The last three weeks have been a different sort of hectic for all of us here at Grofers. For starters, we saw demand go off the charts. We are very proud that over the last three weeks we have been able to deliver essential groceries to over 1 million households in 22 cities. However, we keep getting asked why we are not able to service faster, or have family and friends pinging us to figure why their order was delayed or cancelled. We have seen insane acceleration, brought to a screeching halt by the lockdown, and restarted again in the face of an extreme surge in demand. While the situation is unprecedented, it’s time we provide our perspective on what’s happening with online grocery and the kind of decisions we had to take over the last three weeks.
Size of our impact
Organised grocery in India is really small – like really really small. Kiranas are (rightfully) king as they provide customized service at a very local level and constitute 95% of Indian grocery retail. Kiranas have innovated their way through all of this and their resilience has ensured that they still remain the most convenient bedrock of retail. We know this very well because we work with 4000+ small shops across our country to deliver our products to you. While a lot of them were not able to function completely during this lockdown, they were able to get back to work a lot faster, meaning we could get products to our customers again. We all are in this together to serve you in this time of need.
Online grocery is only about 0.2% of the overall retail market. I think at the end of this crisis we will probably reach 0.5%, but that is still an insignificant share. While we will try our hardest, I don’t think we will be able to serve 100 million urban households in the immediate future. We apologize for not being there today for everyone but hopefully, in due course of time, the size and scale of our impact will go up. At over 1 million households in three weeks, we are just starting up.
Order size limits and essential items
As a lot of our customers started ordering home delivery, the outstanding orders at the beginning of the lockdown became difficult to serve. Many of them contained non-essential items or too many items at one go (items per order went up from 12 to 34 meaning ~3x the time taken for packing an order and ~3x the weight to carry). Although the definition of essentials evolved as the lockdown took hold, our first order of business was to make sure we deliver the essential orders first. This was also necessary as we now had to be more efficient – and picking or packing smaller orders meant we could reach more households.
One big change this has meant for online grocery is that we are going to be focussed on really key products. The choice for customers might be limited for some time at Grofers. There is still plenty of supply to go around, but we have made ourselves leaner to be able to serve more customers. This is the reason customers are seeing more limits on quantities they can buy (in addition to avoiding hoarding by unscrupulous elements). We automatically merged multiple orders by the same customers to save time and had to cancel orders that our system perceived as abusive or, in a lot of cases, simply too big to be serviced in the current scenario.
Handling the queue
Everyday our app is seeing over 1.5 million people trying to order. Given the constraints on how many people we can cater to, we are only able to serve 1 out of 8 customers today. The remaining 7 are certainly not happy. We understand the frustration and are working on adding more capacity to make this work for more people. In the meantime, the decision of how to decide who gets a slot vs. who doesn’t was a tough one. We didn’t want to charge for preferred slots or even reserve them for SBC members as a lot of people are in need of grocery delivery. Eventually, that meant we stuck to first come first serve.
The outcome? Server crashes. Below is an image of the increase in checkouts the moment we started taking orders. While it is every company’s dream to see these surges, the reality is not fun. Engineers love scale but hate sudden scale.
We tried to solve this with a token system that was built on the fly, which didn’t really solve the problem. We have now resorted to randomizing slot opening times to balance the queue prioritization and also manage the load on servers. This means slots for the next 3 days now open up throughout the day in all cities. We know it’s a bit more frustrating than knowing a specific time to visit the app but hopefully as our scale up continues, we will be able to do away with this system and just keep accepting orders.
Stabilizing the supply chain
An emergency situation brings its own constraints on its supply chain. We took this challenge as an opportunity to stand up for our customers. Local authorities and police administration helped us in getting all necessary permissions and curfew passes for each delivery rider amidst the lockdown situation. Today, we have resumed operations in 24 cities and our warehouses are working at 70% strength. The unprecedented situation required that we revamp our safety and hygiene practices for item storage and delivery as well as practice social distancing on work. Commuting became a challenge, so we made arrangements for travel and even started bus services for our people. When FMCG brands didn’t have trucks to deliver goods to our warehouses, we had to start reverse picks on our own trucks and get the goods from their warehouses. An additional 2000 people were hired from industries which were deeply impacted by the current crisis such as textile, manufacturing, and services and would have otherwise faced income losses. We now plan to hire 5000 more over the next 2 weeks.
Right thing to do
As our operations restarted slowly after the lockdown announcement, we were faced with a tough choice. We knew the lockdown would be hard for the underprivileged, so we partnered with Feeding India and Give India to collect donations. However, we could also help by getting supplies to these organizations to distribute. With limited means, we could prioritize customers or our NGO partners. We made the hard decision from the outset to reserve daily capacities for shipping groceries for the underprivileged until other alternatives could be developed. This meant, we sometimes sacrificed customer orders to be able to serve these orders. After a few days, the situation eased up and we could go back to serving only customer orders and have now set up an entirely new process to ship orders for the underprivileged and NGO partners.
These were some of the decisions that we faced over the last few weeks. We will share more of our learnings soon. In the meantime, please take the time to donate at grofers.com/donate and keep washing your hands.