On 6 October 2014, Flipkart ran its "Big Billion Day", a one-day sale. By the company's own count it was a huge day: 1.5 million people shopped, and it sold more than ₹600 crore of goods in about 10 hours.
It was also a mess. The site threw errors and at times was unavailable. Deals sold out within minutes, some within seconds. Some confirmed orders were cancelled. And some prices had reverted to their full, non-discounted rates for a few hours.
The next day, co-founders Sachin Bansal and Binny Bansal sent a joint email to customers. It went through what had gone wrong, one heading at a time:
- Website: "We had deployed nearly 5000 servers and had prepared for 20 times the traffic growth - but the volume of traffic at different times of the day was much higher than this."
- Stock-outs: they had lined up "anywhere from hundreds to a few lakh units" of products, but it was "nowhere near the actual demand".
- Cancellations: many people bought the same product at once, so some orders were accepted for items that had sold out seconds earlier.
- Prices: "We realise that this breaks the trust our customers have put in us."
They wrote that Flipkart had "worked very hard over the last seven years to earn your trust", and then: "Yesterday, we failed that trust." They promised to keep scaling up their back-end systems "so that we do a much, much better job next time".
- 1.5 million shoppers in one day
- More than ₹600 crore of sales in about 10 hours
- Nearly 5,000 servers, sized for 20x traffic growth, and still not enough
Why it matters: A flash sale squeezes huge demand into minutes, and inventory, ordering and prices all have to hold up, not just the servers. When it does go wrong, a plain, specific apology from the top keeps more trust than silence.