Learning from the Network meeting of the Peter Pribilla Foundation

I had the privilege of attending the 10th networking meeting of the Peter Pribilla Foundation on the 5-6 May, 2016 at two wonderful villas around Rome, Italy. Thanks Kathrin Möslein for inviting me again to participate in this wonderful network meetings in picturesque villas. This is not intended to be a minutes of the meetings, but my own notes and learning.

Manfred Broy’s keynote on Digital transformation

Digital transformation today is being driven by multiple forces: technology push, infrastructure maturity, market pulls, and startups that can leverage these business model opportunities. As markets, technology, and competence come together to create new business models, the economy is flooded with startups that could disrupt our lives in more ways than what we can imagine.

The talk brought to the fore three observations in my discussion.

  1. Software is eating our lives

As the digital transformation evolves driving on increased computing power, trnasmission power (bandwidth), and programming; Governments are struggling to regulate these business models. For instance, Skype as a software disrupted the international telecommunication industry that relied on massive investments in hardware at the backend and the consumer end. Blockchain has created an entire monetary system with no involvement/ interference of the State.

  1. From Internet-of-things to Internet-of-systems

More and more devices are being connected to the internet, and more and more data is being collected about every part of our lives. The evolution of the Internet has followed the linear path from (a) http or internet 1.0 that connected computers in a network, to (b) web 2.0 that allowed for interactive content in the form of search and social media, to (c) a semantic web 3.0 that allows for semantic search, including images, videos and other references, to (d) the mobile internet, that focuses on the App Economy – hyperlocal and mass-customized content, to (e) integration of IoT devices and servitization applications that lead the Interactive Industry or what is called Industry 4.0.

  1. Moral questions on how these data is used

As more and more data is being collected and collated by corporations, that are mostly monopolies in their markets, questions remain on the nature of consumer choice on what and how their personal data is being used, definition of trust and transparency of these data banks, and how these changes are affecting our professional, personal, and social lives.

Four sub-groups deliberated on actions, competencies, infrastructure, and promises around digital transformation.

Peter McKiernan and Anne Huff summarized the discussion and left us thinking on two axes.

  1. Has all this digital transformation driven us towards so much personalization and customisation that we excelled in marketing to a segment of one; while we have ended up destroying the social processes that form the basis of creating vibrant communities?
  2. With all these investments in digital transformation, what social problems are we solving in the developed and emerging economies? What are our contributions to sustainable management of our ecological environment, alleviate poverty, and manage active and forced migration of people across national and continental borders? What can we contribute to the improvement of human development, fostering inclusive growth, and evolve meaningful networks of social and economic competencies?

Albert Heuberger talked about the need to integrate research on hardware, software, and open problems. He talked about the various projects that Fraunhofer IIS was working in collaboration with the FAU Erlangen-Nuremberg and the Bavarian Government. His view of the future was to sustain research on

  1. Power consumption economics, including battery technology, to power smart devices that need to be ‘always on’.
  2. Devices, software, and problems that help improve mobility through increasing the digital range of smart devices.
  3. Integration of data from intrusive and non-intrusive biological data like glucose levels, fatigue)
  4. Consumer applications of hyperlocal environmental data, like pollution parameters (COx and NOx)
  5. Long range imaging, including gesture control
  6. 3D displays for mobile phones (VR apps for end consumers)

Helmut Schönenberger and Dominic Böhler from the UnternehmerTUM briefed us about the TechTalents program where they have batches of students and entrepreneurs being mentored by experienced mentors.

Peter McKiernan summarized the two talks about the need for engaged scholarship in the context of business research losing practical relevance. I could summarize the day’s discussion and thoughts as an interaction of two triads.


Our second day began with Mitchell Tseng talking about his rich experience of how the world has evolved in his talk on leveraging individual expertise in the context of global cooperation. As the world moves from optimizing supply chains to global value chains, we need to build three related capabilities

  1. Actively manage the shift from reducing waste, focus on core competence, and being responsive to customer needs to increasing the customer willingness-to-pay, focus on the value communication and delivery, and be responsive to changes in customer value perceptions over time.
  2. In a world dominated by network effects, value providers could realize value from even customer indifference. The old chinese proverb says, “the wool grows on dogs, and the pigs pay for it”.
  3. Rapid prototyping in a globalized world requires organizations to embed the product concept into the prototype and be able to test it across different parts of the value chain and in different cultures.

Hans Koller commented that even traditional businesses like aviation (free flights for passengers paid for by advertisements/ shopping), renewable energy (freebies for consumers who allow for installation of solar panels on their rooftops), and healthcare (providing free healthcare advise/ services in exchange for data collected from patients through embedded devices) are embracing two-sided markets. He also added that such rapid prototyping may leverage modularity (as propounded by Prof. Charles Baldwin) in product design and development. Building modularity across global products and value chains requires well-defined international standards for interfaces.

Peter McKiernan commented that research on value creation from the eyes of the consumers (perceived value) could learn a lot from the research on cognitive psychology literature. The definition of business value creation has over the years evolved from (a) the traditional industrial economics SCP paradigm to (b) Porter’s industry attractiveness frameworks to (c) mass customization and value creation to (d) the experience economy of the 21st century.

Members and fellows of the Peter Pribilla Stiftung (PPS) shared their wonderful work, research, and experiences. Unfortunately, the notes are not part of this document.

The afternoon was centered around two sub-groups working on (a) how the research group could work together in joint projects and (b) designing formats for digital transformation. It was discussed that the network should be largely expanded to include people from outside Germany, maybe leveraging each others’ personal networks. The need to collaborate with each other in applying for joint projects from organizations like the EU was emphasized. The group on designing formats elaborated on the need for an agency that could act as a platform that would evangalize, educate, and build strong networks of organisations that enable digital transformation with those that need their services like the Government, Universities, Schools, non-proifts, and corporations.

The networking meeting ended with summaries by Anne Huff, Frank Piller, and Ralf Reichwald.

We have come a long way from when we started in the last ten meetings. Too much of our discussion was centered around white, middle-class caucasian world. We need to expand our focus to the globalized world that includes a lot of problem. The second problem is that we have been largely academic-centric. We are the product of a system that pushes us to be more theoretical, abstract, and less practical and working with the firms. It is imperative that we move more towards pragmatic application of our energies to solve the big bad world’s problems.

Dynamic capabilities is about how organization’s change and evolve over time. We need to adopt the same approach and ask ourselves, look at our own unconscious biases, shift from the technology level of analysis to the more micro-social levels, include people from more varied disciplines like Psychology and Sociology to educate us.

We have learnt a lot about technology, digital transformation, and new business models. We are so proud that we heard from our PPS Fellows. We have over 50 fellows right now working, and it is heartening to see them do so well in their research and careers.

Thanks to Claudia Lehmann and her team for the wonderful organization.

Comments, observations, edits, and additions welcome.



Building a platform business is hard work, not for lazy people: A response to Prof. Ajay Shah’s column in the Business Standard

I read with interest what Professor Ajay Shah had to say about young men and women entrepreneurs of today wanting to become rich quick, with dreams of laziness in the Business Standard (see here). This note is a response to his observations/ allusions that businesses that run on network effects (a) are not-so-innovative, (b) operate in monopolies, and (c) are built around inferior products.

Building a platform is hard work, not for lazy people

Let me begin with the title – lazy businesses. The implication of laziness is that while there is opportunity and capability to do the hard work, these businesses (and by implication, its founders) are unwilling to work hard. I disagree to the notion that anything develops fast is not hard work. The implication that a business that grows slow is “steady” and the one that grows fast is cutting corners. True capitalism favours entrepreneurs who chase and capitalise on big opportunities, and that too pretty fast. Building network effects is not as easy as he alludes. He quotes the example of Google monopolising cloud-based email due to the network effects it has generated. Google was not the first entrant in the cloud-based email space, there were two large competitors operating when it entered – Hotmail and Yahoo Mail. Gmail entered with a disruption – it offered almost unlimited storage on the cloud, and two, it began with invitation only. It took a while for Gmail before it became open for signup, but given that innovative positioning of “never having to delete your email”, it was quick enough. Behind this innovation at the customer end was the hard work, the painstaking task of building server farms across the world with sufficient security and redundancy built into them. This is exactly the hard work Prof. Shah talks about, innovating around products – the Google innovation was riding on the falling storage costs and leveraging the power of global network connectivity to build a network of server farms, thus driving costs down. The fact that Gmail was able to unseat Hotmail and Yahoo Mail from their leadership positions is sufficient evidence that competition is working, and capitalism is safe too.

Platforms are innovative

One of the key tenets of capitalism is that factor endowments (like capital) flow freely from inefficient uses to the most efficient uses. The fact that the venture capital market is amply fragmented is the first signal that capitalism is working there. Let us turn to the platform business firms that seek these capital resources. As capitalism would have it, money should only flow to the most efficient uses of capital – ask any entrepreneur about raising money, and you would hear enough of how difficult it has been. Raising money has never been so difficult, as each of the business models have been unique. Yes, there have been replication business models that get funded, like I want to be build the Uber of Indian hospitals, but they are few and far between. Each of the business models that are flush with funds from the venture capital firms and angel investors are indeed innovative. May not be in the traditional sense of the product innovation like Google’s server farms, but a lot of them offer unparalleled service innovations. Take the example of Quickr, the C2C used goods marketplace. Though such used goods marketplaces had existed in the past, Quickr has managed to bridge the information asymetry between buyers and sellers in a variety of ways (photographs of products, contact details of sellers, premium services, and enabling within-platform communication through QuickrChat) as an insurance against the platform becoming a “market for lemons.” These innovations have not happened in one day, it has taken them years of competing with similar and local marketplaces and keenly listening to their customers on both sides – buyers and sellers.

Not all platform businesses operate in winner-takes-all markets

Prof. Shah alludes to the suggestion that most, if not all, platforms create and operate as monopolies, once they reach a threshold of network effects. Research in economics shows that there are three conditions for a platform market to become a winner-takes-all market – network effects are strong and positive, multi-homing costs are high for the users, and there are no special preferences for users. He has clearly defined the network effects in his article, and I would skip that part. Let me turn to multi-homing costs. Unlike switching costs which measure user costs of switching from one competing product/ brand to another; multi-homing costs measure the user costs of staying affiliated to multiple product/ brands. A good example of multi-homing costs is the number of emails accounts a user can efficiently own and operate. Even though most of cloud-based email is free to use, and a user can create any number of email ids for herself, what restricts her choices to a few is the costs of logging in to each email id, and making sure you do not miss out on important communication. This multi-homing costs ensure that the market has one social networking site, where people connect with friends, family, co-workers, as well as their business partners. However, not all markets have high multi-homing costs. Users (bargain hunters) do shop on multiple ecommerce sites and maintain their login/ passwords for each of these sites. The third condition for a market to demonstrate winner-takes-all economies is the absence of special preferences amongst the users. In the peer-to-peer networking space, where Facebook dominates the social networking market, professional networking (finding jobs and customers for one’s skills as a special need) has another player, LinkedIn. Passive job seekers would populate LinkedIn, while active job seekers would register with one of the many job sites like Naukri.com. The point I want to drive home here is that having network effects by itself does not guarantee a winner-takes-all economy. Firms expend time and effort in building multi-homing costs and enveloping any special needs to create a winner-takes-all market.

Successful platforms have a superior product/ service core

Though network effects make switching costs high, the history of platform business evolution is strewn with a lot of products/ services that have fallen by the wayside due to poor quality of its core product/ service. We did talk about Hotmail and Yahoo Mail, that did not innovate at the right time and lost out to Gmail. On the other hand, Friendster and MySpace failed due to Facebook’s superior quality and constant innovation. Google+ with the backing of the Internet giant, is an also ran in the peer-to-peer social networking space. Yes, switching costs exists and are non-zero, but given the right kind of strategy adopted by the challenger, that is apart from the superior product/ service, users can, and will shift.

Network effects are hard to build

Prof. Shah’s piece asserts that network effects are easy to build and can be done quickly too. Building cross-side network effects are difficult. How would Prof. Shah like to be the first contributor of a new newspaper, not as established as the Business Standard? He writes a column for the BS because of the existence of network effects – he knows that his columns would be read by the “right audience.” Traditional businesses like newspapers have long known to subsidize one side of its user base, its readers, while making money from advertisers (and in some cases, even benefactors and sponsors). So is the case with the media industry. This is a classic “chicken-ane-egg” problem that network industries have to resolve. There are many ways to solve them, and subsidising one side is just one of them. For instance, Practo has invested heavily in building its practice management software, Practo Ray for its clinics side of the business, so that it could build cross-side network effects. Now that the clinics use Practo Ray, Practo can afford to subsidise patients discovering doctors/ clinics through Practo.com. Tough, hard work buidling and selling practice management solutions to clinics, before the subsidising began. Subsidising one side of users to build network effects is not in itself any bad, but such subsidising should not be at the cost of overall economic well-being. Founders/ VC investors (shareholders) and managers make money because the customers, at least one side of the platform, are willing to pay. And they are willing to pay in return for the value they receive.

In sum, building a platform business with network effects is not lazy work, it takes a lot of patience, investments, and creative solutions to succeed. Yes, they are unlike traditional “pipeline” businesses where value flows from one direction to another linearly. They are different, and in some kinds of ways, fun. They have multiple sides of customers to deal with, and are on the toes all the time to keep the fine balance intact. These are exciting times when traditional pipeline businesses compete with platform businesses.

Comments welcome.

Measuring E-commerce firms’ performance

The week began with the news of the online grocer PepperTap closing its grocery business to become a pureplay logistics firm (see here). And we just read that SnapDeal is recalliberating its performance metrics (read here). So, what exactly is the problem and what can we do about it?

The investor obsession with GMV

Throughout the world, venture capitalists and other investors have used the metric of Gross Merchandise Value (GMV) to measure the performance of E-commerce firms. Everyone manages what they are measured on. So, all E-commerce firms focused on increasing their GMV, that is increasing the gross value of their sales. What this obsession with gross sales does to firms is that there is significant incentive to pursue what I call as “profitless growth”, where only the topline matters, with no attention whatsoever on all other parameters. Especially so, when the entire industry thrived on deep discounting and low customer switching and multi-homing costs. Focus on just one parameter like the GMV might be valuable when the business just sets up to measure the initial traction amongst the target customer groups, but continued focus on the single parameter can lead to misplaced strategies.

Evolving other measures

After all, E-commerce is also a business that needs to provide sustained returns to its shareholders. As with all for-profit businesses, good measurement of performance should include a variety of metrics that reflect the organization’s priorities and strategies. For a consumer focused multi-category retail business, it would be prudent to measure performance on the following four parameters – (a) gross and net (of returns) revenues; (b) gross and net margins; (c) customer addition, loyalty, and attrition; and (d) distribution of sales across categories in line with the firm strategy/ priorities (merchandising mix).

The bane of COD

The boom of Indian Ecommerce industry and its reach to tier II and tier III towns in India could be attributed to the industry adopting “cash on delivery” as a means of payment. With the proliferation of mobile phones and 3G/ 4G coverage across the country, customers with smartphones, and with no access to any digital transaction platform (like a credit card/ debit card/ wallet) can easily buy goods online. And pay for them when they actually receive them through cash. The impact of this on Ecommerce companies is three-fold: adding more number of customers, providing time for customers to actually make up their mind – they could actually return the goods when they arrive with no liability at all (see the recent Flipkart ads), and larger investments in working capital for the industry (either the platform or its suppliers, or both). Therefore, it is imperative that the E-commerce firms measure not just their gross merchandise value, but include the GMV net of returns, or Net Merchandise Value (to account for returns).

E-commerce = discounts

The primary selling proposition of E-commerce firms in India have been around deep discounts. While the idea of a zero-inventory marketplace model (that Amazon pioneered over a decade and half ago) does provide sufficient economies of scale and cost advantages, competitive discounting in the Indian E-commerce industry has over the years shaped customer expectations to the extent of equating online buying to deep discounting. Therefore, measuring gross and net margins of the entire firm is imperative.

Spreading Commerce to the “hinterlands”

The first line of defense Ecommerce firms take umbrage to when someone accuses them of being focused on a single parameter is that “the industry is in its infancy, and we need to broaden our net”. True that the low penetration of E-commerce in India provides a big opportunity for growth, we need to define appropriate metrics to measure the firm’s performance on that front. It is therefore important that firms measure the total number of transactions (as a proxy for volume sales in the offline world), number of active customers (as a measure of customer concentration – or an ABC analysis of customers), number of new customers added (not registrations but at least on transaction), average GMV per customer (as a measure for identifying high-value customers), average contribution per customer (gross profitability), and the proportion of customers whose GMV increased over the past period. In addition to this, we need to also factor in the cost of acquiring customers (CAC), and derive the long term value (LTV) of customers to evaluate performance. As the firm matures, it should strive to bring the CAC lower than the LTV.

PepperTap’s source of worry (read the article on YourStory.com here) was its rapid expansion to new towns where the costs of servicing was far higher than the LTV of the customers in those geographies. As they cut down on the number of cities, their performance improved. Therefore, it is imperative that Ecommerce firms measure and report their CAC and LTV of their customers as a key performance metric.

Alignment with strategic priorities

For a multi-category retailer, the distribution of its sales, costs and margins across categories is a critical parameter to monitor. Firms may prioritize certain categories over others as per their market position and strategic priorities. Successful firms therefore need to measure and monitor their performance across categories, and benchmark against their intent and priorities.

Creating a holistic dashboard

In sum, E-commerce firms would do well to measure, monitor, and report their performance on the four categories of parameters including (a) the traditional GMV, and a GMV net of returns; (b) overall gross margins and net margins for the firm; (c) total number of transactions, number of active customers, new customers added, average GMV per customer, average contribution per customer, proportion of customers registering increase in contribution over the past period, and the cost of acquiring customers; and (d) distribution of GMV, GMV net of sales, gross margins, net margins, number of (net of returns) sales, and CAC & LTV numbers in each category.

Interesting times lie ahead for the industry, as the golden tap of venture capital finance dries up, leading to reduction in discounts and possibly consolidation of firms to leverage the traditional scale economies of a zero-inventory marketplace model.

Ratings, reviews, and recommendations in platforms

In my post last week, I talked about crowdsourcing ratings and reviews to create and sustain credibility of the platform. Almost every platform that operates in a multi-sided market has a mechanism for the users of one side to rate the other side. In this post, I will talk about how to design an appropriate system of measuring the quality of the entity/ product/ service.


The dictionary definition of rating reads “classification or ranking of someone or something based on a comparative assessment of their quality, standard, or performance.”

At the end of every ride, OLA Cabs requests riders to rate the driver/ cab, and the driver ratings are available to the riders when they book the ride. Similarly, Uber has a two-way rating system, where riders rate the drivers and the drivers rate the riders. And the average ratings matter for the driver and the riders to continue using the platform.

The primary (definitional) issue with rating is that it is a comparative score. As a rider takes more and more rides in the OLA system, she is able to compare that particular ride with reference to the other rides in the same system. However, when a Uber loyalist (say for example, my colleagues from USA) takes an OLA ride while in India, he is rating his ride with reference to his Uber in San Francisco benchmark. And when someone who rarely takes an OLA (and otherwise relies on public transport like suburban trains/ buses) would rate his ride with reference to his bus ride. As the references change, the meaning of the same rating changes. Which brings us to the next concern with ratings. That it is always an overall score. The riders may penalize the driver with a lower rating for whatever reason: not able to find your destination, taking a longer route, not having the cab clean enough, or even for this things outside his control like a temporarily blocked road. And the same could be true of a superlative rating – depending on the rider’s benchmark, he could rate the driver a five-star rating in comparison to the crowded Chennai-Chengalpet suburban train, that he takes daily.

This is not to say that ratings are not useful. Over long periods, with sufficient data points, ratings do bring out the true quality and standard of performance. Underlined here is the “long periods of time” and “large number of data points”. Long periods of time provide sufficient opportunities for services with low ratings to improve their performance and sufficient data points provide for cushion against freak low (or high) ratings provided by irrational customers.

One insurance against inclinations to rate a service at either of the extremes (no central tendencies work here) is to decompose the ratings into various service touch points. For instance, the Jet Airways’ service tracker seeks feedback on every aspect of the flying, making the entire responding to the online questionnaire a drudgery. Such long questionnaires would therefore only attract people who have a reason to provide you feedback – who really had a bad experience and want to express their distress, or those who had a superlatively (and unexpected) great experience that they take the effort to fill-in the forms. When the service is as expected (good or bad), one wouldn’t expect customers to fill in long forms (unless mandated). Isn’t this why most of us teachers’ feedback scores have high standard deviations?


As a service aggregation platform, one would want to supplement rating scores with a descriptive assessment (justification) of the rating. For instance, the OLA cabs app would request you to provide the reasons for a low rating by choosing one of predefined set of options. One could not choose multiple options – for instance, it is possible that the driver was late, as well as had his car dirty. This is where open ended responses add value. Again, like long itemised rating forms, open ended questions attract respondents with extreme experiences.

Restaurant aggregators like zomato, ecommerce firms like Amazon.in, and travel sites like Booking.com have implemented reviews along with ratings. Zomato’s review forms require reviewers to provide details of their visit to the restaurant, and the food they ate. In the absence of such information, such reviews may not be relevant to the readers, who intend to use these as the basis for their decision making.

Reviews add value by highlighting specific peculiarities in the product/ service offerings that could not be captured by the ratings. For instance, a sensitive Uber driver who would play appropriate music that is appreciated by the rider would not be a standard data point that Uber wants to capture for all its drivers. However, such an information would be a great input to subsequent riders of that particular driver, who may choose to engage with him about the music. When this becomes a sufficient enough point of discussion in the reviews (enough people write about it about sufficient number of drivers, positively or negatively), Uber might take cogniscance of this to add this to the standard rating form. This is where detailed analytics of the reviews is required.

The dictionary definition of review is very insightful to our discussion: “a formal assessment of something with the intention of instituting change if necessary.” Good analysis of reviews should lead to change, if necessary.

Like the different benchmarks issue with ratings, reviews suffer from an assessment of credibility of the reviewer. It is important that the reviewer is an expert/ has demonstrated that he has used that particular product or service. Amazon.in certifies reviews with a tag “verfied purchase”; and provides the readers of the review an option of rating the review, if that was useful at all or not. Travel sites like booking.com ensure that reviewers have actually booked their stay on that particular hotel and provide the exact details of the reviewers’ credentials to provide the review. In the absence of such credibility, reviews could be abused and gamed in various ways.


Ratings and reviews are good apriori inputs to customers making product/ service selection choices. However, in the case of platforms like Practo, where one chooses physicians (doctors), I am not sure ratings and reviews are sufficient. When the client-service provider relationship is being evaluated (where the service provider is more knowledgeable than the service consumer; unlike a customer, where the customer is more knowledgeable than the service provider), ratings and reviews fall flat. Would you choose your dermatologist based on ratings by other patients, or by the recommendation of your trusted general physician?

The dictionary meaning of recommendation is revealing: “a suggestion or proposal as to the best course of action, especially one put forward by an authoritative body.” Notice the phrase – authoritative body. Credibility not just by consuming the product/ service, but other certifications would be required for a recommendation to be taken seriously. Most popular doctors might not be most efficient. And mind you, the ratings are reviews might just be about the quality of the infrastructure, waiting time to meet the doctor, friendliness of the staff and the doctor, as well as other clinical processes followed by the doctor and her staff. However, while seeking a recommendation for a serious illness, there could be clients who trade-off these against doctor’s effectiveness in curing the illness. Here is why platforms like Practo would require doctors to add their certifications and academic credentials, and mandate that they update them every six months, apart from the ratings and reviews by patients.

So, when you design you platform’s user experience and feedback system, choose carefully – is a rating sufficient, or would you also want a review and a recommendation?

How to build a platform business?

In his recent convocation address at our institute (Indian Institute of Management Bangalore), Mr. Nandan Nilekani stressed on how platform firms have come to dominate the Indian (and global) markets, and the need for our graduating students to understand them well (see http://www.iimb.ernet.in/convocation-2016). In this post, I would focus on categorizing different types of platforms, and some key issues in building a platform business.

Platforms are firms that operate in multi-sided markets. Unlike firms where products and services flow in one direction (remember, Porter’s value chain?) in a pipeline fashion, and money flows in the opposite direction, platform business models connect multiple sets of users. In the traditional sense of the word, a railway platform helps passengers find their trains and vice versa. The train station manager sets the rules, provides the infrastructure, and enables a smooth discovery and transaction between the different sides (trains and passengers). Imagine trying to find and board trains like you would board a taxi in the streets of Mumbai or New York! Generalizing this, the firm that provides the infrastructure is the “platform provider” and the one that sets the rules and norms is called the “platform sponsor”. In some cases, the platform provider and sponsor could be the same firm (like in the case of a railway platform); and in some other contexts, the platform provider could be different than the sponsor (like in the case of Uber or OYO rooms, where the cabs/ rooms are owned by independent entrepreneurs and the rules of the exchange/ transaction is set by the aggregator).

Platforms match different sides of users. In their role as matchmakers, they provide different value propositions – discovery, quality assessment, norms for interaction, setting expectations, and provide feedback – for each of the sides. Let me discuss how to build each of these value propositions (when you are setting up a platform business model) in detail, with examples from established platforms.


This is in fact the first thing to focus on when you set up the platform. Setting up the infrastructure to facilitate interactions is the easiest thing to do. The most difficult part of process is the populating the sides with users. Here is where new platforms encounter the classic “penguin” problem, where users on one side postpone adoption till such time there are enough users on the other side. How would you like to be the “first” person to be listed on a dating site, seeking to find a date? You would affiliate with a dating site only when you are sure that there are already enough members on the other side. Platforms need to overcome this inertia by incentivizing one side to affiliate, in anticipation of affiliation by a large number of right kind of members on the other side. Various platforms have solved the penguin problem differently. For instance, Facebook solved the penguin problem by starting small and being focused on Harvard University students and alumni. Practo  solved it by building and selling their practice management software (Practo Ray) to clinics before opening the patient interaction platform.

Having solved the penguin problem, i.e., having built enough members on both sides, platforms have to ensure that the discovery engine is powered to ensure quality, current, and relevant results. For an interesting take on how Indian ecommerce firms stack up on search results, see this post by Aditya Malik.


In a platform where products/ services/ information are provided by independent parties on one side, it is imperative that the platform ensures quality. It would require verifying the genuineness of the information provider as well as the veracity of the information. For instance, Quickr.com (an Indian C2C marketplace for used goods) positively discriminates posts with pictures of the items being offered for sale than those posts without pictures while sorting the search results. IndiaMART (the B2B marketplace for industrial goods) certifies the sellers with a TrustSEAL, by verifying the antecedents of the seller’s businesses, including their legal compliance, manufacturing facilities, and product range. Verification of quality comes with a cost, and provides the platform with high credibility and enables loyalty of users.

Some platforms use user-ratings and reviews as indicators of quality (like zomato.com, the restaurant discovery platform). Crowd-sourcing of ratings and reviews might provide higher credibility to the platform, but has to be used cautiously. These could be gamed by users. More on this later.

Norms and rules

As a platform sponsor, it is important to set the norms for communication and interaction among users across the different sides. These norms should set the boundary conditions for interactions, like the terms of sale (delivery charges, delivery times, returns policy) in ecommerce marketplaces.

In pure discovery platforms like JustDIal.com (an online yellow pages) or Quickr.com (marketplace for used goods), indiscrimately providing mobile numbers could be abused. Quickr.com has in the past few months introduced a secure chat service whereby sellers and buyers could chat with each other within the platform without having to provide each other’s mobile number. Even when the agreement is reached and the transaction has to be completed, Quickr allows for an anonymous delivery service (Quickr doorstep), where the users need not know each others’ personal contact details.

BharatMatrimony (the online matrimony match maker) allows only paid members to initiate communication with others. What this does is to ensure that brides and grooms who are actively seeking matrimony to become paid members, as the other side is unlikely to respect someone who is “not even willing” to pay for discovering his potential partner!

Setting expectations

The platform should allow for users on both sides to clearly set expectations apriori, to ensure that there are no surprises during the transaction. More the information sought and shared during the discovery phase ensures smooth transition to the transaction and fulfilment phases in the platform. Here is where the platform should ensure that there is a mimimum amount of high quality information available about the entity/ prodcut/ service being matched. Imagine trying to book a hotel room on a travel website without information on the hotel location, types of rooms available on that particular night(s), and the rates! It is therefore an important consideration that platform designers need to keep in mind when designing the infrastructure and rules. For instance, dating sites like trulymadly.com ensure that users provide their facebook pages during the registration process. While actual verification of each users claims on the dating sites might be difficult, linking their facebook account to the trulymadly account ensures that they do not lie (too much!) with respect to critical information about themselves (like marital status). Now you know why the job search portal you just signed in wanted you fill in pages of information, and links to your LinkedIN profile.

Providing feedback

Even after ensuring quality, defining the norms, and setting expectations, there could be some errors. The platform architecture should therefore provide for immediate feedback on all four parameters – quality of the entity/ product/ service/ information; relevance/ adequacy of information; currency of information; and the quality of the discovery, transaction, and fulfilment processes. Cab aggregators like OLACabs request for feedback on the quality of the cab and driver as soon as the ride is completed. However, there is no provision (not that I could find) to provide feedback on all the discovery stage of the platform – the time it took for the app to find me a cab/ auto, or the accuracy of the location services within the app.


Surge pricing: Implications for India

Last weekend, I was in Chennai and I keep using taxi-hailing apps a lot when I am outside Bangalore. About half the time, I was offered rides with surge pricing, ranging from 1.4x through 2.5x of the base fare. And in some cases, for rides in the middle of a hot afternoon, I was offered an upgrade from a “mini” to “sedan”. Riding through the upgraded sedan, I was flipping through the news article, and found an article on Uber’s surge pricing lawsuit (see http://www.bloomberg.com/news/articles/2016-03-31/uber-antitrust-lawsuit-over-pricing-green-lighted-by-judge).

On Monday, 4th April 2016, the Economic Times, Bangalore Edition carried this piece about the Karnataka Government regulation of ride-hailing service (see http://economictimes.indiatimes.com/small-biz/startups/karnataka-nixes-surge-pricing-by-taxi-hailing-apps-like-ola-uber/articleshow/51678792.cms). In a nutshell, this regulation caps surge prices, encourages drivers to operate for multiple services (multi-home), and relaxes norms for drivers to affiliate with a ride-sharing platform.

I was left wondering, what would be the implications for the lawsuit in a duopoly market like India, where OLA and Uber compete. Would the economics be any different? In this post, we will understand surge pricing and its economics, in conditions of a duopoly, and the implications of the lawsuit in India.

What is surge pricing? How does Uber justify it?

The Uber website explains surge pricing thus: (https://help.uber.com/h/6c8065cf-5535-4a8b-9940-d292ffdce119).

“Uber rates increase to ensure reliability when demand cannot be met by the number of drivers on the road.

Our goal is to be as reliable as possible in connecting you with a driver whenever you need one. At times of high demand, the number of drivers we can connect you with becomes limited. As a result, prices increase to encourage more drivers to become available.

We take notifying you of the current pricing seriously. To that end, you’ll see a notification screen in your app whenever there is surge pricing. You’ll have to accept those higher rates before we connect you to a driver.”

The core argument is that surge pricing incentivises more drivers to be available during times of high demand. At the core of this argument is that Uber cannot “mandate” drivers to be available when the demand is likely to be higher, and therefore has to “incentivise” them. Just like tipping the driver to be available. The difference is that the amount of the tip is pre-defined. The effects of surge pricing are well documented in the case study by Chicago Booth School faculty here.

There are two challenges: who pays for the incentive – is the charge on the riders justified? Do drivers like this?

Riders’ perspective on surge pricing

A lot of riders (at least in India) do not like the haggling and negotiating with taxis and autorikshaws for a ride both on price and whether they do want to go to your destination. Most of us using public transport in India are familiar with the famous rant “I have to come back from there empty”. Uber and OLA implicitly promised to eliminate it with its fixed and transparent pricing. Add to that, the ease of hailing a cab to your doorstep/ boarding point, thanks to the mobile app and navigation tools available for both the driver and the rider. In that sense, cab-hailing apps should have their loyalist converts. However, when surge pricing is applied, a typical rider would think twice before confirming the same – she is confronted with the same behaviour the autorikshaw driver on the street would have told her – “this is my price, take it or leave it.” In a sense, it brings out the haggler in the rider, just that now the firm is haggling, and the rider is not even sure the driver benefits out of it (more on the drivers’ perspective later).

The second thing that puts off the rider requesting the ride is the number of cabs available on the map before the surge pricing announcement is made. If surge pricing was indeed designed to get more drivers available, what is happening to all these cabs stationed around my pickup point? Here is where customers begin their multi-homing behaviour (having/ using multiple apps at the same time). She would immediately try the other app – OLA or Uber to see if there is surge pricing there.

It hurts when there is an emergency or a discomfort, like having small children/ elders waiting with you; a flight/ train to take; having to reach for a meeting on time; making another person wait on the other side; or a combination of the above. In effect she is made to negotiate, haggle, bargain for a ride. She does feel she is being taken for a ride, literally.

What does all this result in – a poor experience to begin with, resulting in lower driver ratings. Poor driver, he is being penalised because sufficient number of other drivers were not available. Yes, he may make more money, but at the end of the day, the overall economics may not make sense. Surge pricing acts as a “moment of truth” for the rider, and she resets her expectations. When I am paying 2.1x times the normal fare, I expect the driver to be extra courteous, cab to be clean, and even the traffic to be lighter. On the contrary – surge pricing happens mostly during peak hours when everyone is either getting to work or back home, and on the road; and traffic is likely to be very heavy. All this plays on the riders’ minds and they are lowering the drivers’ ratings.

Drivers’ perspective on surge pricing

A typical driver joined the Uber system (Uber or UberX) with the intention of leveraging her/ his car for monetary gains. Compared to a taxi license, which is highly regulated (check out how to get a cabbie license in London, or what the dominant political parties think of cabbies in Mumbai), or a company employment that can be highly restrictive in terms of business, driving your car for Uber or OLA provides a combination of independence and profitability. At the core of the decision is the value of economic choice the driver has – he can choose when to “switch on”, or become available for a ride, and therefore, how many number of rides and how many hours of the day he wants to ride. This economic choice is also guided by the incentives provided by the cab-hailing firms to the drivers on the number of rides they take per day. Given the very low market penetration in India, most drivers multi-home, i.e., they sign up for both Uber and OLA. And make consolidated economic choices, viz., distributing the amount of time/ number of rides for the two firms so that they can maximise the incentives.

Surge pricing for drivers mimics the pre-disruption world. When there is high demand, I charge more. In fact, OLA (in India) has mandated that drivers should be available through the peak period to be able to earn incentives. While the peak period varies across cities, it is still such a large window that drivers could balance the demand across both the firms (Uber and OLA). And drivers make the choice of going online to that system where surge pricing is high. Even though they do not know in real time when and how much surge pricing is applicable, once the rider has accepted the ride, they would know. And with experience (and a little experimentation), it is easy to estimate. So, if it is likely that OLA would have higher surge pricing rates than Uber, the drivers would shift to OLA, thereby decreasing the supply available for Uber, triggering surge pricing in Uber. While the increased number of cabs for OLA should eventually bring down the surge pricing, it is not quick enough, as multi-homing riders (with both their apps running) are choosing the ride with the lower surge price.

Elasticity of demand and supply in a duopoly

This brings us to the question of how truly “elastic” is the supply, give the duopoly in India? Unlike in markets where there is just Uber and competition is poor, India suffers from a duopoly where the drivers’ multi-homing artificially increases/ decreases supply in one system. And when the riders are also multi-homing, the system stabilises and behaves like a monopoly, albeit with some time lags (than a monopoly market).

Would a flat peak-time pricing work? It may, but given the technology and its ability to discover real-time demand and supply, surge pricing is any day superior algorithm to peak-time pricing. Fixed peak-times are things of the past, when people went to work at the same time in the morning and returned back home around the same time in the evening. Flexible working hours, working from home, working for overseas customers in different time-zones … peak-times are stretched throughout the day in urban metropolises like Bangalore and Mumbai.

The law-suit [Meyer v. Kalanick, 1:15-cv-09796, U.S. District Court, Southern District of New York (Manhattan)]

The law suit against Uber’s CEO, Mr. Kalanick (note, the suit is against the CEO, and not the firm Uber) claims three things:

  1. Uber is not just a technology company, selling apps, but a transportation company
  2. Drivers are employees and not independent contractors
  3. Price fixing by the CEO of Uber, while using fixed prices despite using non-competing independent contractors as drivers

The arguments against these charges by Uber and its CEO are that Uber is just an aggregator, and the drivers as independent service providers have wilfully entered into an economic arrangement to agree to the policies set by Uber to find good quality and quantity of riders. Implying that without a platform like Uber, drivers would find it difficult to discover good quantity and quality of riders. And vice versa for riders. The transportation contract is between the rider and the driver. If this were to be entirely true, then the rider should have absolute choice of drivers/ cabs, as well as drivers to have absolute choice of which rider to take. Given that Uber make the match, and provide you “one” driver-rider combination, this true contract is questionable.

Drivers as independent contractors may be defensible with the argument that drivers have the choice when to login to the service. They may choose to switch off when they want, after fulfilling some minimum conditions. If they were employees, the firm would mandate more than a set of minimum conditions and behaviours; and would not provide the driver with the absolute economic choice provided in the current arrangement. It is to Uber’s economic policies that they have signed up to, and that explains why drivers with different economic expectations can co-exist in a single platform.

The price fixing charge is defensible with the argument that since drivers are independent contractors, it is important to “incentivise” them rather than “mandate” them to be available in peak demand times. To argue that these drivers do not compete is flawed, as the extent of surge pricing is determined by the supply of drivers in relation to the demand. In order to be available on Uber, each driver must maintain certain service quality, and do a certain number of rides. There is definitely competition amongst the drivers – they would want to be available where demand is likely to be higher than supply; and be there before other drivers. For accusing Uber of price fixing under anti-trust laws, Meyer should establish that in spite of varying demand and supply, Uber maintains the same price, coordinating with independent contractors (drivers) who do not compete.

Here is where surge pricing comes to Uber’s rescue – it is their most effective defence against price fixing charge.

Implications in India

The Indian law is uncertain, to say the least, on the regulation of platforms. The (in)famous case of a Uber male driver raping a female rider in Delhi is a case in point. Uber was first banned by the Delhi Government, the ban revoked by the High Court, only for the Delhi Government to subsequently not approve its application. Uber and OLA then went back to court and the courts agreed to revoke the ban when they promised to replace diesel cars with CNG vehicles. What this means in terms of legality is that Uber and OLA are in fact, undertaking on behalf its independent contractors.

The courts and everyone else in India would be waiting for the judgment of the class action suit in the USA on how this market pans out. Given the size of the Indian market, classifying drivers as employees, and Uber and OLA as transportation companies would kill the platform business model. While I do not believe that it would not go that extreme, interesting times lie ahead on how the courts and regulators interpret the developments.

100% FDI in e-commerce – will prices fall?

On the 29th March 2016, the Government of India allowed 100% FDI in Indian e-commerce firms. While there is reason to cheer about the fast-growing sector getting more access to much needed funds for fuelling growth, there are three interesting developments in the notification.

  1. The Government has explicitly defined what is a marketplace model, as different from an inventory model.
  2. The consequence of this definition means that marketplace ecommerce firms cannot have a single retailer selling more than 25% of the retailer’s sales.
  3. The definition also means that the retailer cannot provide discounts and promotional offers on their own, directly or indirectly.

The impact of these three definitional changes would in the short run, require marketplace e-commerce firms to discontinue price discounts they offer directly or indirectly. Amazon’s promotional funding to sellers, PayTM’s cash back offers, or Flipkart’s big billion sale have to end. Will this mean they would stop offering discounts? I do not believe they will. They will find other ingenious ways of providing the customer with discounts, given that they would have access to larger source of funding through the FDI investments. More on that below.

It’s the sellers that matter

This definition of the marketplace model would clearly lead to interesting dynamics on the seller side. For instance, an SMB seller who would otherwise be listing his goods across multiple e-commerce companies would now be wooed by more and more marketplaces, as they seek to expand their base of sellers. Do you realize that the firm that owns the site www.amazon.in is actually called Amazon Seller Services India Pvt. Ltd.? In order to expand and sustain their broad base of sellers, these marketplaces would now have to offer discounts and freebies to the seller side, rather than the buyer side as it was apparent in all these years of growth. These seller-side offers would eventually translate into lower prices for buyers in two ways.

One, in the traditional sense of the word, the seller bargaining power would go up; sellers’ multi-homing costs (costs of simultaneously offering their products and services across multiple marketplaces) would come down; and the volumes would go up. Larger sellers therefore, would invest in technology to manage their multi-homing costs, automate a lot of processes, outsource specialised functions like last-mile delivery to focused service providers, and would grow their own sourcing networks. Smaller sellers on these marketplaces would have no incentive to be remain small, and would either get gobbled up in a consolidation game or become second-tier sellers to the larger sellers operating on the marketplace e-commerce retail. This consolidation and growth of sellers on the marketplace would result in lower costs through economies of scale and scope, which the seller would eventually pass on to the buyers.

Two, the consolidation of the seller market would lead to fierce competition across sellers; and the basis of competition between the sellers is likely to be only price. Other differentiators like product variety/ features and brand are likely to be owned by manufacturers/ marketers (like Samsung), whereas service differentiators like distribution network, logistics and related customer service are likely to be managed by the marketplace. The only bases of competition for the sellers to compete would be (a) optimisation of inventory to reap appropriate economies of scale and scope, (b) managing distributed inventory through accurate prediction and forecasting of demand and supply, and (c) reducing costs through faster inventory turns as well as leveraging their bargaining power with manufacturers as well as retailers/ ecommerce firms.

Good times are here to stay (for the consumers)!

So, in effect these regulations do not necessarily mean that the prices in the ecommerce retail would rise and match the offline prices. There may be small adjustments; but in the long run, the discounting would shift from the retailer to the supplier. And the consumer would continue to enjoy lower prices (offered by the sellers) along with superior customer service (provided by the retailer, as this would be the only basis of competition across  marketplace e-commerce competitors).