Agarwal talks about innovations from the Amazon stable, of which Seller Flex, Amazon Pay and I Have Space are the result of “listening to customers" and “inventing to address their needs," while other more transformative ones, such as Echo and Firestick, are the results of “imagining the impossible." Edited excerpts from an interview:
It has been six years since Amazon formally entered India and you have been helming its operations ever since. If you were to decode your experience into the good, bad, and the ugly, what would that be? How do you think about the consumer of the future?
One of the things that we firmly believe is, instead of trying to predict how the future would be different, we gain a lot of leverage by trying to find how the future would remain the same with time and with geographies. And at Amazon, we had the conviction before we came to India that those three pillars of experience are primarily large selection, competitive prices, and fast and reliable delivery. We haven’t really found a geography or a customer set or even across the last 20 plus years that I’ve been at Amazon, that customers have said I wish Amazon gave me access to fewer products, at higher prices, and slower delivery. The good part is that India is no different from the rest of the world in these three important dimensions.
Customers in India, like customers globally, care about selection, low price, and fast and reliable delivery as much as any other customer in the world does, and that was great for us because all the investments we have made over several years can automatically be used to serve the customer in India.
At the same time, I can guarantee you that the consumer of the future, just like the consumer of today, would be divinely discontent. They would be very unhappy, they would always ask for better experience, they would push the boundaries in terms of how technology should be serving them. In terms of India, I think the consumer would enjoy level playing field. Mobile internet in India is a tremendous leveller in a country with significant heterogeneity in pretty much every dimension. If you think about income levels, social structures, languages, cultures, the mobile internet really levels the playing field both for sellers and consumers.
With Go stores opening in cities like San Francisco plus Amazon’s purchase of Whole Foods last year, is the future looking more hybrid for the world’s largest retailer?
In some cases, we have tried to use the offline convenience to serve an online customer better.
Amazon Easy is a good example. They trust the neighbourhood store person more than just shopping online and we are leveraging the trust of that person to help the customer use the online convenience of infinite selection but bring that selection inside the store through a terminal as an example so they can find what they’re looking for, pay offline where the trust is higher and get the online convenience of delivery at doorstep.
You can actually take your Amazon app inside certain stores, scan certain QR codes and get access to Amazon customer reviews and Amazon images and Amazon videos to enhance your shopping experience and buy the product offline. One of the biggest learnings I have had in India is that the incumbent ecosystem that exists in the country’s neighbourhood for decades is so well-positioned in serving customers, that you are better off partnering with them than to go against them.
What is the culture of innovation at Amazon? Can you give us examples of some key big ideas that you are working on, especially locally?
I think of three principles to sum up our culture of innovation at Amazon—obsess on customers, invent on their behalf, and think long term.
Example of things that we’re doing in India by listening to customers are innovations such as Seller Flex. We were listening to sellers as customers, and they also told us that they are businesses that have been operating for generations, have low cost real estate where they keep their own warehouses in and it would be great for them to figure out how they could further lower their cost of operations by leveraging them.
So, it gave us this idea that why not take our fulfilment software to their premises instead of asking to keep their selection in our premises. And that was the whole process of Seller Flex and that innovation has become such a powerful driver of fast and reliable delivery in India that we have actually exported it to other countries around the world. So, that’s an example of listening to customers.
Big needle-movers require you to envision and imagine the impossible that even customers cannot talk to you about and that is this word that Jeff uses in his shareholder letter which is ‘wandering’. Look at AWS and Echo.
Customers didn’t tell us they need a cylindrical device in their kitchen that they can talk to. There are places that we are wandering and we like it. Another example of wandering is I Have Space.
And in case of India, we are going a step further.
We are working very hard because the landscape is so complex in language. You will see more people use two languages in the same statement. Natural language understanding (and) speech recognition are significantly more complex, which require deep learning to figure out how to serve customers better.
Today, Echo in India can do lot of complex stuff, you can make two languages work, to say you are running out of atta. You don’t have to say wheat flour.
We have put artificial intelligence now in millions of households in India through your Fire Stick that you can talk to, look at your channels and play a movie.
The desire for shopping experiences will intensify. How are you looking at leveraging technologies like Augmented Reality and Virtual Reality for enhancing such experiences?
A lot of things around you are going to change, from technology to how people use and regulations and so on. Data is increasing and hence ability to use data to have computers decipher algorithms using Machine Learning, deep learning, Artificial Intelligence is increasing at a rapid scale. So, the ways in which you could create extremely personalized experiences in a very heterogeneous customer landscape at massive scale are becoming far easier and democratized in a way that anyone can do it.
I would point to Echo as a great example, I don’t know if you read Jeff’s most recent shareholder letter, he talks about listening and he talks about wandering. big inventions are things that maybe customers also don’t know to ask. Customers didn't’ tell us they need a cylindrical device in their kitchen that they can talk to. These are things that came about through, as Jeff calls, wandering.
We are seeing advanced markets going for click-and-collect models, examples of which include ASOS and Boots, eBay and Argos. What do you think of this model as you grow your franchise here?
We have, for example, a partnership with Shoppers Stop and one of the key things that we have done with them over the last year is to ensure that customers have visibility of Shoppers Stop selection on the Amazon App. Why just pick up Shoppers stop order there, why not Amazon order while I happen to be there. You don't have to go, it comes to your door in 2 hours and you can try out many things in that and while you're shopping for grocery in person, you can pick up Amazon products, and those are all valid experiments worth doing in much the same way.
Experts posit that the consumer journey in 2026 will increasingly look like a pretzel that twists, turns and loops back on itself. Consumers can start and end their shopping experiences on a mobile platform, in store or online. It will be a fluid movement and it will be even harder for retailers to keep up with or predict because it will include a growing number of devices and touch-points. What does Amazon think of this and is it working in that direction? It is said that Artificial Intelligence will democratize everything?
About 20-plus years ago when we launched the website, it was a very basic algorithm; no AI did data correlation to predict what you might like based on past purchases. Now what machine learning allows you to do, instead of trying to foresee the algorithm and program your computers to execute the algorithm, you can serve your computer a lot of data of observations and ask the computer to decipher the algorithm.
And the best technologies are ones that are mostly invisible and can work as normal utility. Echo is a great example. Echo is inspired by Star Trek in some sense.
And it is all about Machine Learning and cloud in the back end. But, the convenience with which you can talk to Alexa in Echo and just have a conversation is the kind of experience that we are aiming for.
In case of India, we are going step further. We are working very hard because the landscape is so complex in language. You will see more people use two languages in the same statement. Two languages combined in one. Hacks in different vernaculars, colloquial terms. It makes the understanding extremely difficult. Natural language understanding, speech recognition is significantly more complex which requires you even more ground troop and deep learning to figure out how to serve customers better.
Today, Echo in India can do lot of complex stuff, you can ask it to buy garam masala for you, you can mix two language works to say you are running out of atta. You don’t have to say wheat flour. And lot of these interesting things seem very natural, but doing that requires Machine Learning and technology behind it.
And just looking at our website, every bit of your experience has been guided and made better by AI in some ways. Personalizations on Amazon are far more complex. Simple things like predicting your delivery because you're shopping with us reduce steps in the checkout pipeline, prioritizing your payment methods based on the likelihood of success so that your payment checkout is likely to be efficient, trying to get sellers' catalogues to be data-quality perfect in some ways, so looking at images and deciphering the colour from it so we can remove defects in the catalogue. You don’t buy something which looks pink but title says blue, so we would just correct the title to say pink because that's what the product is. So simple things, search, prioritize products that might be closer to you, etc, to looking at ways that make packaging more efficient or suggesting different packaging last minute so less air is shipped, and defects are avoided. I can go on and on. Every aspect of our life.
We have put Artificial Intelligence now in millions of households in India through your Fire Stick that you can talk to, look at your channels and play a movie.
India is a key market for Amazon and you have left no stone unturned to prove it be it giving capital allocation signals to competitors from time to time or upping your game here. Now RIL is also entering the game. All of you have deep pockets of capital, huge local and global expertise, how does one differentiate and claim leadership especially when the end consumer is so fickle-minded and there are many to throw discounts and loyalty cash -backs to woo them?
The customers can compare and hence the most important outcome that we want is that in a completely transparent world, we want customers to choose us. And we take it as an article of faith that if we are disproportionately better on the things that matter most to the customers, which is selection, pricing and fast and reliable delivery and convenience in general on a consistent basis, they will choose us.
Can you give me some indicators on your India performance till now?
We don’t breakout by country but I can give you some broad indicators of growth here, When we look at selection, we are the largest marketplace in India with more than half a million sellers, small and medium enterprises in many cases offering more than 200 million products to customers--far the largest selection and not just segment of selection.
Similarly for fast and reliable delivery we know that we have the largest selection of products available for same day, one day, two day, two-hour delivery in terms of pin codes and number of products.
Additionally, the Prime members in India are the most engaged Prime members anywhere in the world. Which is a huge vote of confidence in the kind of experience they are getting and the differentiator that Prime has become in India. And what this means is the probability of a Prime member to engage with shopping, video, music etc., all these benefits on a monthly basis is significantly higher than any geography in the world including the US. So we are very excited by the loyalty we are seeing from members. They are shopping more, they are engaging more, they are spending more hours watching TV and listening to music. All of that is great and we love that and we're continuing to grow that.
How does Amazon think about owning the "future of e-commerce" narrative in India ?
When I think about India, I am very encouraged by two things. Our ability to offer the largest selection to customers and do it in a way so that they can get great value and get it delivered conveniently--especially fast and reliably--is completely aligned with the idea of a Digital India. And then when I look at Make in India, our offering like Global Selling--which, by the way, has become a very significant part of the business and Indian manufacturers are now selling to Amazon global marketplaces and we are essentially taking Make in India global in a very scalable manner--is very aligned with the priority. I feel very heartened that our business priorities are very much aligned with the country's priorities, and India will see over time that Amazon is actually making Indian economic growth faster. And not just because Amazon directly provides access to selection but neighbourhood stores in their localities started acting as Amazon logistics partners and built a livelihood and brought selection not possible to bring it by itself.