Business Impact

Collaborating through Streams of Information

JP Rangaswami, chief scientist at Salesforce.com, explains why his company’s new tool for collaboration could pave the way for productive game-like interactions in the office.

Mar 11, 2011

Salesforce.com jumped into cloud computing before the term even existed. The company was founded in 1999 to offer businesses a customer-relationship management (CRM) service that ran online and didn’t require software to be installed on employees’ computers. But last year, Salesforce entered a decidedly more crowded market: the one for collaboration tools. Like other such services, Chatter takes elements from Facebook and Twitter and puts them into an application that helps employees assist each other much more efficiently than they can by endlessly e-mailing documents to each other. The company’s chief scientist, JP Rangaswami, explained to Technology Review’s deputy editor, Brian Bergstein, why Salesforce thinks Chatter is unique, and how collaborating in the office might become much more like a game—in a good way.

TR: What does Chatter do differently from the countless other collaboration tools that have existed?

Rangaswami: The challenge was not in designing a system to collaborate, because as you said, there are millions of them. The challenge was how to take a collaboration mechanism and associate it with systems of record. Systems of record are the replacements for books and ledgers of the firm, and have to be built with a very high degree of security. But the systems of engagement, the ways that people communicate with each other, had to be built on diametrically opposed security principles. And you don’t really want all your customer data being e-mailed around. You wouldn’t want a list of all your best customers being extracted from your system of record and then mailed by somebody who’s leaving the company.

Chatter works securely with Salesforce systems of record—and it can be made to work securely with others, such as software from SAP or Siebel. And it is based around data rather than around messaging.

What does that mean, in practice?

It allows you to follow, in Twitter terms, things in addition to people. I can follow a customer complaint, I can follow an order, I can follow an invoice, rather than just a person. I can choose which events I would like to subscribe to. And groups can be created, to allow project teams to work together. The UI by which the human in the enterprise interacts with information is a stream—curated by my own network.

Why does it make sense in a business for information to be filtered by someone’s network?

People buy from people, people sell to people. There’s something very social about the engagement process by which any sale is carried out. Because it is a social interaction, tools that allow us to reduce the noise of having separate silos are valuable.

What is an example of the kind of problem that your customers say they need Chatter to solve?

“Help me find the things and people that will help me close the sale: Where is the collateral that will help me answer the question the customer has? Who are the people who have the skills to be able to do that?” So you can think of it in terms of search costs being reduced.

If more and more of the work of a company is done in these information streams, where employees pose questions to their networks and get answers in response, then the ways of measuring employee performance might need to change, right?

We publish something called Chatterlytics: who’s most likely to comment back and reply within a given time, whose answers have been seen as valuable. Terms like “relationship capital” or “social capital” or “human capital” start having meaning. We’ve used them for 20 years, but without the tools to measure them or understand them. It’s what we’ve learned from the Facebooks, from the Twitters, from the LinkedIns and the like: okay, now we know something called clout. Not how many followers you have, but also what kinds of followers you have. That’s where “gamification” of the enterprise begins.

I think I see where you’re going—that if people were eventually scored on whether their colleagues saw them as good collaborators, they would be motivated to keep it up, to improve their scores?

There is an altruistic element, [but people confronted with a request for assistance also ask] “If I do this, will it make my job easier? If I do this, will I get peer respect? If I do this, will I get recognition?” These sorts of questions are human. And what we’re now finding is that we could see a world where, for this project, you could say, “You need these three badges.” [Users of the social-networking service Foursquare earn badges for doing certain things, such as visiting a certain bar enough times.]

I am extrapolating, so you understand where it is going, rather than where it is today. [But] we are already working on many of these measures, and we have groups of people focused on where badges would be meaningful.

Talking about making this into a game might provoke some skepticism from people who think that’s a fad, or that business isn’t a game.

The people who push back and say it’s a fad are the people who still don’t understand how Wikipedia could come about, who still don’t understand the sheer volume of literature that says that in prediction markets, play-money markets appear to have very similar degrees of accuracy as real-money markets. Or don’t understand that the new generations coming through are very heavily predicated toward peer respect and recognition.