This is my second summarizing post on the excellent Enterprise Search Summit 2010 conference. See my earlier post for a summary of the workshop day, and first full day of the conference.
Day 2 kicked off with an excellent keynote from Peter Morville of Information Architecture (for the WWW), Ambient Findability and Search Patterns fame. I have seen Peter speak before, and I am a bit of a 'fan boy' of his work, but I will not summarize his keynote here as Peter put his slides up on SlideShare.net, so I will just embed them below:
Next up after Peter was a panel discussion "In search of search" led by Sue Feldman of IDC. The panel consisted of:
- Haroon Suleman, Lead Enterprise Search Architect, Mercer
- Jim Cassella, Manager, Global Application Architecture and Strategy, Colorcon, Inc
- Michael Mills, Management Consultant, Kraft & Kennedy, Inc
This was an excellent discussion, but with a representative from a large professional services firm (Mercer), one from a law firm, and one from a large supplier to the pharmaceutical industry, I felt that they we were definately representing the high end of enterprise search, i.e. industries which "get it" and which make my poor old retail organization pale into insignificance where it comes to information management in general and search particularly. Takeaways from this discussion on running a search procurement and implementation included:
- the tension between user requirements for a single point of access to all information and the reality of 12 major repositories with 15 different search applications (!)
- The problems with security and the new search engines allowing 'un-secured' content to be found in inappropriate circumstances
- The importance of getting a preferred vendor to do a Proof Of Concept with your data (not their own test sets), but acknowledging that you should pay for such a POC
- Sue noted that is not really about "digital natives versus digital immigrants" but its more about the value proposition and ease of use, especially for the very busy senior execs
- Pervasive search analytics - search meets BI
- Social networking in many enterprises is EMAIL ! Therefore be prepared to search it, and 'social graph' it.
- E-discovery is now a bigger industry sector than enterprise search, but work on e-discovery tools soon works its way back into mainstream enterprise search technology
Security came up again and again in this discussion and in the questions asked by the audience. In sophisticated environments, such as law firms where 'security by exception' is the norm, the search vendors tools are often lacking in their ability to provide security trimming of results sets that can cope with the requirements.
Next up was Jeannine Bartlett presenting Early and Associates 'SIX' Metrics frameworks. SIX stands for Search Integration eXperience. I am not going to expand much on this session as its their proprietary methodology - but it was a very interesting session, so if your interested, give them a call !
After some more networking at lunch time, and I have to tell you the networking with fellow practitioners is an excellent reason to attend this conference, my first afternoon session was one being presented by my very last consulting client, Statistics Canada. Kathy and her colleague Andrea presented on the StatsCan journey, on how they have gained senior stakeholder buy-in and executive sponsorship, how they undertook a phased program of analysis and tactical upgrades to their search environment, and how this has paid off for them. Their user community is now reporting much greater 'customer satisfaction' and that is not just based on their small improvements to their Ultraseek environment and their user interface work. It also based on their back end work on sorting out metadata and content collections etc. Check out the StatsCan specialized search tools page and note that they include considerable help content and e-tutorials. Just one question though Kathy - how come Martin got his name on your slides but I didn't ? .......... :-)
My final session of the conference was "Sustainable search at Du Pont" by Alicia Shortlidge. Alicia has been involved with search at her company for ten years, and described herself as search team manager, project manager , search sys admin and developer ! Out of a central App Development Consulting Group team of ten, only Alicia and one colleague are full time dedicated to search, and yet they have built 18 specific 'search applications' on the same search tool platform (my notes seem to suggest Recommind again, but they are a bit scribbled !) which all use the same index. They are basically specific search applications for single major business units, developed to meet special customer requirements (under a charge back model). They also use a Google GSA for their intranet. Currently they have Librarians manually tagging content with metadata, but hope to move to an auto-classification model, freeing up the Librarians to do more effect QA / QC on the automated tagging.
Security was a big issue again, with the use of the search engine as a 'security audit tool' - turning it on and finding stuff in unsecured and inappropriate repositories.
I missed the closing keynote, having to set off for La Guardia, but I have the slides from "The future of Enterprise Search' by Leslie Owens of Forrester. To qoute from her very first slide:
"The broad category of enterprise search is dead. Today's knowledge workers demand role-specific, contextual search, wherever they work."
So, there you go: 'Enterprise search is dead, long live, erm..... enterprise search...?"
No, seriously, I see her point and I understand. Nothing stands still in enterprise informatics, and the nature of enterprise search is morphing. I think Forrester will let me show you a graphic from one of Leslie's slides showing their interpretation of Unified Information Access:
OK, so I like the general concept here, structured data and unstructured content brought together by connections and processes of search and retrieval. However I will have to do some more reading of Forrester's content before commenting further, but what do you think of this ? Use the comments section to let me know.
So, in summary Enterprise Search Summit 2010 was an excellent conference with good sessions and even more so, some excellent networking opportunities with very intelligent practitioners; Ed, David, Jami and Christina, amongst others, thanks for improving my knowledge ! Lets not forget the vendors who attended, who were all very happy to discuss their products and search in general. I have to say, based on the other attendee's I spoke to, Microsoft if your reading, you missed an opportunity by not having any FAST people attending (although there seemed to be a surfeit of ex-FAST people !).
If you interested in the subjects around enterprise search and missed last weeks conference, take a look at the Enterprise Search Summit Fall, taking place in mid November in Washington DC, I can thoroughly recommend it :-)
I am 'back at the ranch' after an excellent trip to New York for the Enterprise Search Summit 2010 which was a lot of fun and very informative.
A big thank you to my friend and colleague Martin White of Intranet Focus Ltd, a true search expert, who got me involved, and to Michelle Manafy, editor of eContent Magazine and conference chair (and her colleagues) for her excellent job organizing everything.
So even though I had to go buy an old fashioned note pad, due to failure of both my Asus Netbook batteries to take a charge, I made copious notes, so this could be a long post........ Actually I won't give too much detail, as I am sure Michelle would rather you sign up for ESS Fall, the autumnal version of the conference being held in Washington D.C. !
I will start with some observations, before diving into the details:
- Security is a really big deal for enterprise search implementers !
- There is no "one ring to rule them all" in enterprise search - most attendee's seem to have more than one search engine / application, and they are OK with this
- Google Search Appliances and SharePoint 2007 are the poor children compared to the products from more 'established' enterprise search technology vendors
I will examine each of these issues as we work through the conference chronologically. I will break my summary into two posts to keep it more readable. I ran a workshop on Monday morning, on making the most of the search facilities of SharePoint 2007. I should have been a double act between Martin and I, but he could not make it due to a family emergency. I had a great audience, who were very participative and happy to share their experiences, and there was some serious expertise in the group, so a great big thanks to you all !
In the afternoon I attended a deep dive into SharePoint 2010 and FAST 2010, and it was a long and very useful workshop provided by:
- Miles Kehoe, President, New Idea Engineering, Inc
- Mark L. Bennett, CTO, Software Engineering, New Idea Engineering, Inc.
- Jeff Fried, Senior Product Manager, Microsoft
- Natalya Voskresenskaya, Co-founder, Arcovis LLC
They have all just co-operated on a new book: Professional Microsoft FAST Search: Customizing, Designing, and Deploying Search for SharePoint 2010 and Internet Servers and they are a thoroughly great set of people.
Tuesday, the first day of the full conference saw excellent keynotes from Dr Marti Hearst from UC Berkley and David White a senior analyst at Aberdeen Group. Marti put her presentation on SlideShare, so I have embedded it below:
What I liked about Marti's message was that emotional response to User Interfaces is important, and that small details matter when your trying to design an "interrupt free engagement". She also made the excellent point that humans are very social beings, and computers are not, so "don't try to personalize search, socialize it". An excellent keynote.
Google's Rajat Mukherjee had a quick 15 minute slot between the keynotes, and although his focus was obviously on Google's products, a very nice point, which seemed to surprise some attendees, was one about organizational or corporate culture. Within Google, everything is considered open and accessible, unless there is a very good reason to lock it down, so by default all content is shared with everyone. When you consider that they do a lot of 'dog fooding' (i.e. using their own products) that makes sense when you think about the features of Google Docs and Wave.
David Whites keynote was based on Aberdeen group survey of enterprise search implementations and presented in a 'best in class' versus, well lets be polite and say "less than optimal". One thing I would like to draw out from this presentation is the following, all the leaders in implementing best in class enterprise search have a dedicated team. Got that ? Let me shout it again, a dedicated team !
First of the breakout sessions for me was Searching the Past, as session on search and Digital Preservation by Christine Maxwell. I have some experience in long term digital preservation while working for the Open University, and with ESA and NASA on the Mars Express / Beagle 2 space missions, so I found it fascinating ! I will simply pass on some links to interesting projects that Christine brought to our attention:
My next session was 'Remembering and using the information you have got' by Ayelette Robinson. My favorite soundbite from Ayelette was "better business through better information access" - yep, amen to that sister..... :-) Ayelette is in Knowledge Management for a law firm, where search is seen as a key decision support tool. Her company uses Recommind as their search tool, and is also using Lexis-Nexis tools for auto-classification (generating metadata from the contents of a document, and classifying the document appropriately). They are big fans of the auto-classification and it is working well for them (but my take on this is that lawyers have very structured documents).
Next up was Avi Rappoport of SearchTools.com on enterprise search architectures specifically Federated versus Aggregated search. If you don't know, Federated (also called 'brokered') search sends the queries to existing search engines / indexes, whereas Aggregated search attempts to construct a single large index, even though it maybe indexing multiple content sources / systems. As per usual, there is not one size fits all, and different approaches fit different contexts. You can always drop Avi a line if you want to know more on this one.
'Improving Findability behind the firewall' was next on my agenda, presented by Bob Boeri, and ECM consultant with considerable expertise (!). Bob sees enterprise search as an element of Enterprise Content Management, and noted the importance of standards, for metadata, taxonomy and other elements of content management strategies, and the positive impact they can have on both precision and recall of your search results.
My last session was a three handed affair, so my notes are bit all over the place, but we can break it down to:
- The Tribune Company (as in the news papers) using tools for Ontology Management
- Raytion on search analytics
- Vivisimo on 'information optimization: efficient searching reduces expert training costs"
Wow, so that's the first day written up for you. Just to summarize, many of the attendee's of my workshop and many other presenters touched on the complexity of their search environments, with multiple search engines / products in use, and nearly everyone noted some issues with security - be it 'early binding' / 'late binding' or security filtering against LDAP directories and how to present results to users. More on these topics, and notes from day 2 in the next posting.
OK, so I like the general concept here, structured data and unstructured content brought together by connections and processes of search and retrieval. However I will have to do some more reading of Forrester's content before commenting further, but what do you think of this ? Use the comments section to let me know.
