Break On Through (to the Other Side)

I and Hari Varma (my co-founder at Vasista) started working on building a cloud ERP in early 2011. It was a very modest initial setup with a couple of interns and the two of us cooped up in a small basement office in Banjara Hills, Hyderabad. Having worked in the software industry for more than 12 years in the Bay Area for various enterprise software product companies, I knew that building a serious ERP from scratch was a tall order. But luckily for us, there were several open-source ones that gained traction across enterprises. We finally honed in on an open-source ERP called Ofbiz that was part of the Apache family.

Our idea was to initially create a SaaS-based ERP that is well-suited for SME clients in India — a sort of go-between SAP on one end and Tally at the other end minus the deployment and maintenance hassles. Some of the pluses of Ofbiz were it had a mature data layer (Entity Engine) backed by a very good enterprise data model and a solid Service Layer. An in-built MVC framework provided a basic UI widget framework with a subpar look and feel. Since Ofbiz catered mostly to US and European markets, Indian business processes/rules especially related to accounting and HR had to be built from scratch by us.

We rolled up our sleeves and tinkered around with the framework and during this period we also hired our first FTE (then fresher) who turned out to be one of our best technical hires. He stayed with us for more than 5 years and became a top-notch full-stack developer. We also hired an accountant/function consultant to help with testing and small rollouts. In the initial year, we didn’t have any paying clients but luckily for us, Hari (who is a Chartered Accountant) had some clients from his earlier consulting gig who allowed us to use their data to do small pilots with our product.

Our original business model was to offer accounting and payroll services and give away the product for free to SME clients. But during the initial pilots, we realized that the market was not mature enough to build a scalable business that offered these types of services. Moreover, we would be taking big risks related to compliances since it was common practice for companies in this segment to not be very transparent in reporting all transactions.

But we were encouraged by how the product was shaping up and we were able to get the core modules (Accounting and HR) ready for the SME segment. So we quickly pivoted to selling our SaaS product to larger enterprises and as we knocked on doors to get our first paid client, we got a break with one of the large state Dairy companies. They were looking to replace their 20+ year old Sales and Distribution Management system written in FoxPro! We built out the solution in less than a month working literally day and night. In addition to building the solution we also had to train users to make entries in our system using the browser, mouse, keypad, and the whole shebang! This turned out to be even more challenging since most of the staff were close to retirement (or came back as consultants post-retirement) and were reluctant to take the leap from DOS interfaces.

We overcame these challenges and after a successful parallel run for one month where we tallied all the orders between the two systems (over 3000 orders were being processed per day) they pulled the plug on the old system. With the successful pilot, we had an order from them to build out the rest of the core modules for their supply chain.

We worked hard for the next year or so understanding this new (at least to us!) supply chain and building/rolling out all the core modules in the remotest corners of the state. We ramped up the technical team by adding a couple more FTEs (freshers) and with this small team, we had the core Dairy ERP product ready. This was christened Milkosoft by the client, a bit tacky but clearly conveyed the message and the name stuck!

Digitizing the milk procurement processes posed some unique challenges for us. Unlike in western countries where the farm sizes are large, in India, the vast majority of collection happens from small to marginal farmers who typically own 1 or 2 cattle. It’s quite amazing that large dairies procuring 100K+ liters daily are able to aggregate these 1 to 5 liters of milk from each farmer (located in remote areas) twice a day. The milk movement happens on a daily basis from the source to the factory hundreds of miles away while ensuring a cold chain at all times. The billing and payments cycle is typically on a weekly or bi-weekly basis and payments reach farmers via intermediaries at the village collection centers.

A lot of initial technical challenges involved making sure that the application performed as fast as their old desktop programs even when operated from remote areas with spotty internet connections. Integrations with various testing and weighing machines (IoT even before the term was popular!) wherever available to automatically capture data and ensuring material balancing across the milk movement supply chain. We also developed a flexible pricing engine which is quite unique for this industry. The pricing primarily depends on the FAT and SNF content of the milk and various additions and deductions are applied based on quality parameters. This pricing engine can handle any price chart used across the industry without the need for any code modifications.

The next challenge was rolling out the solution across the procurement supply chain. We were asked to do a UAT at one district where entries had to be made in the old FoxPro application and our system and the entire billing process tallied exactly to the last paisa. This was a tall ask since there were 12K-15K data records that needed to be punched every fortnight from manual sheets of paper. The data operators would typically do this in the last 3–4 days of the period. Due to some vested interests data entry into our system invariably lagged making it impossible for us to tally the systems in time. We even developed data migration utilities from the old system to eliminate the additional data entry but this was not acceptable to them for UAT. And so this continued for 3–4 billing cycles and our team was getting increasingly frustrated.

We finally got a lucky break when during one period, after all, entries were made in their old system, there was a major system crash and they were not able to recover any data. This forced them to scramble and expedite the remaining data entries into our system and generate the billing from our system so that they didn’t miss the payment deadline for tens of thousands of farmers! There was no looking back from here and our application soon became the primary system at this location. To their credit, these users become our product champions and helped with the training and rollouts at all the other districts. Soon our solution was automating a lot of data capture right from the source at the village collection centers and ensuring material balancing across the milk movement supply chain.

Our initial success with this large dairy helped us grab orders from other large dairies and before we knew it we had become a pioneering SaaS Dairy ERP provider in India. Our first SaaS implementation is now a decade old and is still going strong by generating thousands of sales orders and purchase orders across thousands of villages daily. This is a testimony to the clear product-market fit that we were able to achieve and the tremendous value we’re generating for the client. Interestingly this client had a fairly large IT department, before our solution was deployed, to manage all the earlier desktop applications and now most of them have either retired or re-deployed to other departments.

The product got further matured in our next major implementation where we focused on integrating stores and purchases with dairy production processes. We built configurable templates to handle any type of recipe-based manufacturing process. A large dairy deals with thousands of raw materials and hundreds of finished goods and we were able to handle all these with our flexible core engine and provide key material balancing reports to the production managers.

Our HR module got further enriched as we started handling HR activities at these large government institutions. We built a sophisticated rule engine that handles literally hundreds of complex benefits and deduction types across our clients. An administrator can easily configure new types without the help of any developer (no-code solution!). Complex leave rules and approvals management and integration with biometric devices to enforce various attendance rules impacting payroll were all added in due course of time.

In our endeavor to add more value to this supply chain, we looked closely at the pain points prevalent among various stakeholders of these companies. We saw first-hand the struggles faced by the sales operations teams and customers of these large dairies. The order-to-cash cycle is very short for this highly perishable supply chain. While the volumes are large, margins are razor-thin for the various stakeholders. Any efficiencies that can be brought through technology interventions can add huge value to these companies. The traditional order-taking process through a call centre was very error-prone and often causing excess/short supply leading to a lot of downstream friction. Additionally, poor real-time demand visibility to the production department meant just-in-time production capabilities were not getting leveraged optimally.

With android mobile penetration beginning to pick up in the cities (circa 2014), we saw an opportunity to leverage this trend to streamline the order-taking process. I had worked earlier (circa 2009–2010) developing some hobby mobile apps for Android and iOS. These free apps were moderately successful with more than 500K downloads (with zero marketing efforts!). With this background, I was able to quickly whip up an app (Native Android app + APIs + Backend Services) in a week. We gave it out to a few retailers who instantly loved it since they now had the power of placing orders at their fingertips and access to their entire sales ledger 24×7. Quickly word spread and within a couple of weeks, all retailers got the app installed. The Director (effectively CEO) was thrilled to see all these retailers thanking him for this new initiative and he soon shut down the call center to ensure all orders got booked only through the mobile app. It was a thrilling experience for all of us to see this level of adoption so quickly.

We were directly experiencing the impact of cloud and mobile for the first time in an industry that is typically a holdout/ laggard in the technology adoption cycle. We made the elephant dance!

Check out the next episode of our product journey here.

For the original Medium.com article click Here