As a particle physicist at MIT, Mike Miller needed a better way of juggling the massive amounts of data generated by the Large Hadron Collider. So, together with fellow MIT physicists Alan Hoffman and Adam Kocoloski, he built a cloud service.
Dubbed Cloudant, the service evolved into a startup backed by $20 million in seed funding, providing a means of storing and analyzing "Big Data" not just for particle physics research, but for all sorts of websites, mobile apps, and other commercial ventures. Miller became the company's chief scientist, but, this being a startup, he wore other hats too. For the first 15 months of Cloudant's existence, he also ran the company's customer relationships, recruiting businesses to the service, teaching them to use it, and working to keep them using it. But then, as the startup matured, Miller passed this job to a new hire, a former trader at one of the big London financial houses.
>As Miller introduced him to the company, the trader's first thought was that Cloudant wasn't just running an internet service.
As Miller introduced him to the company, the trader's first thought was that Cloudant wasn't just running an internet service. Because it was helping such a diverse collection of other companies run their own businesses, the trader said, Cloudant also controlled something that looked a lot like a venture capital portfolio. "The great thing about a cloud service is that the majority of the companies you help run are new," Miller says, "and indirectly, you end up with this incredibly diverse portfolio of companies spread across so many markets." Cloudant wasn't investing money in these companies the way a VC would, but it was still very much interested in their success, and because it ran at least part of their tech infrastructure, it could track their performance.
Then Miller realized how valuable this arrangement could be. "It occurred to me that this could give cloud services an edge when it comes to M&A," he says, referring to that eat-or-be-eaten aspect of big business: mergers and acquisitions. If you run a cloud service, he explains, you can immediately see when a customer's business starts to take off. You can see how much a customer is spending on infrastructure. And you can see how much that infrastructure gets used by the world at large. "You have an in on how popular something is," he says, "and it could give you the inside scoop if you were interested in diversifying, buying a company or moving into a new market."
Miller and Cloudant have never actually acquired one of their customers in this way, but they've still experienced the phenomenon firsthand. As is so often the case of the multilayered world of cloud computing, Cloudant is a cloud service that runs on top of another cloud service -- Softlayer, the service owned and operated by tech giant IBM. In other words, IBM could see into Cloudant's business, and in February of last year, it acquired the startup.
First and foremost, cloud services -- from Cloudant to Softlayer to Amazon EC2 and Google Compute Engine -- are places where a world of businesses and independent developers can build and run software without setting up their own computer servers. But for the companies that operate them, these services can double as the most effective of market intelligence tools -- at least in theory. After all, Amazon's cloud services already run as much as one percent of the internet. As cloud computing grows even more popular in the years to come -- and as it consolidates around a small number of enormous companies -- this could become a distinct advantage in the tech world's never-ending battle for the future.
"There's no question that companies that run clouds see things before others," says Graham Weston, the CEO of Rackspace, the company whose cloud operation ranks second only to Amazon in terms of size.
The phenomenon isn't discussed all that much in the press -- perhaps because cloud operators worry it may seem like they're taking advantage of their customers. Some operators insist the information available to them really isn't that useful. But, like Miller and Weston, other insiders very much see the potential of cloud services as intelligence tools, including Lance Crosby, the CEO of IBM's Softlayer, and Lucas Carlson, who oversees cloud computing at internet service provider CenturyLink. "This is something," Carlson says, "that's often talked about in the cloud community."
>In a broader sense, you can use a cloud service to see technological trends before the rest of the world does.
If you run a cloud service and you see a customer's business taking off, you may have the inside track on acquiring it -- you have an existing relationship with that customer -- but you can also grab a head start on cloning its technology. And, in a broader sense, you can use a cloud service to see technological trends before the rest of the world does. You can see when large numbers of internet users gravitate to mobile messaging applications, or perhaps even when software developers who build apps on your service start using a new open source technology. You can use this information on your own, or you could use it to help outside clients, as IBM indicates it will do.
"Before Softlayer, IBM had no insight into pre-A-round [startups]," says Softlayer CEO Lance Crosby. But this has changed, and for Crosby, that means IBM can spot the kind of trends that Fortune 500 companies may one day care about. "Ten years ago, they weren't asking about startups in Silicon Valley," he says of IBM's Fortune 500 clients. "But they are now."
For Miller, this kind of thing is analogous to the way Google uses its web search engine to understand what's happening in the consumer world. Thanks to its search engine, Google knows what people are looking for and what they're visiting, and it can serve ads accordingly -- and perhaps even use the information to move into new markets. But a cloud services may provide insight on another level. Cloudant's customers have included heavy hitters such as Samsung, Fidelity, and Novartis, as well as many startups that have been acquired by bigger names, including Orchestra (now owned by Dropbox), Omgpop (bought by Zynga), Oculus VR (bought by Facebook),and Bright.com (bought by LinkedIn).
"This can be a huge differentiator," Miller says. "I don't know how much we've seen this realized yet, but cloud computing is a new market and a new model."
This phenomenon could lend added importance to Facebook's acquisition of Parse, a cloud service specifically for running mobile apps. Facebook is now building and acquiring an army of individual mobile apps to compliment its primary social networking service, and Parse could help decide what direction to move in, providing a window into new technological trends and showing what smartphone apps are the most popular. Ilya Sukhar, who oversees Parse at Facebook, says the company could use the service to identify what apps are generating a lot of traffic, but he indicates that the company is not doing this, and he says Facebook is not delving into an specific customer data.
>'If you look at what Facebook is doing, imagine if Amazon wanted to do that. Nearly everybody uses Amazon on some level.'
Mike Miller says that, thanks to its cloud services, Amazon is in a prime position to diversify its operation in much the same way that Facebook is. "If you look at what Facebook is doing, imagine if Amazon wanted to do that," he says. "Nearly everybody uses Amazon on some level." Traditionally, Amazon doesn't acquire many companies. But very often, it clones technologies offered by others. And as Google expands its own cloud services in big ways, it may also be in a position to use cloud intelligence to its advantage. (Amazon and Google wouldn't discuss how they use this type of information internally, though a spokeswoman for Google Ventures, the company's VC arm, said that Ventures does not do intelligence sharing with the company proper).
Jason Pressman, a managing director with Shasta Ventures, agrees with Miller's assessment, saying he'd love to have the information available to the big cloud giants. "They would know at the same time that the company knows the application is taking off," he says. "That would be a very very powerful competitive weapon for any investor who wanted to play in these ecosystems."
Others disagree, arguing that anyone can get market intelligence looking at the public statistics from things like the Apple App Store and web-tracking sites like Alexa. Peter Levine, a partner with Andreessen Horowitz, says the data available to cloud operators doesn't go that much further. "You'll get some information on how many servers an app is using and how much an app vendor might pay for the service, but you don't really get a full understanding," he says.
But still others push back on these arguments. They're adamant that cloud services can provide a depth of information and an immediacy you can't get elsewhere. Alexa's data, after all, isn't that recent or that accurate, and the App Store merely shows you what apps are getting the most downloads. If you know how much a company is spending on infrastructure, Miller says, you know a lot more. "You don't know their overall earnings, but you do know what their spend is -- because you dominate it," he explains. "It's not quite a board seat, but it's close."
Is this is an invasion of privacy? Miller doesn't think so. The fact of the matter is that, if you run your operation on a cloud service, the company running that service is going to know how much money you're spending. That's just the nature of doing business with another company. "This isn't about inspecting network packets," says Miller. "To me personally, it doesn't seem like a terrible thing to recognize the overall infrastructure spend and recognize that this is probably correlated with either business or consumer demand and know that this is a business opportunity."
>'You just have to look at the bills. You can just look at the total spend from these companies, rather than their server workloads.'
As CenturyLink's Lucas Carlson puts it: "You just have to look at the bills. You can just look at the total spend from these companies, rather than their server workloads." But on top of this, a cloud company can also see how much your infrastructure is being used. And that's what makes the cloud different from, say, just selling software or hardware to your customers. Whether the business side of a cloud provider has the right to use this data -- or even share it with outside outfits -- comes down to the provider's terms of service.
And those terms of service are typically broad enough that it's just not clear whether the cloud provider will do with this type of data, says Michael Overly, a lawyer with the Foley & Lardner law firm, who advises businesses on cloud computing contracts. For businesses that have enough clout to set their terms of service contracts, he advises them to add a line stipulating that their vendor "will only use client data for the performance of the services for client's benefit." If the provider won't add that, he says, "we get concerned."
In any event, as time goes on, the largest cloud companies will only have a more extensive window into the ever changing world of tech. By 2020, says research firm Forrester, cloud services will account for about 15 percent of the money businesses spend on the hardware, software, and services needed to run their operations, and the market is pooling around giants like Amazon and Google and IBM. For Mike Miller, it's merely obvious that these giants would use the insight provided by their cloud services to their own advantage. "It just underscores another reason why the cloud is about the service and not the software that drives the service," he says. "Services are the new business."
Additional reporting by Robert McMillan
