Just how big are the digital monopolies?

The continued rise of digital platform monopolies – the dominant organisational form of contemporary capitalism – is likely to further concentrate economic gains and will drive debates about how data should be governed.

I guess we’re talking about Facebook, Amazon and Google here, yes?

Turnover of a company is the wrong thing to look at, should be wages plus profits to equate to GDP. However, ignoring that – $30 billion Facebook, Google $100 billion, Amazon $135 billion.

Call that $300 billion among friends. US economy is around $20 trillion. That, overmeasured – and I’m comparing global to national GDP as well – is some 1.5% of the economy. Concentration, eh?

51 thoughts on “Just how big are the digital monopolies?”

  1. True, but others would say they are the gatekeepers to much else. Online advertising and search are utterly dominated by these three and Microsoft. It may be a small part of the economy, but it’s a rapidly growing one

  2. This isn’t about search engines and the like so much as the platforms that are increasingly used to host corporate digital infrastructure (broadly the Cloud). So not so much Facebook, but Amazon AWS, Microsoft, Google and IBM. The value of turnover of these companies may be relatively small (even if absolutely large), but their share of the world’s data is growing fast (and the annual volume of data generated by business is growing even faster). If current trends continue, businesses will be increasingly dependent on those firms to handle their data, and data will be the most important component of their business.

  3. Bloke in North Dorset

    Are Amazon in the same class as Google and Facebook?

    Yes, Amazon uses our data but mostly to sell us stuff we didn’t know we needed or even that it existed. I’m sure they use it to sell advertising but that’s not their main revenue line.

    Whereas as with Facebook and Google we are the product.

  4. It wouldn’t matter if they were mere monopolies in providing basic social networking, search, web hosting, etc. The problem is that they quickly become politicised. All three have the power to block objectionable content or individuals. As per your previous post, Facebook is guilty of allowing people to say that maybe Hillary ain’t that great; Google’s search algorithms have the power to make awkward topics unfindable. Amazon’s power is less entrenched, but it does host an awful lot of websites, which it could block at any time.

  5. It’s all been said by the previous commentators. In a nutshell, name the biggest competitor to each of those companies (I’m gonna add Apple to the mix), and tell me what market share they have.

  6. @BiND: Amazon is the biggest player in cloud computing, followed by Microsoft, IBM and Google. Their operating margin on that part of their business is about 25% compared with 3% for their US domestic retail businesses. They aren’t making as much out of selling us stuff as they make out of hosting other people’s data and applications on their infrastructure.

  7. As it stands today I am not forced to use Facebook or Amazon and when I do so, I do it out of choice (eBay and other internet retailers often being cheaper).

    Google however is a different matter, since it does offer a fast and efficient web index that is miles ahead of the competition (the nearest of which is Bing and that is just a complete pile of dogshit).

    The other search engine offerings seem to be either proxies for Google with some interesting features (e.g. DuckDuckGo) or specialised such as IMDB.

    Is this a problem? I’m not sure. I’ve tried to reduce my use of Google since the “Diversity Memo” car crash from a few months ago, but it’s surprisingly difficult as Google owns so much of the prime acreage and despite their bias they do deliver technologically good services (Google Search, Maps, YouTube, etc.) for “free”, with the downside that you’re browsing data is being monitored and analysed with a depth and precision that would make the Stasi of former East Germany salivate with envy.

    Not sure what to do really.

  8. @JG
    I don’t think DuckDuckGo is a google proxy. It seems to have its own search index, and gives different results to google in a different order. It’s also far more limited than google search.

    To try to limit the build up of private information about me, I use DuckDuckGo for most searches. A lot of the time I’m only looking for a couple of obvious pages (an organisation’s official site, a news story or a wiki page) and DuckDuckGo is perfectly adequate for that.

    I can always fire up a new private window (with VPN) and use google on that if I want to do a more extensive web trawl, but DuckDuckGo is pretty adequate.

    Not sure why I go to all the trouble to protect myself from Google given that I have an android phone though….

  9. Not sure why I go to all the trouble to protect myself from Google given that I have an android phone though….

    …and if you’re using Chrome, especially on Android then Google still gets your searches even on DuckDuckGo anyway.

  10. The issue isn’t so much about whether you use Google or Amazon’s websites so much as the fact that if you useful any business or make any payment, chances are that business will be running it’s operations hosted by one of the large cloud companies, so in theory,. Google or Amazon has access to your data.

  11. @Alex

    This is presumably a sector where any serious player is going to scale, and a consequence of that is that a limited number of players is always going to be likely, right? Of course there are downsides to extreme concentration (anything from the effect on freedom of speech to lower resilience in the face of a state-sponsored cyber-attack) but I’m not sure what can be done about it.

  12. @JG

    Indeed! Quite like Opera myself. Even aside from the privacy issues I don’t like Chrome – don’t even have it installed on my PC.

  13. Indeed! Quite like Opera myself. Even aside from the privacy issues I don’t like Chrome – don’t even have it installed on my PC.

    Completely sympathise. I’ve been using the TOR browser much more of late, not because I wish to purchase drugs from the dark web, but more because it provides a level of anonymity to the clear web that other browsers don’t offer.

    Google of course blocks TOR traffic. Doesn’t like the competition I guess. Fuckers.

  14. In fairness to google, I was mostly triggered into action by taking the “privacy audit” they offer and indeed advertise. Thought “wait, do I really want you to be keeping all of this stuff?” and had a rethink of habits. Though I only took the audit because I’d become concerned just how personalised their ads had become. Incidentally a mate of mine stopped using Facebook this year for the same reason – if they knew enough to advertise *this stuff* to you, what the heck else do they know?

  15. @MBE: I don’t think there is much we can do except be aware of the concentration and potential downsides. It isn’t so much the content that the firm’s get through browsers and search engines, but the fact that as infrastructure providers for a lot of current and future services, they will have access to the data for very substantial parts of the economy. They will be hosting all manner of government data, payment transactions, securities settlements in the cloud whereas that data is currently held in private systems. The business models of Google and Amazon generally give the company rights over client data, Microsoft and IBM less so. As the article states, who gets what rights is going to be a moot point in future. Google and Amazon are going to want to exploit the data, the others perhaps less inclined to do so, but will probably charge more for hosting as a result.

  16. There’s very few large players in the UK in betting exchanges, Stock exchanges or Metal exchanges. Somehow this isn’t regarded as a problem.
    Only one IT supplier to the government’s national benefit systems – now that IS a problem imv, and has resulted in over-run costs.
    But lefties don’t want to look at that kind of dominance.

  17. @Bongo: “Only one IT supplier to the government’s national benefit systems”

    Only one at the moment, but that is hardly a monopoly. The DWP simply chooses to be screwed by one set of consultants at a time. Getting five sets of consultants to do the same probably wouldn’t make things any better.

  18. @Alex There are 3 large IT companies supplying housing benefits software to local authorities, and because they have to compete for the 300 or so customers, they had all incorporated the 2-child legislation about 1 year ahead of time. That’s a bit of a rose-tinted view of competition but afaik the Universal Credit software can’t yet handle families of 3+ kids where one was born after April.
    As you can imagine, I’m a fan of devolution to the lowest realistic level.

  19. John Galt: I use Google for search, but their maps are crap. I use Bings for maps as they actually use actual maps, but their search is crap. Just as I use an optician for eye tests and a dentist for dental work.

  20. MyBurningEars,

    “This is presumably a sector where any serious player is going to scale, and a consequence of that is that a limited number of players is always going to be likely, right? Of course there are downsides to extreme concentration (anything from the effect on freedom of speech to lower resilience in the face of a state-sponsored cyber-attack) but I’m not sure what can be done about it.”

    Precisely. There’s almost a natural shift in almost any market towards small numbers of players because of industrialisation. The only businesses that are made up of lots of tiny competitors are very labour intensive, like hairdressers.

    You could have pointed to the dominance of Radio Rentals, Granada TV and Redifussion in the 70s, or that most arcade machines were made by Taito, Atari or Capcom. Or that home computers eventually ended with just Atari and Commodore. There’s 3 game console makers. There’s 2 graphic chip companies. 3 major CPU companies, 3 smartphone operating systems.

    And let’s not forget this beauty from the Guardian: https://www.theguardian.com/technology/2007/feb/08/business.comment “Will MySpace ever lose its monopoly”.

  21. @BonM4, MBE

    It is not the monopoly that is at issue here. It is the right of the infrastructure hosting companies to access the data they are hosting for others. Assume that all the cashless payment systems run through the cloud, and that Amazon or Google are the companies hosting that service (Netflix already runs on Amazon hosted cloud). Then assume that Google cuts a deal with the payment company to get access to its data in return for lower hosting charges. Suddenly they have a better source of data than the websites and search engines. Would companies give up their data so easily? Maybe not. They might want to do their own analytics. But there is always a Ryanair who will cut a deal to squeeze the last penny out of their costs.

  22. Alex,

    If Google, Amazon or Microsoft ever did that, their cloud businesses would be dead within hours. People would take their business elsewhere. Which is why that’s not what they’re doing.

  23. @BOnM4: You would like to think so, but that is what happens with some of their clients at the moment. In fact it is a selling point of IBM (who probably sell the AI/data analytics as a service than the others) that they *don’t* retain any rights in the data whereas Google and Amazon do. All negotiated per client, but companies sell their client data all the time. Selling to the cloud hosting company is a little simpler.

  24. Alex,

    “You would like to think so, but that is what happens with some of their clients at the moment.”

    Who has paid for hosting on S3 or Google Cloud and found their data being used?

  25. @BoM4: It doesn’t just happen.

    First you get Amazon to run their data analytics service
    https://aws.amazon.com/products/analytics/
    Then Amazon offer to buy the data from you and snslyse it themselves. It needs to be a big corporate app to make it worthwhile, not some startup website that runs in the cloud.

    Google do something similar but the don’t have the same volumes as Amazon.

    IBM specifically target big data applications in the cloud and bring in their Watson (AI) and analytics services, but unlike the other two all rights in the data remain with the client.

    This is an area of considerable focus and growth, and the writer of the original article was right to point this out as an issue. With the growth of technologies such as blockchain (which is far more than bitcoin), rights over data will become moot and as the author says the companies that do the most cloud hosting, particularly where they add BI and analytics services, have an advantage if not quite a monopoly.

  26. Alex,

    Why does Amazon need to pay for data? Didn’t you say earlier that they had rights to it? Or, don’t they? What’s the deal. What do you know about what companies and what they do with regards to Amazon?

    With regards to blockchain, why does that affect data rights?

  27. Several industry leaders have folded already in these fields. Apple almost went down and I wouldn’t put money that Microsoft will avoid IBM’s fate.

    Google keep ahead by inventing entirely new fields. Once that pace dwindles they’ll be vulnerable.

  28. @BonM4:

    Different users, different terms of service. AWS/EC2/S2 and all the add-ons aren’t just bought on credit cards over the net. As corporations move their data centres into the cloud and buy services off Amazon and Google they negotiate custom deals that involve data analytics and data usage rights. Effectively they can expect a better price for their hosting service if they sell data to Amazon or Google as part of the deal (and Amazon or Google want their data).

    Blockchain is relevant because it is a distributed ledger shared by a network of interested p[arties/businesses and generally likely to be hosted in the cloud. Future blockchain applications are likely to feature key business transactions (bank transfers, securities settlements, trade documents such as waybills and air waybills). All of which will be running on systems hosted by the big data companies. Ownership of the data will be a big issue. It is easy to see a world where Amazon/Google have access to details of your buying habits without you ever using their websites.

  29. Alex,

    OK, I’m getting bored with this now. Tell me one customer on S3 or Google Cloud with T&Cs that say that Amazon or Google can access their data. Or show me one contract that either company has that says that. Can you do that, or would you like to admit that you’re bullshitting?

    As for blockchain, no, it has zero relevance to cloud. Cloud hosting is an abstraction layer, managing the physical aspects of hosting for you. And what I know of financial companies, they’re nervous about things like cloud. They like to have their own data centres with their own servers running in them, under their control.

  30. To answer the first part, I don’t have sight of any contracts, but ask any customer of Amazon or Google who also buy analytics services from their cloud provider. That clearly gives Google and Amazon access to the data. Anecdotally, Google and Amazon retain rights in the results of their analysis, IBM do not. I don’t know about Microsoft.

    Blockchain has plenty of relevance to cloud. Just as cloud computing has displaced large systems, expect blockchain to be a big part of future systems. They may be different levels in the architecture if you think of cloud computing simply as infrastructure or a platform, but blockchain applications are likely to be implemented as cloud offerings, not least because of the vast amounts of data they will generate. Such data is eminently suited to the analytics software developed by Amazon, Google and IBM.

    Sure, banks like to keep their systems in house, but (1) that doesn’t stop them using hybrid cloud solutions already and it doesn’t mean they won’t use the cloud in future, (2) there is nothing that says that the future payment systems will operate through financial institutions, (3) there are many more non-financial blockchain applications, and (4) yes, if you look it up Amazon, Google, IBM and Microsoft all offer to provide blockchain support on their clouds, and yes they have projects under development.

    As for your accusations of bullshit, I admit that this is a little prospective, but that is what I do. I am a data architect at one of the firms listed above. Whatever current practices are, the ownership of data by cloud platforms is a looming issue.

  31. Alex, I’m afraid BoM4 has it right. Amazon will do some customer analytics stuff, but it’s all consumer-focused and advertising-oriented. If I rent an Amazon server, or use one of their SaaS platforms to crunch my company’s accounts, Amazon wouldn’t know how to interpret the data, let alone have any interest in owning it.

    The only data Google or Amazon have any interest in acquiring is consumer behaviour data: who is likely to buy what, and when. They already get most of that themselves.

    As for blockchain, I haven’t seen this level of hype since the inventor of the Segway promised that whole cities would be redesigned around it. It’ll find its niche, but in most suggested use cases it isn’t any better than the status quo; and often worse.

  32. Andrew M,

    Yup. There’s a quid pro quo using services like Facebook and Google – you get it for free, but they know about you and advertise to you with it.

    And yes, there are good, practical uses of blockchain, but they’re far fewer than many people are hyping. I like the Chai Wine Vault – allows traceability of fine wine to reduce counterfeit wine.

  33. @AndrewM: Then I guess you are either a few years behind the curve., or a hell of a lot smarter than people whi I assume are paid a lot of money to throw even more money at a combination of cloud computing, AI and analytics. Some of it has paid off handsomely, some of it less so, but it is far from hype. Cloud computing is making serious money and the crossover with analytics doesn’t just work for consumer data and Amazon shopping.

    As for blockchain, yes there is a lot of hype, but there is also a lot of good stuff. Come back here in five years and we’ll ee whose right, but at the moment I see lots of use cases that are ripe for blockchain and most of those are likely to be cloud or hybrid cloud implementations. So lots of transaction data potentially available to cloud operators. Whatever your view of the prsent, that is a more than likely future.

  34. Alex,
    I just don’t see Amazon’s interest in owning whatever data it is that your company needs to crunch, other than marketing data.

    If you run an airline, Amazon might like to know who has booked flights to Greece, so it can flog them Greek phrasebooks or sun cream or whatever. But all the other boring data about pilot rostering, plane load factors, flight computer logs – that’s all very useful to an airline, but of zero interest to Amazon.

  35. Bloke in Costa Rica

    Andrew M et al.:That’s really what it boils down to. All this talk of monetising ‘data’ without knowing anything about its semantic content is ridiculous, which is why the cloud-mongers don’t do it. The companies using the services can leverage the big data facilities offered by the providers (Hadoop and friends, for example) but they know what their data represents.

  36. Alex, as the article you linked makes quite clear, the only accusation levelled at AWS is that its fat margins effectively cross-subsidise Amazon’s retail operations. Other retailers balk at funding their competitor. But there’s no suggestion that Amazon are sneaking into the AWS servers at night and poring over their competitors’ sales figures.

    Your same article even points out that Netflix is happy to use AWS, even though Amazon Prime Video competes directly with Netflix.

    As for Oracle supplying a blockchain service: during the gold rush, the suppliers got rich selling picks and shovels. But where’s the gold in blockchain? Most suggested use cases boil down to signing messages with a private key, without reference to previous blocks: so no chain.

  37. @AndrewM: But you will also find that retailers will generally not use AWS as a cloud service provider because they need data analytics but they don’t want to use AWS’ analytics suite. As an example Home Depot use Google because they are not perceived as a threat to their retail business. Other companies use IBM because they have an explicit clause excluding IBM’s rights in the data when a client signs up for analytics. AWS and Google do not.

  38. Alex,

    “Well waddya know. Just when you say blockchain has nothing to do with the cloud, Larry Elllison goes and releases a blockchain cloud service. Suggest you let Larry know he’s wasting his time and money.”

    Which doesn’t mean that blockchain has anything to do with cloud. Your example is like thinking that shops have something to do with cruise liners, because cruise liners have shops on them.

    As I said, and clearly, you have no idea what you’re talking about, cloud is just an abstraction layer. If you knew anything beyond buzzwords you’d read on places like ZDNet, you’d know this. Logically, anything you can do on a cloud, you can do on a computer, because all the cloud is, is an abstraction layer to computers.

  39. Alex,

    “As an example Home Depot use Google because they are not perceived as a threat to their retail business.”

    Unless you work for Home Depot, you don’t know that. Home Depot have not publicly spoken about why they chose Google. Nor do you know about why companies sign up with IBM.

  40. @BonM4

    Don’t you ever tire of being wrong?

    While AWS is far and away the market leader, chief information officers at some retailers say they chose to go with Microsoft’s Azure, Alphabet’s Google Cloud, or IBM’s BlueMix rather than entrust important parts of their business to a rival.
    “You’ve got to think about who you’re renting your cloud from,” says Paul Gaffney, senior vice president for information technology at Home Depot. “If we announced tomorrow that we were renting data centers from Lowe’s, everyone would think we were crazy.” Home Depot, which began migrating functions such as purchasing to the cloud earlier this year, is working with Microsoft and Google.

    As for IBM’s customers, you are right. I don’t know what they think. I only know what IBM’s senior management tell me that they think.

  41. Alex,

    “Don’t you ever tire of being wrong?”

    You’ve failed to back up every single claim you’ve made. Unfuckingbelievable.

    “While AWS is far and away the market leader, chief information officers at some retailers say they chose to go with Microsoft’s Azure, Alphabet’s Google Cloud, or IBM’s BlueMix rather than entrust important parts of their business to a rival.”

    No, that isn’t what they say. Again, you are wrong. It’s about not funding their rival who they perceive as using cloud to subsidise the store.

    “As for IBM’s customers, you are right. I don’t know what they think. I only know what IBM’s senior management tell me that they think.”

    Ah, you’re an IBMer. I should have figured that out from the sharp-buy-useless vibe.

  42. @BonM4

    Wrong again it seems. That isn’t my text you’re refuting in the first paragraph. That’s the text in the article linked above. Search the Web and you’ll find plenty of similar statements.

    Likewise, search the web and you’ll find reports saying that 90%+ of all server IP traffic/data/workload will be cloud based by 2020. Traditional data centres will be little more than a rounding error
    And yes, blockchain will run in the cloud, and yes, blockchain users are very likely to use analytics on the vast amounts of data, and no, they most likely wont want the details of the analysis to end up with Amazon owners of AWS who currently hold over 30% of cloud market.

Leave a Reply

Your email address will not be published. Required fields are marked *