Curation in the Enterprise: Actionable information


In Steve Rosenbaum’s Curation Nation, Esther Dyson quotes Bill Gates as saying “The future of search is verbs”. Esther goes on to say that nobody really looks for something per se, they look for things in order to do something.


When Marc Benioff elucidates his vision for the Social Enterprise, he stresses the importance of having information you can act on. So for the last few months I’ve been spending time thinking about what makes information actionable, and whether the social enterprise helps or hinders in this regard.

For information to be actionable, it must have at least four characteristics:

It must be accurate, and verifiably so
It must be timely
It must be comprehensive
It must be comprehensible

These are necessary but not sufficient conditions; to make information truly actionable, the information must be accompanied by tools that allow you to act on the information. In a perfect world, you would also have tools that allow you to monitor the effect of your actions and to receive feedback. But at the very least, you need information that is accurate, timely, comprehensive and comprehensible, with some ability to act on that information straightaway.

In order to understand the role of tools here, let me give you an example of information that is not conveniently actionable: the voicemail. While at BT I spent a lot of time with Martin Geddes (I have a lot of time for him), exploring views on how to reduce latency, particularly “human” latency, from communications-enabled business processes. And Martin’s favourite example in this space was the voicemail. Someone leaves you a message. The information in that message is all dressed up with nowhere to go, because all you can do is to take down, transcribe or otherwise memorise it, then you have to swivel-chair into some other mode in order to do something with the message. So Martin was asking why we don’t design voicemail differently, so that the scenario changed. Like this “Hi there, this is Basil. I’m sorry you weren’t in when I tried to deliver your bicycle at 2pm today as promised. If you’d like to speak to me to arrange another delivery at no extra charge press 1. If you’d like me to return as soon as possible today press 2. If you’d like to cancel the order, or to speak to an agent about it, press 3.”

I thought Martin was on to something, but between us we couldn’t make it happen. But the principles remain sound. Communications-enabling business processes are part of the future, and will form part of the social enterprise. Whe designing these processes, it makes sense to look at where the latencies come from, and to minimise, perhaps even eradicate, them. Making information actionable is a key step.

So let’s look at how the Social Enterprise makes information actionable.

Accuracy and veracity

In a social enterprise, the risk of inaccurate or false information is reduced in two ways.

First, because of the transparency implicit in the social enterprise, we get Linus’s Law in operation: given enough eyeballs, all bugs are shallow. So you get the value of inspection, something that security experts have known for a while. There are two provisos to bear in mind: one, Linus’s Law exhibits network effects, so you need critical mass before real value is derived; two, not all the information in an enterprise will be made transparent; in social enterprises, as in families, some things are more private, more confidential, than others.

The power of inspection is therefore one way to make information more actionable, reducing the likelihood of error or falsehood. This power is further amplified by the existence of a second phenomenon, the multidisciplinary team. Social enterprises are fundamentally non-hierarchical in practice, given that everyone and everything is a node on the network. So the historical concept of silo-ed business units is weakened, with considerable “lateral” activity, not just within the enterprise but beyond it, involving customers and the supply web. This predilection towards multidisciplinary teams engendered by the social enterprise also goes a long way towards reducing the risk of inaccuracy, regardless of whether that inaccuracy is an act of commission or omission.

[An aside. I have never stopped marvelling at how people believe something just because it was spat out by Excel and made its way into PowerPoint, as if these were robust systems incapable of being compromised. One born every minute, as Phineas T used to say.]


“Venkat” Venkatraman, an old friend and Boston professor, told me many years ago that businesses used to be hierarchies of product and customer, and that they were morphing into networks of relationships and capabilities. I’m paraphrasing him, but the core idea is significant, and forms the heart of the social enterprise.

One aspect of the social enterprise that is underrated is its nonlinearity. Social enterprises don’t follow strait-laced linear “serial” processes. Instead, aided and abetted by network effects, and no longer straitjacketed by silo thinking, information flows laterally from unit to unit, within the enterprise, beyond the enterprise. Massively parallel. This makes the whole organism work faster, since the traditional up-one-hierarchy-suitably-filtered and then down-another-hierarchy-also-suitably-filtered approach cannot hold up proceedings.

Besides, the very concept of everything in the network being a node increases substitutability. You no longer have to wait because someone in the serial process is busy or on vacation or something, the social enterprise has no time for all that.


As discussed in an earlier post, the construct of a social enterprise facilitates identification of true domain experts rather than signal boosters, sidlers, spreadsheet jocks or suchlike. It goes further, the social enterprise actually helps ensure that the domain expert can participate where and when required, without worrying about organisation charts and titles and roles.

Consequently it is more likely that the information is generated by a domain expert, close to the action as it were. If the original information was not actually generated by the domain expert, it is reasonably likely that he or she will annotate the info, comment on it, tag it, augment it in some way.

The ease with which people from other disciplines, cultures, contexts can participate also means you get other valuable, hitherto unintended, benefits. You say tomahto and I say tomayto. Cultural distinctions can be made and shared, translations are more likely, the availability of commentary from other disciplines helps increase the likelihood of comprehension.

Comprehensiveness and context

Historically, there were two types of systems: systems of record and systems of engagement. Computing departments built systems of record, substitutes for the books and records of the firm, robust, secure, safe, hard to attack, and consequently hard to use as well. Communications units built systems of engagement, which, while reliable, were open to all, with simple security needs and exhibiting network effects. Convergence, the coming together of computing and communications, had the unintended consequence of bringing these disparate design and security philosophies together for the first time.

People working on the Social Enterprise at pretty much pioneered the work on making this work; they were used to delivering services on a multi tenant utility model as their core business for over a decade, so this was nothing new.

As a result, the social enterprise is built with a deep understanding of entitlements and permissions; as long as you have the right to see it, you can get to the source information, the order in question, the compan profile, the complaint, whatever. So the context of the information is carried with the content. You can drill down as needed, or see the summary. In the social enterprise, summaries are just convenient representations, the source data is never far away.

This increases comprehensiveness. But it’s not all. The presence of the extended network, the nonlinear participation of customers, partners and staff, the crowd sourced ability to inspect and reduce error, the support for tags and folksonomies, every one of the characteristics I have described earlier today accelerates and enhances comprehensiveness.

The social enterprise is therefore designed to provide more accurate, verifiable information, faster and more effectively when compared with the traditional enterprise, easier to understand, more complete and in context.

The social enterprise provides actionable information. But as I stated earlier, actionable information alone is not enough, you need the tools to do something with that information. Which will be the subject of my next post.

Comments welcome as ever.

4 thoughts on “Curation in the Enterprise: Actionable information”

  1. Great reflections on this topic which is very close subject to what we’re doing- building an enterprise content curation platform (Eqentia).

    Context comes first, content 2nd. When context is well defined, the right content flows. And with social, the right content will lead you to the right people. That part is the actionable part if the context is strongly aligned with your interests.

    We also believe that Curation is a means to an end, and not the end itself as being portrayed by several pure curation tools out there, especially the consumer-focused ones. The word “Curation” has started to lose its true meaning because it’s being overused, misused and abused. We should always ask – what is the benefit we’re getting out of curation?

    Looking forward to your follow-on pertaining to actionable tools.

  2. If we’re all nodes with permissions to curate content, content is the new hierarchy, and contains permissions. Then what tools can handle permissions. Do tools map back to super users and domain experts (oh my, I’m going in circles).

  3. What about trust? Information with all the 4 characterstics that you listed – without trust would be underused or misused or lead to a wrong conclusion?

  4. Prakash, verifiable accuracy reinforces trust. Not just the accuracy, but the ability to verify it. The ability to get to the source, check source credentials where relevant, understand source data and context.

Let me know what you think