Web 2.0, Social Media and Decision Management

One of my favorite BI bloggers, Cyril Brookes, had an interesting post today – What Does Web 2.0 Really Mean for BI’s Future?. He starts by pointing out that “Successful BI enables the executives and professionals to make better decisions” to which I would add that successful enterprise decision management enables the systems in an organization to make better decisions. I would agree with him, though when he says that this “is a numbers play; presenting them, assisting their assessment, finding issues hidden in them, empowering action to resolve issues, and sometimes automating them…There will always be numbers in corporate BI”.

Cyril goes on to point out that there are various aspects to “Web 2.0” including finding people with similar interests, sharing experiences, and creating knowledge from tacit resources inside people’s heads. Cyril discusses how these might complement BI but how might they complement enterprise decision management? There are two sides to this – how does EDM use social media outside the firewall and how does it use it inside the firewall.

Social media outside the firewall allows your customers, suppliers, partners, members of the public, whoever add their knowledge and opinion to yours. Once they have done this the question, from an EDM perspective, is what can you do with this information?

• You can mine it for insight
  The growth in text mining technology makes this ever more practical. You can also make it easy for people to build explicit links within their comments, allowing you to mine those linkages directly. For instance, you could use the product links people include along with text analysis of the “tone” of their note to see what cross-sell or up-sell you might use. This is harder than using structured data but increasingly practical. You might also find that the mere fact of participating in social media is a significant data point in customer segmentation models.
• You can enhance results with it
  When you give a person a decision, whether that person is a call center representative or the user of your website, you could include relevant entries from social media with it. For instance, if you recommend a cross-sell you could include positive comments about the product made by other customers.
• You can automate responses to it
  Again, using text analysis, you could decide that certain kinds of comments or interactions or contributions are indicators that outreach is called for and automate email or phone or social media-based responses inviting prolific posters to participate in some special program, awarding loyalty program points based on interactions etc.

Social media inside the firewall allows your staff to collaborate in new ways that are relevant to EDM.

• Manage feedback on the results of a decision service
  Making it easy for those who deliver the results of a decision service to comment on how it was received might give you useful feedback on the results. Particularly when you are using adaptive control techniques to try multiple strategies, this could be very useful as the numbers might point to a particular strategy but your comments might show that it had a very negative impression on those who declined it, something it may be hard to put a dollar value on.
• Discuss assumptions in rules and models
  In the context of decision management it could be very useful to allow managers and analysts to discuss the assumptions inherent in specific rules and models using social media. The “wisdom of crowds” approach might allow you to tap into the tacit knowledge of more workers than the format rule or model review process could.
• Exception handling
  Documenting and discussing how an exception was handled and how that worked can be useful as you try and gradually automate more and more decisions. Understanding how a class of exception could be handled and what the consequences were is critical to getting the more complex situations automated effectively.
• Mashup reporting and decision management
  Allowing users to combine the results of decision automation with reports and search results that clarify or extend the decision as well as peer opinion and discussion might help with adoption and with usage. Being able to see the numbers that should drive a change in the rules, the discussion of how those rules were developed and the rule management environment itself in one place can only be an effective tool.

While it might seem that EDM, with its focus on automation has little to do with people-centric web 2.0 technologies, clearly there is a role for them to be used together.

A final comment – I agree with Cyril that human nature is not really going to change even when the current generation of workers is replaced by those growing up on Facebook etc. Like him I also think “they will be much more ready to participate in corporate collaborative activity” so making more of these kinds of things practical in terms of having a critical mass. I also think that this audience will be much more willing to “mash” things together than their parents are.

“May you live in interesting times”