Ever since 2010, Analytics has found its way in the Gartner Top trends of the year in some form or another.

This is indeed a great feat and signifies public interest and curiosity in analytics. But unlike other technologies, analytics does not seem to mature and move out from being just a trend!  So what’s happening here?



We have often asked this question about why do companies, especially the new age online marketplaces like Amazon, Ola, OLX etc. delay profitability.  Why should they operate in losses and till when?

Today let’s try to find the answer to this perennial question.





So, the traditional and industrial way of thinking is that the more you create something, the less would be its incremental value for users.  For e.g. a special edition Ferrari not just commands a specific value but people would be willing to pay more if one more unit is produced, the reason being that it is limited. But if the same Ferrari is produced in mass, then the value of an incremental unit may start to decrease after some time.

This is the reason why, the traditional marketing and industrial thoughts revolved around creating exclusivity to increase value, and precisely why the industrial age firms would strive to make profits as early as possible, because otherwise the incremental value of the product and service could decrease and  keep going down as the scale of operation goes up.

Now, the digital or platform way of working is exactly the opposite.   The digital platforms or marketplaces work in such a way that the more people use them, the higher is the value of the platform for an incremental user. 

For e.g. people used OLA in the initial days received less value, because there were limited drivers leading to longer wait times and less connectivity to locations. But today the value to a user is much higher than what the early user received-  there are more drivers, less waiting time and more locations covered.  So the more this platform is used, the higher the value people using it will get.

In the case of OLA  As time passes by the value provided by the platform becomes so huge that it will become difficult for drivers and passengers to abandon Ola and switch.  Hence, today OLA management is in a position to reap or plan to reap profits. They can now increase their commissions.

All platforms should in their initial days, focus on ONE SINGLE attribute and that is ‘growth’, the growth of users and growth of transactions.  Profitability, increased commissions, etc. act as frictions in the expansion of a platform and hence they should   Which is why the endeavour for any platform in the initial days is to reduce friction by way of making the registration process very easy,  attractive incentive schemes for joining and the focus is on increasing usage of the platform.

A few days ago Amazon invested into a 2-year-old autonomous tech development company Aurora that is ed by ex-Google and Tesla executives. Amazon is the largest investor here.

Amazon is also testing a robot called Scout!
It’s a six-wheeled autonomous delivery robot that delivers packages door to door in the neighbourhood.

Why is Amazon doing this?


Before we understand this, let’s take appreciate that Fulfilment is one of the key strategic needs of any commerce marketplace.

We discussed in a few of our earlier episodes of digitalDNA about how e-commerce marketplaces are investing in logistics, supply chain and offline integration.

Amazon, for example, has aggressively been building out its warehouse network to the point that it now has warehouses within 20 miles of nearly 50 per cent of the U.S. population.

But at the same time, fulfilment expenses also bleed the marketplace the most, again, in case of Amazon about USD 34B was spent on fulfilment in 2018 (against a new revenue of USD 232B, that’s about 15% of net revenues gone only in fulfilment expenses)

The fulfilment costs have doubled from 2016 to 2018.



Now given that Amazon’s Prime business is worth over USD100B, and the fact that the consumers have become accustomed to almost free and lightning fast deliveries, Amazon and other com marketplaces do not have a choice but to find alternatives.

McKinsey predicts that autonomous deliveries will slash retailers’ shipping costs by 40%.

At the same time, it’s also important to understand that Amazon is fast expanding into new areas like groceries or prepared food delivery, where the dynamics of fulfilment are diverse.

Look at this chart to understand how Amazon can face competition in future:



One can observe that Amazon will face competition in future from Logistics companies like DHL as well as from contemporaries like Walmart who are partnering with Google’s self-driving cars already for deliveries and of course the upcoming startups in the e-com marketplace space.  (Check the Chart in the above video)

Given this situation, Amazon’s is deploying this 2 pronged strategy to counter all present and future competition:

  1. Network-driven fulfilment and logistics, which are connected by various warehouses and are connected with diverse technologies like drones, robots and autonomous vehicles
    2. Own as much as possible of this fulfilment network. That is why unlike Walmart, Amazon is investing and not just partnering into autonomous vehicle technology.

This will give Amazon a cost advantage as it scales up plus pre-empt new competition, as the entry cost would be too high, as also open up a new area of monetization- it would offer deliveries services to other marketplaces and local retail outlets, retail chains and even offer integration services to logistics companies.

So expect Amazon to be a top logistics company in the coming future apart from a leading commerce marketplace.

We all know about Spotify, the largest music streaming company in the world with about USD30b as valuation. Spotify started in 2007-8 and is available in 80 countries with about 200 Million active monthly users

As per my estimate total size of this market globally is around USD 17-20 B and about 55-65% of this is garnered by 4 players Apple, Spotify, Tencent and Amazon

Spotify and similar other players like Apple Music, Pandora, Gaana or Saavn thrive on largely 2 monetisation models- freemium content for monetisation, charging people for a subscription to provide them with premium content and additional features and secondly charging advertisers for ads to be shown in various ways across the platform.

This monetisation model and the supremacy of Spotify is now challenged by a new player, Tencent music, which was started in 2016, and in just about 2 years it’s valuation has soared to about 28-30B USD. very similar to that of Spotify’s.

Tencent Music Entertainment Group is China’s biggest online music and social platform which runs digital streaming apps such as QQ Music, KuWo and Kugou as well as karaoke app WeSing, which together have about 800 million monthly active users

If you see the chart, you can observe that Apple Music has about 50m paying users, Spotify has about 70million and Tencent music has 30m paying customers. Now the interesting part is that not only this feat by Tencent achieved in just 2 years, but also that only 3.6% of its user base is paying vs about 46% of users of Spotify who pay.

This illustrates the huge velocity of Tencent music and huge untapped potential. Let’s understand how Tencent Music monetises its platform.

The point to be highlighted here is also that the core content (music) is the same across all platforms, hence one needs to be very innovative in thinking and forging a monetisation model. And that’s what Tencent Music is doing.

Spotify, Apple and other players depend upon advertising and premium subscription for monetisation, but Tencent perceives the service as an ecosystem and hence wants to earn from various other modes.

Tencent’s monetisation model is based on 3 challenges:

The core content is same across all platforms- so little differentiation.
People need a social experience – opportunity to create engagement around music
People, especially in China, are sensitive to how much they can pay for a song- Huge diversity in terms of the value of a song/music piece.

70% of the money that TENCENT EARNS (2018, 1st half year, revenues crossed USD 1 Billion) comes from Virtual gifts that listeners give to music stars or singers who sing via karaoke, and from premium karaoke tools, singing with friends, as well as by charging an access fee for live events streamed on the platform. Only 30% of there even comes from premium subscription and ads.

The social engagement around music is what creates differentiation, locks value for users and creates a network effect. Spotify and other large players do not have a major differentiation in terms of content and they are not locking in user value- hence they face a danger of users easily switching and moving away from them in future to a competitive product.

That’s a huge win. As a music marketplace Tencent earns the money from the value and engagement it creates rather than plain advertising (which commoditises the content and engages people with a brand and an external website)

It seems that Tencent should not worry about competition from other music streaming companies, rather it’s real competitors are social platforms, who are eating away it’s users’ time and have similar platform dynamics at work.

Music Streaming Profitability

So if I were to put my money, I would put it on Tencent and its business model. To summarise :

1. Tencent has built an ecosystem of apps, and together they integrate well and empower each other.

2. Tencent is focusing on locking user value by way of creating engagement opportunities around music content and hence increasing the switching cost for users.

3. Tencent has a huge untapped potential to monetise and grow, unlike the competition.
User-generated content

4. The backing of Tencent’s various other products like QQ and We Chat etc.

5. 10 years and 80 million users, and Spotify is still not in profits. Apple does not release the profit numbers but my take is that Apple is also in the same boat.

There is a lot to learn from Tencent for Spotify, Apple, and our own Gaana and Saavn.

Within 20 seconds of the kick start of Alibaba group’s annual 11.11 Global Shopping Festival, TOTAL GMV of gross merchandise value surpassed USD155 millions,  in 2 minutes it crossed USD1 Billion and the day closed with over USD30b of gross merchandise sales value.

Now, Just let that sink in, appreciate the number of users, transactions, customer care calls and emails and maybe hundreds of terabytes of data generated at an unimaginable speed.

Singles day is a clear example of success with data.

Data by itself is of little use, and Big data even more useless, unless you leverage High-performance computing to process this data, use analytics tools and AI to get actionable insights and deliver action and impact.

Alibaba’s TMall, B2C e-commerce marketplace with top brands and over 600m users have an abundance of data on consumer buying habits and sellers’ marketing strategies and decided to put this data to use in 2017.

And as a result, Tmall Innovation Centre (TMIC) was started as dedicated research and development (R&D)

Some of the world’s top brands like P&G, Johnson & Johnson, Mars, Samsung and the likes have partnered with TMIC to create new products and variants for China based on consumer transactions and behavior captured on the platforms.

Let’s look at 2 examples of how TMIC used data to help sellers and brands. Listerine is a well-known brand but it was facing difficulty in cracking the Chinese market, where the users reacted poorly to its antiseptic formulation. 

So, Johnson and Johnson worked with TMIC to learn that Listerine’s target audience—young Chinese women—prefer mouthwash with fruit and floral flavors. As a result, Listerine debuted two new products in China during Alibaba’s 11.11 Global Shopping Festival, or Singles Day.

Similarly, TMIC and MARS chocolates worked together to find that Chinese consumers would be willing to Try chocolate with some spice AND LAUNCHED SNICKER- SPICY IN CHINA.

From its debut in Mid 2017 till mid-2018, sales of snickers spicy surpassed USD1.4m

Not just with the big brands, but Alibaba repeated this success with smaller brands as well.  The watch division of TMall studied consumer behaviour of users and invited young watch designers from the Alibaba ecosystem to create and launch 17 new designs on the singles day.

So, Alibaba group is using data in 3 ways here :

  1. help sellers sell more, both small and large
  2. help brands develop new products
  3. speed up R&D, innovation and go to market.

Now, the story of new retail will be incomplete without Artificial Intelligence or AI.

In the words of Jack Ma DATA AND HPC are the mother and father of ai.

So, AI helps Alibaba ecosystem in 5 major ways:


  1. Recommendation system-  Alibaba has developed a software called “E-commerce Brain,” which uses real-time online data to predict consumer wants, and the models are constantly updated for each individual through AI to take into account purchase history, browsing history and online activities. Now, this recommendation system also works across digital kiosks and magic mirror installed offline in delivering consistent user experience.

2.  Customer care- Any common queries and questions asked are automatically redirected to a computer system called Ali Xiaomi (Ali Assistant). Other than answering frequent questions and delivery status, it can help users find the right products when provided with a text or voice description or even a photo. It is said that during singles day sale- the chatbots handled 95% of the customer service queries. 

3.  Pervasive Personalisation-  what if each brand and seller on a commerce platform could personalise its store as per individual users based on their taste & preferences?  Of course a rise in conversion rates, and that’s what is exactly happening here

4. Supply chain & logistics- from predicting what products will sell more, appropriate pricing strategies, inventory preparation to best route and transportation for deliveries- for online and offline sellers – just about everything a seller could want.

5.  Security – Face recognition coupled with pattern mapping and predictive analytics is already helping secure the marketplace and install more confidence in users.

So, this was the story and strategy of Alibaba which is slated to revolutionise and transform retail as we know, and everything that we have observed and discussed in these 2 episodes  point to just one thing- Alibaba is focused as hell on improving and unifying consumer experience and helping sellers be more innovative, efficient and profitable.

That’s what we call as NEW RETAIL

With over 1 trillion USD spent online  which is about half of the global commerce sales, China is at the forefront of commerce and with about 60% market share, Who can be in better position than Alibaba to forge a strategy for the future of retail, which is aptly christened as NEW RETAIL by Jack Ma.

Now we shall discuss New Retail is two parts.  In this episode, we shall discuss the O2O (online to offline) and in the next episode, we shall discuss how Data is central to the New Retail Strategy.

So, Future of retail is not a question of online or offline, it will be a matter of integrating consumer touch points for seamless experience in the most efficient way.  And that is what ALIBABA is attempting in its new endeavour of the New Retail.  Through New Retail, Alibaba is trying to revolutionise 4 areas of retail as of now:

  1. The supermarket
  2. The Malls
  3. Mom and Pop Stores  and
  4. Auto Retail

Before we begin let’s understand that Alibaba group has 3 major pillars from a retail perspective-  Taobao- a mass retail marketplace (more like a micro business to consumers or c2c), Alibaba- connecting Chinese exporters with the world (b2b)  and TMall -a  curated trusted brands marketplace (more b2c)

So, let’s come back and talk about the 1 area of retail to be transformed and that is the supermarket-

The supermarket transformation is christened as HEMA stores.  Shopping at Hema is a smartphone-powered experience—you can do it from home or in the store. When you’re in the store, you’re able to scan a barcode with your phone to get product information. Payment is also cashless, done through the Alipay platform embedded in the Hema app.

The fast 30 minute delivery for those who live within three kilometres of the market,  is Hema’s key USP. Each store serves as its own warehouse and logistics centre that collects, fulfils, and delivers customer orders as fast as they come in, online or offline.

2.  The Car Retail: 

china is the world’s largest Car market, and the retail part of it can really become more convenient and effective. Alibaba is creating auto vending machines across the country.  It’s a model that lets customers browse makes and models inside their app, choose a car they want to test drive, pick it up from an unmanned vending machine, and drive it for up to three days. After experiencing the car in a no-pressure situation, they can make an appointment to visit a dealer when they’re ready to buy.

3.  Mom and Pop Convenience Stores- The local kirana shops, as we call them in India are generally plagued with inventory related insights.  There is absolutely no insight into demand prediction, loyalty, up-marketing and inventory management, and RoI. Alibaba launched a program that is called Ling Shou Tong or LST,  which digitises the inventory management of each store and integrates these businesses into a central warehousing and logistics system. It also provided them with an analytics platform that anticipates customer preferences and lets proprietors know what they need to order, how much, and when. Ling Shou Tong also modernized the insides and signage of the stores. Early results point to more efficient operations.

4.  The Malls-  These are the giants that attract a huge volume of footfalls and window-shoppers, but in recent times are witnessing lower volume of shoppers.  Alibaba has invested in Malls and the stores in these malls are equipped with “Virtual Shelves,”

As per Google 43% of consumers exit a store without buying in frustration because they did not find what they needed, or lack of information about what they needed.

So with virtual shelves, if you don’t find your size or colour in stock, you can still select the product you want on a screen, punch in your size, colour, and other specs, scan with your app, and have exactly what you want to be delivered directly to your home.

Even the changing room or washrooms of malls can be a New Retail experience. Step into the ladies’ room, and while you’re waiting, check out the “Magic Mirror” on the wall to experiment, virtually, with new makeup colours. And if you Like what you see? You can buy it from the nearest vending machine

So in the next episode let’s look at the DATA part of the New Retail strategy of Alibaba. Thank you and see you soon.

You know Amazon as the dominating force in e-commerce but did you know that in 2018 Amazon garnered up more than USD 4b from advertising revenues?

By my estimate, this would be about 50% of Instagram ad revenues.

In fact, Amazon already occupies the 3rd space in the US from an ad revenue share perspective after Google and Facebook at 4.2%


While the figures do not portray a threat to Google as of now, let’s look at a few key trends and the context, indicating an undercurrent of a new war of sorts between Google and Amazon:

  1. If you look at Facebook’s growth, it grew in a similar manner with USD4-5B in ad revenues in 2012 to USD 40b in 2017.


2.  The trends suggest that more and more people are now beginning their search on Amazon when they intend to buy something, rather than on a search engine. In fact, the no. of people starting their search on Google with purchase intent has been declining over the past years and increasing on Amazon.

(Another research shows how Google’s share will dip between 2018 and 2020, while Amazon’s will rise in the same period, and Facebook may remain consistent- implying that Amazon will largely eat into Google’s share)





3. The third trend points to the fact that Google’s ad revenue market share is already slated to dip from 41% to 37% this year, this is despite the fact the overall advertising pie is increasing.



So from an advertising point of view, Amazon is already a threat to Google largely as also to Facebook and poised to be a very strong third channel for advertisers.

Now let’s look at this competition from a different angle, and that is about how Google is trying to become a threat to Amazon with its Shopping endeavour

Google shopping marketplace is now launched in India and overall it’s strategies are very interesting to observe.

So, we know that about 55% of shoppers will begin their search on Amazon, what is more, interesting is 70% of the shooters will want to cross-check the price on other shopping platforms too before finally buying. Google is trying to tap this opportunity through its aggregation platform where it displays products and prices of products from different marketplaces like Amazon, Flipkart etc.

So it’s a great start to get the attention of shoppers and engaging them even if they start on Amazon and then potentially divert traffic to Amazon’s competition. or fulfil the purchase via Google ’s shopping actions, in which Google gives users access to the universal shopping cart which they can fill with products while searching from any marketplace and then pay via Google pay.


Google is hitting on the convenience USP of Amazon. Users buy from Amazon not because its the cheapest, but because of its easy and intuitive buying process. Google wants to make this process even quicker and easier.

Only about 5% of sellers on Amazon make more than USD 1LAC AS annual profits. The long tail of smaller merchants may not find Amazon as a very attractive option. Google will target these small merchants to allow them better visibility on its platform


So, If you see, the war between the 2 giants is at various levels, from Voice assistants to advertising to search to e-commerce. In fact, it’s a platform or ecosystem dominance war and it would an interesting one to observe how the competition plays up.

Most of the new age businesses like Uber, Facebook or Airbnb, are launched as platform business models, wherein typically they connect 2 sides of a platform, like buyers and sellers, or riders and drivers and so on.

The challenge is to understand which side to grow first.

It’s a classic chicken and egg situation. 

For e.g. in the case of Uber, Its platform is two-sided, connecting people who need rides with people who have rides to offer.

In general, platform models work on getting the supply first, for e.g. In Airbnb if there is no supply, people will not come. However, if the platform has listings available, then there is value for people to come to the site. After this- the demand will accelerate supply and supply will accelerate demand.

In the case of Uber, it’s quite evident that the supply side is limited, and that’s what Uber should attempt to grow.



But how do you get the supply side going?

Here are a few ways to build traffic and traction for your platform in its initial days:

  1. Piggyback  When Airbnb wanted to get listings on their site from homeowners. The founders of Airbnb started thinking about where would people go if there was no Airbnb?  The answer was clear- Craigslist!  They found volumes of people who are renting out their apartments via craigslist. The property owners had nothing to lose by listing in Airbnb, and suddenly Airbnb was blossoming with huge volumes of listings.  The supply side is sorted.


2.   Choose your suppliers carefully: In order to get a huge volume of suppliers or listings, platforms should          very carefully choose their suppliers. They should invest in moderation and filtering out the bad eggs.                  Supply will build the demand so platforms should have zero tolerance here. 

       Etsy, for example,  went on to craft fairs across the country to identify and onboard the best vendors


3.  Focus on user experience:  Help your supply side sell better and look better:   Just getting the listings was       not enough, those listings needed to look very attractive and provide a great user experience for dead side         well.  Airbnb hired professional photographers to go to property owners’ homes to take inviting pictures.            This made the site much more attractive than the competition.

   Uber also started with ‘Black’ cars first in the US, wherein they accessed a pool of professional drivers and         enrolled them. 

   This created a professional experience for riders and Uber instantly had a huge fleet                   running  on       the road. 

These great user experience spread via word of mouth and a the platforms start receiving referral customers and repeat customers.

Remember-  If you are not getting repeat users, you will not get referrals.  That’s why building a great supply side with great user experience is the no.1 task for any platform-driven business.


4.  Find strategic gaps  E.g. Uber did not launch in all cities on day 1. They identified which cities show the widest gap in terms of supply and demand of taxis. Then they identified which time bands show this wide gaps, and launched in specific cities and specific time gaps only.


Airbnb followed a similar strategy with its rollout, launching in Denver in 2008 to coincide with the lack of hotel space during the Democratic National Convention and adding new cities at times when they had major conventions or other events.

Launching in situations of high demand and low supply also helps startups acquire the right type of customers—the early adopters

So, these are a few ways and best practices that few of the top platform companies adopted to get their first thousand or few thousand customers. 

I had this discussion sometime back with a few colleagues and I thought of sharing the key thoughts from there

Most of know a bit about blockchain and we use ola or uber regularly as well.

And interestingly cab hailing service is one of the industries that have a huge potential to get disrupted

Not just Uber and Ola but companies like Oyo or  Urban Clap etc. also work on a platform model of business, wherein they facilitate an exchange of value between 2 parties. In the case of cab-hailing companies- rider and driver become the 2 parties.

While doing this, they create and locks-in immense value for both parties in terms of driver and rider reputation, safe payments, easiness of finding cabs and customers, remembering your music playlist and so on.

This is why the riders and drivers eventually become dependent on the platform and the platform charges a premium through surge pricing, transaction cuts etc. means. which make your ride more expensive and less profitable for drivers.

We will understand Platform businesses in detail in one of the coming episodes.

Now, imagine if riders were to find and connect with drivers directly, without any platform in the middle to make the match and yet enjoy all or most valuable benefits that a platform gives them.

Of course, the rides will be cheaper and earnings for drivers would be better. There would be more privacy and data accuracy as well.

How is this possible?  Blockchain OR A decentralised platform can makes this possible

So, imagine that we have a suite of distributed applications all connected to a decentralised platform, let’s name it SMART.

All a driver has to do simply is access this suite of applications: one to verify insurance, one for background checks, one for licensing, etc. These apps are provided by certified vendors. Once all these boxes are checked, the driver appears on the network as “available.”

ride-hailing process decentralisedSource: DACSEE

Riders need a coin wallet. The rider appears instantly on the other end of the SMART network and is matched with a driver automatically.

There are no commissions or premium charges, but just a one-to-one transaction. Driver and rider both are happy.

This is not fantasy when we talk about Bitcoin, there is no one in charge there, and it works in a similar way. So there is a good use case for blockchain already.

Having said that, we still may have to give a very small fraction of each transaction to a few entities for marketing, validation and support activities, but I guess this will still be a very efficient system than what we have now.

This would also be interesting because I see that instead of one or two big platforms we may see many such platforms and this would usher in a different type of platform dynamics and competition at play.

So looking forward to meeting you again with a topic that has an impact on our businesses and lives. Till then keep tuned in.

I have been asked this question many times, and I thought of bringing this up today and document through our DigitalDNA video series.

I have been asked this question many times, and I thought of bringing this up today and document through our DigitalDNA video series.

First and foremost, my personal belief is that -there are other more important attributes than time to post, which social media practitioners and content marketers should address.

Let’s discuss those attributes and metrics in one of our upcoming shows.

So, there are various studies already available on this subject, but the fact of the matter is that one size, or in this case one time will not suit all.

Different industries and their consumers will be active at different time bands, so there is no single generic answer to this and I am not going to provide that as well.

So, before we advance ahead, let’s understand by what we mean by the best time to post:

We essentially mean that the best time to post is the time :

  1. When the most number of people are online
  2. When there is the least competition
  3. When there is most engagement possibility

Again, personally, I don’t believe in points 1 & 2, for one because they are contradictory and second that they are not in our control.

Let’s understand how:

Facebook insight gives you a great way to find out the time band when most of your fans are online


1.this does not imply that all of them will see our content  and

2. this implies that most of the competition would also be active around this time band, and competition does not only include our direct competition but also all those brands which target this audience.

So, the real test of what time is best to post, is to check engagement level by time band.

As you can see, the engagement is low at 11AM (which is the peak time for Reach), while engagement is maximised at a non real peak time, which is at 1PM.


Facebook Post Engagement By Hour

 The bottom line is clear- look at engagement time bands for posting your content.

Posts navigation