How much data should a Google make public? And when do the scales tip for the business to be not worth its while?
Several interesting questions have got thrown up over the past few weeks when it comes to big tech firms. There are, in the US, the hearings on whether Big Tech—a Facebook, Apple, Amazon, or Google (FAANG, without Netflix)—needs to be broken up. And in India, the issue of curbing data monopolies is a live one with, as it happens, the government inviting suggestions—August 13 is the last date for submissions—on the report on sharing of non-personal data.
While Google and Facebook aren’t ripping off customers—how can they since they provide their services free?—the issue here is whether they are using their power, including over data, to stifle competition. Do the insights Google gains by offering services worth billions for free give it an edge over the competition in, say, the advertisement market? And if they do, how undesirable is that? More important, is it worth breaking Google up to fix this? Think of, for starters, the Android ecosystem it has created and how that has dramatically lowered the costs of mobile phones.
When a final decision is taken on whether or not Big Tech firms need to be broken up, or their powers circumscribed in some other way, it will have interesting lessons for the Competition Commission of India (CCI) and other regulators. CCI cleared the Facebook investment in Jio Platforms in very little time in June, and was quick to dismiss the allegations of predatory pricing by RJio some years ago. This is not to pass judgment on CCI’s rulings, but just to say that thinking on competition issues is evolving the world over, and issues that were dismissed as not important earlier are being relooked with a different lens today.
The issue of data monopolies and the powers of Big Tech, of course, are intimately connected. Sure, the issue of Facebook buying Instagram or WhatsApp is about using its money to buy out the competition, but data monopolies can, the argument goes, have a larger impact on stifling competition. In fact, one of the charges against Amazon (on.wsj.com/30AbVuT) is that it used data from its suppliers to build products that competed with them! Given how much it sells on its platforms across the world, knowledge about what sells and at what price-points is a big advantage to have. But, think about it a bit more, and how much of this can Amazon possibly do? After all, if it does too much of it, no one will want to sell on its platform.
The issue of data monopolies is what India is hoping to address through its sharing of non-personal or anonymised data. To use the Amazon example, once the principle of sharing is accepted, Amazon will have to part with its data—on payment of a sum that a Non-Personal Data Authority (NDPA) will determine—to anyone that asks. Ditto for Google and Facebook, WhatsApp, etc. As for data available with public authorities like the GST Network or civic authorities—how many cars travel to Mumbai’s BKC during rush hours?—this will also be shared, possibly for free.
Conceptually, this is easy to do, but it throws up several questions that are less easy to deal with. If Retailer A applies to the NDPA asking for the top-selling items on Amazon’s website, duly anonymised of course, will this be given? It is data that is proprietary in nature, but it is non-personal data (since it is anonymised), and sharing it will help boost the competition and, in a sense, the data belongs to individuals, not Amazon; put together, all these criteria suggest such data must be shared. Yet, the fact remains that if it is shared, it takes away a lot of Amazon’s USP, its ability to produce/procure and deliver, in time, what the customer wants, at desired price-points, and in different parts of the country.
What is the top-selling garment in the country, in what fabric, what design? Where can you procure the best-quality silk trousers? Marketers would normally spend a small fortune in trying to figure these questions out, but now this can simply be asked for via the NDPA?
It is obvious, as Infosys co-founder Kris Gopalakrishnan—he headed the panel on sharing of data—has said in an interview (bit.ly/2PzNHL7) that simply giving out raw data won’t help, it is what you do with the data that matters, and the business model that you build on top of this data. That is, if Retailer A takes Amazon’s data dump, it won’t help; Retailer A will have to work on the correct algorithms to be able to tweak out the valuable consumer insights that Amazon got by investing billions of dollars in setting up its sourcing and distribution network. But surely a top competitor will have the ability to develop the algorithms to extract the relevant data; and surely there will be lots of boutique firms who will do the data analysis for a fee? Gopalakrishan, after all, says Nasscom-McKinsey has estimated the data business could be as big as $500 billion over the next five years.
It is important, as Gopalakrishnan says, to share data equitably and to ensure that data monopolies don’t stifle the competition. But a lot of this data is collected through huge investments. Is that something that should be shared just like that? Why not force Coca-Cola to make its secret recipe public? After all, that too helps cement its market dominance. These are issues that the government, and even the Gopalakrishan panel, will need to take another look at.
To begin with, start making government data available; this could be the number of diabetics or the sales of analgesics in a particular PIN code, or traffic patterns, etc. As for data collected by private firms, a much larger debate is called for as well as a study of its pros and cons to be able to understand whether it is a public good. An individual selling the details of what she bought, read, travelled, ate, etc is one thing; a company being forced to part with customer data it has acquired is quite another.