable.

A lot of the stuff you have to code for. But, yeah, a lot of the validations do come built in.

So what are some of the projects that you consider to be your competitors in the ERP space? And what are the features that you think would convince a user to choose ERP next? So if you if you see the ERP space, the biggest

competitor we have is Odoo, which is based,

out of Belgium. And, it it's probably got a a community that's 10 times bigger than ERP Next. But other than Odoo, I think ERP Next is the most actively developed, project,

in this space. And Odoo is also developed in Python, so that's,

that's also interesting.

So, basically, what really sets, ERP next apart is I think, I think it's much better as a project because, you know, obviously, I

it's me. So, but but, generally,

it's much more modern in the sense, Odoo. And Odoo started,

and and there's and then we also have a unique business model. Right? We don't,

make any revenue from customizations. We don't make any revenue from partners. I mean, that's how usually business oriented open source projects make money. So we make money on the cloud, which is the end end user. And,

for the entire self hosted ecosystem, I mean, that's completely kept separate from our commercial activities. So we have a strong incentive to,

make the application very easy to set up and easy to use and easy to implement because, on the cloud I mean, that's the reason people are gonna subscribe and pay for our application. What versus Urdu,

where it derives 85% of its revenue from its partners. So basically, the partner faces the end customer. A lot of the improvements and customizations that the partners do don't end up in the core product. And another,

problem I see with Odoo and a lot of commercial business,

commercial open source, business apps is that they are fraction fragmented into apps. So, basically, Odoo boasted there are 4, 000 apps built on on Odoo. And say if you want to build something, so if you want to choose an app so if you're an if you're a user, that's, number 1, that's terribly confusing because now you have to go through, like, 4000 apps to search for what really works for you. And if you find something, you'll find 3 competing vendors giving you the same. And,

and depending on what you choose,

so by doing this, you know, everybody's replicating the effort. And, and and just because the whole ecosystem is designed to make money for partners, then people don't have

a a completely free, approach to the way the things happen. So in Odoo, you will most likely come across paid apps, and that's the basic way people make money. And even Odoo makes money because from the current version, they have become paid. So,

so that's 1 of the big problem that I see with Odoo Ecosystem. With the ERP Next, we have decided we'll make it a monolith application. So when we were designing,

we're implementing ELP next for a school, we decided that it's not gonna be an app on ELP next. You know, it's gonna be a part of the core because even the school has accounting, even a school has payroll. A lot of things are deeply integrated. And somebody

using,

and somebody else

improving I mean, so the improvements you make for schools also, you know, extend for the whole product. So so so that's 1 of the 1 of the core areas where we are separate, from Vodu. And there's another big big difference is that updates are not part of Odoo. So if you if you if you install Odoo, you cannot update it because the updating tool is a paid, is a paid service. And if you buy the enterprise license, you will only upgrade. But we have built a tool which we call the bench, which is our

deployment tool.

So it's it's just running 1 command, the bench update, and it's going to, update your entire app. So that's that's a big difference.

And for somebody who does want to run ERP next on their own systems, what are the requirements for that, and what does the scaling story look like for, for somebody who needs to scale it up to a large deployment?

Yeah. So so like I said, we we have a deployment tool. We call it the bench, which basically comes with an opinionated production setup.

The way, so when you deploy an ERP next instance,

we use NGINX. I mean, the our production,

deployment

comes with, you know, pregenerated scripts for NGINX. We use supervisor and then Gunicorn to manage the web processes.

We,

our background

workers are managed,

by Python r q, and we have Redis. We use Redis for caching. We use Redis for the queues, and we use

Redis for Redis for 1 more thing.

Yeah. So we basically have 3 Redis in yeah. So we use Redis for PubSub with a node instance. So we also have a node instance running when you install the Flappy framework for real time socket type communications. So basically, when you're viewing a document, somebody has used a document, you get a a small push notification saying who else is viewing the document. So all that real time communication happens through a socket IO, through socket IO from a node node process that also runs on, on the server. So this is a fairly complicated

setup. We recommend you use a clean,

system in Ubuntu or a Debian or a CentOS. I think the we have Ansible scripts that are,

that can do, like, a 1 click install on any of

And and these guys, you know, virtual machines are the norm. Right? And you can go on Google Cloud. You can go on Amazon, and it's very easy to set up DigitalOcean and set up a virtual machine and line out. So,

so, basically, just

just fire up a virtual virtual machine and, you know, run our install script, which is based on Ansible. And you and you get a production setup ready. In terms of scaling the for an ERP application, the bottlenecks are on the database size. So, basically, you,

if if you're hitting really large scale, you know, database optimization is something where again, from we learn a lot from our own cloud customers. So we have a lot of users,

who have a lot of data, and we are constantly optimizing our, database index indexes and all of that. But I mean that I mean, that's where you have to really, work on scaling the database side.

What are the database systems that you support in ERP next?

Yeah. So so so we support MariaDB. Out of the box, we don't support any other system because we also have, certain queries that we've written. So we so Flappy comes with its own ORM, but the ORM doesn't include, complex joins and querying. So so for complex joins and querying, we just write SQL. Lot of people think that is a problem, but we think,

I mean, it's for for an ERP kind of applications, sometimes writing,

direct SQL queries,

is is so much easier, and and it leads to so much cleaner code, and also performance. So so we are tied to MariaDB,

for now. I've definitely been in a number of conversations

discussing the relative merits of actually pushing any sort of logic into the database. I know that a lot of people are sort of allergic to the idea, and they want to keep everything encapsulated in their code base and not actually rely on the database itself for doing any sort of logic. But there's also a lot of potential performance gains possible

by actually leveraging the capabilities of the database and pushing some of the business logic into that in the form of triggers or stored procedures or things like that. But it does create some complexity in terms of making sure that the database is versioned appropriately, and it can also oftentimes

lock you into a particular database vendor, which I know at least for a number of self hosted or,

self managed applications, people generally want to give you a larger option of the database back end. But it also comes at a trade off of not being able to gain all of the potential benefits from your database system. Yeah. We don't use, triggers and stored procedures, though. I mean, we we use direct

SQL for querying for joins. Basically, we use SQL. But but yeah. I mean, I I agree it's a trade off. And and it's also perception, right, that, that MySQL and MariaDB has lesser features and then Postgres has more features. But but but that's changing because, the rate at which MariaDB is being updated is also also very good. And then it's backed by all all the biggest companies are using MariaDB or MySQL,

for all their apps. So so far, we haven't come across a use case where we we have regretted taking the decision. Sometimes people ask this question on the forums, but, you know, these days, it doesn't really matter, you know, at least for for the kind of users we we'd be looking at.

Yeah. I think a lot of the public perception around MySQL was also tied to the acquisition by Oracle and the way that they were managing the project. But now that MariaDB has split off, there's definitely been a lot of goodwill captured by that move and by the project

itself. Yeah. And and and they're making a lot of improvements. I mean, it's constantly they're pushing new features. And, yeah, we we we're pretty pretty happy with the way the things are going with Myrdp.

And on the marketing site for ERP Next, it's advertised as being for small and medium businesses. So I'm wondering what the characteristics of larger businesses are that might not make them a good fit for the features or structure of ERP next.

It's more than in larger smaller and medium businesses. It's it's only because it's easier to sell to smaller businesses. Right? So if if you look at the ERP industry, it's not been disrupted for ages. I mean, it's because the larger industries are still tied into the SAPs and oracles of the world because it's mission critical. I mean, SAP is still running r 3. Right? I mean, it's like an ancient piece of software. But people don't take risk with ERP software. So it's very difficult for a new vendor to kind of push into this space.

And it's even difficult for an open source,

kind of

platform because open source is still,

it's great for servers and it's great for back end stuff, but it's still for business software, open source. For for the ERP industry itself, it's it's less than 1% of, of the total deployments would be probably open source based, at least in terms of revenue. So it's it's a very small industry.

Enterprises have still not warmed up to open source for enterprise, and it and it's it's terribly hard to sell to enterprises. And 1 of the things, me, personally, I I I I wouldn't sell to save my life. Right? So so it's so much easier. I have a product dial out writing code, and that's what I like doing. And and so putting it on the cloud and letting people try it out for a 30 day trial or self hosting it and deploying it on their own is is is the best way for people to try. And these are mostly small businesses, innovative businesses,

businesses who are in the new age kind of, you know, whether it's open hardware or whether it's, you know, ride sharing and you know, all those kinds of businesses,

climate related businesses like solar. We have farms,

farm equipment companies. We have 3 d printing companies. So it's a biotech company, Genentech. So it's the more the newer then the newer kind of industries are actually using ErpX.

So what are some of the most interesting or unexpected ways that you've seen ERP Next put to use?

Like I was saying, you know, we we have a lot of these new age businesses. So recently, there is a,

so so so there is an Uber like company in India which is, using ERP Next. And they're they're building actually yeah. They're building a full ride sharing system on top of ERP Next. And along with rides, they're also doing deliveries. They're doing

meals.

I mean, it's it's all based it's all based on the common platform. So I had a chance to visit them a cup, a couple of weeks ago, and that was really exciting.

And, you know, we we are always surprised, by the by the kind of companies,

you know, that use ERP next. I mean, like I said, we have 3 d printing companies,

electronics manufacturers. And and and surprisingly on our cloud, the the highest number of paid subscribers are IT service company.

So even though the ERP is traditionally built for a manufacturing kind of a company, but apparently, there is a lot of, market at least for the kind of,

positioning we have for service companies too.

Yeah. As I was looking through the list of features, particularly the CRM capabilities, I was trying to think about ways that I could potentially put it to use for managing the podcast and the guests that I have on here because that's 1 of the problems that I've run into is being able to keep track of all the people that I've contacted with. And for the time being, I've been using a Trello board, but it doesn't really do a good job of being able to capture their contact information along with the stage of the workflow that they're at. Yeah. You should definitely have a look. You know, I'm happy to help. You know, if you ever need help, just ping us. You can ping me, and I'll be happy to help you. And are there any common issues that you have people contact you about when they're trying to implement ERP Next for their businesses?

So so we are a very active community. We have, I mean, we have more than 900 issues over on GitHub. Most of the issues,

if you see,

especially when so so most of the issues come whenever we release a new version. So people migrating to the newer version, that's the that that's when we have the highest amount of support. So right now, we're between versions. So we are very relaxed.

The issues do come up on installation upfront, but those have also kind of so maybe we moved everything to Ansible based scripts. And, you know, you can easily, edit those scripts and you can,

tweak it to build your own

deployment. So installation things have gone, have become much lesser. But the the people who use this application are also not tech oriented people. You know, these are small business owners or

managers and small businesses who really don't know what a lin Linux. They've never interacted with the Linux command line, and they're just probably this is their first real

exposure to setting up something like this. And, you know, a lot of newbie questions around that. A lot of questions like, so so so basically, that's 1, you know. And the other is around customization. So once they hit a so in the middle, it so so if you're able to set it up, I think using it is fairly simple because, like I said, we are a cloud first company. And then the questions come around customization, where people want to have custom scripts. They wanna have different formats for their invoices. They wanna set up custom workflows and then you eventually create their own models and extend the product. So a lot of questions around customization.

And are there any interesting projects or features that you're working on for release in the near future that people might be interested in hearing about? So 1 of the biggest barriers for small businesses to use,

such a thing, especially like I said, you know, these are end users who really don't understand the Linux command line is is the whole client server setup. And, I know you're talking about cloud and software as a service and all of that. But the vast majority of businesses, I think, are still offline in that sense. You know, if you're using an accounting application or a point of sale, you know, it really makes sense to build an offline first application. So so what we're,

experimenting right now is to build a full offline version of ERP Next or at least a part of ERB next, we'll call it maybe a small accounting app, and then put it up for download. And then, and so we're building this off Electron,

the Electron framework in JavaScript. We probably

integrated with our current

architecture, where we the kind of models and the top type concept that we have. We'll still use the framework to to kind of, design the models, and then it will spit out maybe a GS app that we can then compile and then, you know, it will be available for download. So so if you're a small business, you you just download it and start using it for invoicing and your daily accounting. And you'll be surprised there is no good open source,

accounting application for small businesses. I mean, they I mean, you have, you have leisure SMB and you have some stuff, which is really ancient, but there is no real modern accounting application, which is really what which is widely used for accounting. So we're thinking we'll that seems to be a much bigger need in that sense. And we have learned a lot over the years about small businesses with the LP Next. We learned about the functionality. And we're really excited about going this offline first way. And then and and then if you build this offline first, then we'll also build the mobile apps in that sense. So so so so that's our,

that's something we're currently experimenting with. And so that would be for people to be able to use it as a desktop based application and not have to require interfacing with any sort of servers at all. Is that correct? Yeah. So it'll be a desktop based application which will be based on Electron, still be, written in JavaScript. And, you know, it'll be offline first. And then we'll also have,

an option if they wanna if they if they have a Flappy server in the back end, they can make it into a multi user application. So so we wanna build it that way. Build it offline first, but also make make it like a multi user thing if they wanna do

it. And is there any sort of minimum size of business that you think it would make sense for using ERP next? Or is it really something that is useful for any scale of business even if it's just a single user? So it depends on the majority of the business according to me. You know, if you're hurting with data, you know, like you say, you know, you you're probably hurting with your contact list, for example. You know, that that's when you really start looking in ERP. You know, you could be a single person or you could be, mal you know, you could be a company. Even as a company, you could be happy with Excel sheets, you know, maybe if you are a 20 people company. But,

you know, that's, but at some point, you you you will realize that things are not in order. Like, either you're missing deadlines or, you know, you are you're you're falling you're falling back on certain stuff. And that's when you want things to be organized, you go for an ERP. It really doesn't depend on size as much. I mean, after a certain size, it does it becomes a necessity. But, smaller businesses are as complex as a larger business. So, you know, even if you are a single, if you if you're a furniture manufacturer,

if you're if you're a carpenter and, you know, you're you're not too much different from a furniture manufacturer in that sense because, you know, still, you know, you're buying material,

you know, you have, you you have products. You know, you have you have deadlines, and you have customers

you have to keep track of. So so it actually depends on, you know, do do you feel the need for it? Are there any other topics that you think we should cover before we start to close out the show?

No. I think it's been good. We've covered a lot. Okay. So for anybody who wants to follow you and keep up to date with what you're up to, what would be the best way for them to do that? I'm there on Twitter. We have a very active forum, to discuss.

I I could type it out. Yeah. Yeah. You can just fill in the contact information. So for anybody who wants to

see what you're up to, they can just look at the show notes and those will be on the website and they'll also be included in your podcast app of choice. And so with that, I will move us on into the picks. My pick this week is just going to be WordPress because I've been using it for actually rebuilding the website for the show, and it's made things a lot simpler

than trying to build it all out custom. So for people who are interested in having a fairly easy to manage and deploy website with a fair amount of customization capabilities, WordPress is still a good option. And with that, I will pass it to you. Do you have any picks for us this week, Rishab? I just finished reading this book today.

I don't know if, the, Ready Player 1. It's a book on gaming and, geek culture. I don't know if you, if you read it. I have not read it yet, but I've heard a number of people recommend it. So I'll definitely have to check it out,

sometime soon. Yeah. So that's my recommendation. If you if you, you know, if it's it's a geek fantasy novel and it's and I I I just found out Spielberg making movie on it. So if you grew up with computers and gaming and, it's it's amazing. It's really nice. Great. Well, I really appreciate you taking the time out of your day to join me and talk about ERP next. It definitely sounds like an interesting system and 1 that I'll likely be checking out to see if there are any ways that I can use it in my business, and in my podcasting. So I appreciate it, and I hope you enjoy the rest of your day. Thanks, Ramesh. Thanks for having me. It was really nice talking to you. And,

yeah. And I hope, wish you all the best for more more podcasts. Great.

The Python Podcast.init

Summary

Brief Introduction

Interview with Rushabh Mehta

Keep In Touch

Picks

Links

The Python Podcast.__init__