TN Hari: Is the end-game near in online grocery?

Quick commerce is changing the way consumers as well as existing online grocery players are looking at this space. I am tempted to stick my neck out and share a hypothesis on how the end game in online grocery might look like.

To imagine the future, it would be helpful to look in the rearview mirror and see how this space has evolved, and the strategic choices that won the day. Every winning strategic choice has been, and will be, about deeply understanding the consumer pain-points, and psychology.

The first round of winners (and losers) was decided by the strategic choice that startups in this space made around inventory. The choices available then were: a) an asset light model where the company didn’t hold inventory, but tied up loosely with existing brick and mortar companies in this space to fulfil customer orders, and b) create a supply chain where the startup owned and held inventory.

The first model allowed quick scaling and, quite naturally, was backed by Tiger Global, and Softbank. The fatal flaw in this model was that it lacked a good answer to the question, ‘why does a customer shop for grocery online?’ The customer shopped online to avoid a visit to a physical store, and hence it was important to accurately predict what could (and could not) be delivered. This was not possible unless you owned the inventory.

Therefore, the inventory-owned model became inevitable. The other problem with the asset light model was that the unit economics were very adverse. Besides, a company with this model wasn’t really an online grocery company, but a logistics company. Moreover, the cost of running this last mile logistics would never have been covered by the margin structure that they could negotiate with the brick and mortar supermarkets. Bigbasket won the day because of making the right strategic choice of owning inventory.

The next big realisation was that, besides the planned shopping for grocery two or three times a month, customers had top-up requirements throughout the month in small quantities. This was true especially in the category of fruits and vegetables, and other perishables. The ‘next day delivery’ that Bigbasket was known for, would not work for the top-up purchases. By their very nature, these purchases were impulsive and could not be planned in advance, and nor could they wait for a day. Bigbasket started off with an ‘express delivery’ model to address this need. The ‘express model’ attempted to deliver in 90 minutes.

But this had to be shelved due to poor economics, and probably legacy thinking that came in the way of making this work despite all pointers that this was an important customer need. As a compromise, Bigbasket subsequently created a same-day model with delivery in 5-7 hours. Quite obviously this model had better unit economics than the express model; and for a company that was deeply wedded to unit economics, this was an attractive proposition. There was also a confidence that customers would be accepting of a 5-7 hour delivery time for top-up needs.

This confidence was probably misplaced.

Grocery is the mother of all categories, and bigger than all the other categories combined. So, it was very unlikely that the battle for this category would end so quickly. Soon, online grocery was disrupted again, and another bunch of young entrepreneurs promised an audacious delivery time of 10-19 minutes. This received a mixed response. There were some who said that this put delivery partners in harm’s way, and there was no need to do that because this wasn’t like a medical emergency. The unit economics was tough to pull off, and this was quite obvious to even a casual observer. The assortment that a company committed to deliver in 10-19 minutes could afford to carry was small and was in the range of 2,000-3,000 SKUs, compared to nearly 100,000 SKUs for a full service online grocery company like Bigbasket.

Despite all the scepticism, this bold move redefined the paradigm, and customers now began to evaluate the speed-assortment trade-off with a different yardstick. If 90 percent of their requirements could be met with a limited inventory that could be delivered with speed, then customers seemed open to switching to substitutes that were available rather than being very particular about their choices. The question was, what was the optimum number of SKUs? Was it 3,000, or was it 10,000?

So, the assortment in grocery is likely to boil down to this magic number. If the number settles at say 8,000, then the value of a model with a much larger inventory, but longer delivery times, is likely to diminish. So, grocery is likely to get sharply ‘verticalised’ — i.e. any merchandise outside this optimal assortment of high-frequency purchase items is unlikely to be valued a great deal. Customers are likely to turn to Amazon, or anyone that can think and execute like Amazon, for all low repeat-purchase items (like say general merchandise).

My prediction is that the end-game for online grocery would be about being able to deliver an assortment of roughly 8,000 SKUs (give or a take a few hundred) in 60 minutes.

Companies are likely to approach this end-game from two directions; namely a) from 3,000 SKUs 10-19 min delivery to 8,000 SKUs 60-min delivery, and these will be companies such as Zepto and Dunzo, and b) full service companies like Bigbasket (100,000 SKUs with 5-12 hour delivery) getting to 8,000 SKUs 60-min delivery.

The 10-19 min folks will have no choice but to eventually expand their assortment to 8,000 and increase delivery time to say 45-60 mins, because the current model on a standalone basis is not sustainable; and the additional wait time from 19 minutes to -60 minutes is not a big deal for customers; whereas, the reduction from 5-7 hours to 60 minutes is a significant reduction. Quite obviously the 10-19 minute delivery had just been a foot in the door in a crowded market, and these players will do everything, including expanding the assortment and selling private labels, to help improve the overall economics.

Like in any business model, this model too would evolve based on a set of customer feedback loops that would precipitate the decisions.

TN Hari is author of ‘Pony to Unicorn’, and advisor at The Fundamentum Partnership.

Views are personal and do not represent the stand of this publication.