The 451 Group's Client Conference - Information Management Panel

We've just concluded a lively panel discussion on the "Changing Face of Information Management" at the 451 Group's client conference here in Boston. Nick Patience, Managing Analyst and Research Director for Information Management lead off with a brief presentation that covered four major trends he sees: sentiment analysis, enterprise search, the current credit crisis and eDiscovery in general. Among the highlights:

• Sentiment analysis is a crowded space with many new players; much of the current focus is on social networks

• Enterprise search is going to have to evolve; the question is "how?"

• Information management will change in the enterprise as a result of the credit crisis; again, the question is "how?"

• eDiscovery is starting to separate into high and low-end markets, and will see signifiant growth & investment in the coming years

Following this presentation Nick sat down with Nicole Eagan, Chief Marketing Officer at Autonomy; Stephen D. Whetstone, Vice President of Client Development and Strategy at Stratify; and myself. The highlights, from my perspective:

• eDiscovery: Stephen predicts more litigation then ever, especially in the wake of the recent financial issues. He sees the enterprise working hard to reduce the cost of capturing and managing information. I pointed out that cloud computing was one possible application for that, assuming that security issues can be mitigated. Litigation is an occasional problem (for most companies) and so spinning up a cloud of servers to process and analyze all of the data seems like a prudent approach. There was also discussion of the importance of using policies to retain data correctly, etc. In response to a customer question about being more 'proactive' I mentioned that retention via policy is just one step; it's more important to understand what activities lead to litigation and focus on identifying those ahead of time.

• The Credit Crunch: more of the same. Nick asked what we expect to see in terms of additional regulation. I said that I expected a lot more, including specific requirements and possibly official government presence within certain high-risk enterprises. Stephen seemed to agree. He noted that Sarbanes-Oxley defines responsibility, but is vague about execution. That lead to the Federal Rules of Civil Procedure update of 2005. He expects another logical progression to follow, with more specific responsibilities and processes being part of it.

• Enterprise Search: I noted that, despite the typical marketing message, most enterprise search solutions are, in fact, limited to one or two silos. As a result, the enterprise now has more silos than ever and needs to start bringing them together to create competitive advantage. I also pointed out that search needs to start connecting systems together instead of just serving end-users. Nicole noted that search is just "one of 200" features their collection of software products offer and implied that it isn't more important than the application itself. One surprising turn: after years of passionate condemnation of federated search, Autonomy is apparently now in favor of it - for structured data, at least, which she seemed to indicate causes problems for their technology. This is no surprise, as they claim only to help understand the "other 80%" of data, i.e. the unstructured data.

We didn't talk about Sentiment analysis, but I would have made the following points if we had:

• Document level sentiment, which is currently the most common implementation, is not very useful - for many reasons, ranging from the fact that many articles are not about one thing (e.g. "Best Cars of 2008") to the use of sarcasm.

• Sentiment focused on entities is much more valuable - isn't the key to know what someone thinks of X? Attivio will have more to say about this next year...

• The popularity of sentiment is at least partly because it helps "structure the unstructured", and the fact of the matter is, users appreciate structure, especially when it is used to drill into a larger set of data (structured or unstructured).

Anyway, it was an excellent panel and I thank Nick and all at The 451 Group for asking me to speak.

A final word on the conference tomorrow, after the final sessions.

TrackBack URI for this entry

Comments (0)