Unified Information Access Blog
Welcome to Attivio's Unified Information Access Blog. Join us for discussions on topics ranging from enterprise search solutions, information access insights, Agile software development methodology to programming with Java. We hope you'll find the articles informative and participate in the discussions by leaving a comment.
Looking deeper into a recent report on findability and also to expand on my previous post on the topic:
Surprisingly, survey respondents report that "multiple search engines are the norm in most organizations". To put a finer point on it: "There is a separate search solution for e-mail, Web content, wikis, Blogs, ERP systems, CRM systems, intranets, File shares, and so on." Less surprising is that this irritates users.How has enterprise search fallen into the same trap as the database and become yet-another-silo? My two cents worth is that the legacy players are unable to escape their origins and are, in fact really only good at one thing - namely, whatever they started out doing. To wit:
| Origins | Strengths | Challenges |
| Web Search | Ingesting web, CMS content or simple database content, query performance, reporting | Relevancy across multiple sources/types of content, security, access to enterprise content |
| Intelligence | Recall on a single, focused set of content (like patents) | Performance, precision, use on diverse data sets |
| eCommerce | Front-end integration, management of the site and categories, reporting | Performance, scalability, relevency, use on multiple silos, freshness (update frequency) |
| Linguistics | Whatever their particular focus is, e.g. question answering, clustering | Relevancy, especially across multiple sources/types of content, security, ingestion performance |
| Standards | Ingesting and manipulating the standard, integrating the standard with other systems | Dealing with content/interfaces that don’t match the standard |
One thing you'd expect to fall out of this is greater interest in federated search, i.e. having one search box that reaches out to the silos and aggregates the responses. And indeed, 60% of member and 30% of non-member respondents rated this capability "imperative" and/or "significant". (Members have supposedly "exhibited greater levels of familiarity with many technologies".) Interestingly, googling federated search turns up not one paid listing by a legacy enterprise search company! Could it be these companies aren't interested in being part of a cross-silo solution?
Another answer is that federated search has simply never delivered on it's promise; it's quick and easy but doesn't satisfy - mostly, in my view, because it is hard enough to deal with the real differences in silos of content within a single software architecture, let alone across multiple search architectures.
Both members and non-members of the AIIM Association seem to agree on certain things. (If you are following along at home, look at the charts on pages 17 and 18). For example:
• Social network features like voting and bookmarking draw poor importance ratings. This seems likely tied to the overall poor ratings for findability, i.e. what's the point of voting on poor results in a multi-silo environment?
• Behavioral search is similarly given poor interest ratings, likely for the same reasons.
• Commonly available features like concept search, text analytics, entity extraction, clustering, etc, get middle of the pack importance ratings. (Non-members are even a little tougher on some of these.) I believe this is because these features are either over-emphasized (as the "answer") or simply don't work in a customer's environment. Many of these features are built on very specific sets of data like web pages or news and may require significant tuning and/or integration in the enterprise.
One big disagreement is around the value of metadata. Members rate it very highly - ~75% call it "imperative" or "significant", but only ~37% of non-members agree. My guess here is that members work at companies that have greater investments in information access, and as a result have access to more and better metadata. This is a clear call to action for legacy enterprise players: helping to find, field and use metadata is still an important task for relatively new users.
I'll have a final word on the findability study in my next post... This e-mail address is being protected from spambots. You need JavaScript enabled to view it if you have any questions or comments in the meantime.

Recent Posts
UIA: Boosting Customer Loyalty (Or How NOT to Annoy Your Customers)
Clash of the Titans: Price and Performance for Data Warehousing and Analytics
Enriching Unified Information Stores with Text Analytics
Introducing Key Phrases
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
