Category: Search

Ten intranet search case studies

It is very difficult to track down good case studies of intranet search implementations that when eight come along together it was well worth spending a couple of hours to read through them, They are published in the Nielsen Norman Group 2017 Intranet Design Annual. I will review this report from an intranet perspective in due course, but for now I just want to look at some of the highlights, and low lights, from the way ten very different organisations have provided intranet search. In the introduction to the report the authors note that search is starting to be taken seriously (my words, not theirs) and that is very good news.

The stand-out case study is that from IBM. The IBM intranet is on a vast scale in size, reach and user numbers, and search has always played an important role. The description of the way in which the Blue Pages people directory has evolved is worth buying the report for on its own account, and then hand delivering copies of the pages to your HR Director and your CEO. Of course a huge amount of research and development went into the development of the directory but the result is probably the definitive people/expertise directory for you to benchmark your own vision and achievement. If you want to learn more about the technical background to the Blue Pages track down the papers authored by Ido Guy when he was at IBM Research Haifa. However in the case of IBM it’s not just about the technology but also about vision, governance and user research. The technology is in fact bespoke and built on the Amazon Web Services platform (interesting!) but IBM Watson is about to be implemented. One of my current clients is about to use IBM Watson and I will be very interested to see how well it works as a people and expertise finder.

Other companies which have an interesting story to tell include Bank of America, Encana, Goodwill Industries, Jet Blue and the Kerry Group. The range of search software is wide, including Google Search Appliance (Bank of America), SharePoint 2010 (Goodwill Industries), PostgreSQL (Jet Blue), SharePoint 2013 (Kerry Group) Solr (Latvian Railways) IDOL 7.5 (Santander) and Office 365 (Tourism New Zealand). Which means that Bank of America and Goodwill Industries are in for an interesting 2017 and maybe Santander needs a Plan B.

The not-so good news is that when you look down the teams running the intranet there is never a mention of who is taking responsibility for search, and in particular search analytics, even though every other aspect of intranet management is listed out. This probably means that IT are taking the search lead but IT need to be able to call on user feedback and business requirements to put the search log analytics in context – which of course assumes that IT are actually running any search analytics.

Martin White

So how does search work?

When I wrote the first edition of Enterprise Search in 2012 I only provided enough of an outline about the inner workings of search to support the recommendations I was making in the rest of the book. Comments from many readers encouraged me to write about what I would call the technology of search in more detail in the 2015 2nd edition. In many respects writing about the technology was the most difficult section of the book as I needed to work through how best to present the technology in a way that made sense to a business manager. The phrase ‘the technology of search’ is a misnomer, because it is actually about the mathematics of search as delivered in some quite sophisticated software programs, often using a blend of computational linguistics and applied probability theory. My contention has always been that a search team has to understand much more about how search works than the teams supporting any other enterprise application as most of these are build around a relational database plus a lot of extras. If enterprise search is complicated then the step up to text mining is quite substantial. The subject is covered in depth by ChengXiang Zhai and Sean Massung and Deep Text is a very good introduction to the topic by Tom Reamy.

Over the last few months I have noted a number of blogs that provide very good descriptions about how search works. These include a series by Daniel Tunkelang published on the Query Understanding website and by Maish Nishani on the Olasearch website. Patrick Lambe has written a very good 15 page summary of search technology in Behind the Curtain: Understanding the Search and Discovery Technology Stack on the Green Chameleon website. In addition Charlie Hull and Udo Kruschwitz are in the process of writing a book on this topic which should be available from Now Publishing in 2017. If you want a really deep dive into the technology then the recently published book Text Data Management and Analysis by Professor ChengXiang Zhai and Sean Massung runs to 500 pages, which gives you a sense of how much detail there really is!

The importance of understanding what is going on under the bonnet has been emphasised by two recent blog posts, one from Mark D. Anderson and another from Marcel Meth. Both these blogs illustrate that sorting out problems in search applications requires a significant amount of technical knowledge and a lot of patient detective work. First you need to know when there is a problem, because search does not break, it just fails to return the results you were expecting, and that requires you to know what you were expecting. Second you need to work out where in the complex series of search processes and sub-processes the solution to the problem may lie. One of the attractions of the Google ESA was that the underlying technology was behind a very robust security firewall so there was no point in having technical expertise on the search team because there was nothing they could do except dust the server casing from time to time. That is now just history as the Google ESA begins to fade away and replacing it is going to need more than rack space and a feather duster.

All the signs are that search, in its various forms, is gradually being recognised for the critical role it plays in enhancing business performance, especially in supporting decision making and identifying expertise and knowledge. Taking advantage of search does require an investment in understanding how it does what it does. That will also stand you in good stead for the arrival of artificial intelligence. I would argue that it is even more important to understand the principles of AI than of search because you are going to need to trust the black box you have invested in. A good place to start is the 3rd edition of Artificial Intelligence – a Modern Approach by Stuart Russell and Peter Norvig. This is now available in paperback but runs to over 1000 pages. So start with search and be well prepared for an interesting future.

Martin White




Eight good books on search management

Over the vacation I reorganised my office, including adding a bookcase specifically for the 60 or so books I have on various aspects of information retrieval and search. In the process I updated the list of books on the Enterprise Search book website, highlighting eight which I regard as a core collection for search managers. These are listed here for convenience. All the books have been reviewed on the Intranet Focus blog and some have a dedicated website. They cover the spectrum from a deep dive into the technology, the optimisation the user experience, search evaluation and text analytics and mining. With a couple of exceptions they have all ben published in the last two years, which indicates the level of interest in the subject. Publishers publish to make a profit!

Deep Text 
Tom Reamy, 2016, Information Today (Review)

Designing the Search Experience.
Tony Russell-Rose and Tyler Tate, 2012 . (book website) (Review)

Enterprise Search
Martin White, 2nd Edition 2015. O’Reilly Media (book website)

Interactions with Search Systems
Ryen White 2016. Cambridge University Press (Review)

Relevant Search
Doug Turnbull and John Berryman. 2015. Manning Publications. (book website) (Review)

Search Analytics For Your Site.
Louis Rosenfeld. 2011. Rosenfeld Media (Review)

Text Data Management and Analysis
Chengxiang Zhai and Sean Massung. 2016. ACM Books (Review)

The Inquiring Organisation
Chun Wei Choo.  2015. Oxford University Press.  (Review)

I am also looking forward to a new book on the underlying technology of enterprise search from Charlie Hull (Flax) and Professor Udo Kruschwitz (University of Essex), to be published by Now Publishers

Martin White

Website search – “Must do better!”

Although most of my projects are internal enterprise search engagements I have just been assessing the website search of a major consulting firm. One of the elements of the project was to compare the website search against those of other professional services firms, The results were very depressing. In total I looked at around 20 firms, and the consistent message was that website search is not important enough to pay any attention to, even though to me the quality if the search experience is an indicator of how important the firm regards attracting new clients. This is because for a professional services firm the major challenge in a highly competitive market has to be finding new clients or new business from existing clients.

As I look through my notes there are some common problems with the sites.

  • Almost invisible search boxes tucked away on the edge of the page
  • Search box either too short or opens up Google-like on a new page with no warning or search help notes
  • Very little date information on the results and no ability to sort by date to show off how much recent information the firm had available
  • No highlighting of query terms in the results summaries
  • Filters and facets in abundance with little consideration of their value in focusing in on highly relevant information
  • A common filter is file format, which assumes that “web pages” (for example) are of especial importance to search users
  • Poor expertise search. All too often “people search” is primarily a name search feature
  • Very limited use of promoted content to highlight core firm capabilities

Those are of course common to many websites but when I consider the amount of investment in the design of the website and the consultant hours spent on writing briefing notes etc for publication on the website I do wonder why there is invariably not the slightest sense of user centricity about the site and about the role that search should play vis-a-vis the navigation.

A specific problem I came up against was the way in which query terms which are noun adjuncts are handled. For example [risk management], [corporate finance] and [interim management]. Most professional services firms (and it is the same with universities) use these two word noun phrases quite extensively. Of course in a perfect world site visitors would use quotation marks as a proximity function. However my experience with some university websites suggests that search users may be assuming that they can query [chemical engineering] and the search engine will be smart enough either to check “Did you mean School of Chemical Engineering?” or to have promoted content that is specific to the School of Chemical Engineering just in case. These are the sort of issues that user testing and high-quality search log analysis will uncover, but in website search and in enterprise search the lack of commitment to search evaluation is usually very obvious.

As well as the wealth of material on the Nielsen Norman Group site (though it’s not the best example of site search!) Chapter 9 in Information Architecture for the Web and Beyond is a very well-written 50page overview of the business case for website search and implementation good practice.

Martin White

Could Google transform enterprise mobile search?

The presentations at the 2016 Search Solutions conference, organised by the Information Retrieval Specialist Group of the British Computer Society on 30 November, were uniformly excellent. For me the most interesting was given by Dr. Qin Yin from the Google Research Laboratories in Zurich. To explain why I need to provide some context. In general the enterprise search experience on a mobile smartphone is poor. As the desktop UI has become more complex with multiple filters and facets it has also become more difficult to provide a sensible UI for a smartphone. Responsive design is not at option. There have been attempts over the years to develop mobile enterprise search UIs.  Autonomy offered one briefly and Sinequa does have a mobile solution but I remain unconvinced of the value of using a smartphone as a terminal device for enterprise search applications. It may be possible but is it desirable?  Another search product that is now largely history is desktop search, overtaken by improved (in relative terms!) Office search and the challenge of indexing (in my case) 10GB of stored data.

Qin Yin’s paper was about the work that Google has been undertaking on offering users the ability to search content that is already on their smartphone, in effect providing the mobile equivalent of desktop search. Now I know that Apple offer the Spotlight search through Siri but that is not a good experience in my view. Google is now offering the initial versions of what it describes as App Indexing, using its Firestorm suite. In effect you can now search through content you have downloaded into apps on your smartphone. The index is held within the smartphone but is revised periodically when (rather like with software uploads) from Google when there is sufficient bandwidth. There are some case studies on the Firestorm site.

Now for another bit of context. People tend to want their internal enterprise search to be as good as Google. We know that is technically not possible but at least it provides a default benchmark for the user experience. What occurred to me is that as this mobile internal search improves users will start to say “why can’t I have the Google App search on internal content”. I discussed this with Qin Yin after her presentation and the answer is that in principle you could, especially as Google ramps up its enterprise cloud offering. This would give the user effective access to people and expertise directories, policies and task applications, especially where the mobile desktop is getting rather crowded and you just can’t find the app with the information you need. Offline access would also be facilitated.

Now I may be off track with this but I think that as the application develops, along with Google’s enterprise cloud service, there could be a new dimension to enterprise mobile search. Although Apple have something that is sort of similar in concept  (but certainly not in execution) it is not in the enterprise information services business at all. I would be the first to admit that some of the systems architecture diagrams and descriptions from Qin Yin passed me by.  Mobile app development is not among my skill sets and I have to say that it was only half-way through the paper that the enterprise angle occurred to me and I transformed into a very active listener.  Even so I would recommend that you keep monitoring the way in which Google develops this approach so that you can be prepared to respond to the “Why can’t our mobile enterprise search be like Google?” question.

Martin White


Agnes, Brian, Ellen, Jane, James, Janus, Kurt, Kristian, Mark, Michael, Paul, Sam, Susan, Tony and Wedge

If nothing else you will have to admit it’s an unusual title for a blog post. If you are a relative newcomer to the intranet community you may be unaware of the role that these people play in supporting the exchange of knowledge and good practice. Without exception the major consulting companies pay little attention to topics such as intranets, information management and search. Some do offer advice on social networking and collaboration but at a level that is targeted at senior managers who are probably the last people to network socially and collaborate. As I was writing reviews of outstanding reports from Jane McConnell and Sam Marshall yesterday their commitment to the wider community was very obvious. This post lists some of the people who in various ways and for many years have transformed our understanding of intranets, team working and digital workplaces through publishing reports and promulgating good practice and who have to make a living whilst doing so.

  • Agnes Molnar is an enterprise search evangelist with a very good knowledge of SharePoint search
  • Ellen van Aken curates a collection of 300 intranet promotion videos alongside her consulting work
  • Jane McConnell understands digital workplaces better than anyone else and publishes and annual survey of progress
  • James Robertson writes books, runs workshops and conferences, gives out awards and challenges conventional wisdom
  • Janus Boye runs communities of practice in Europe and North America and an annual conference in Aarhus
  • Kurt Kragh Sorenson also offers communities of practice and runs the IntraTeam event in Copenhagen
  • Kristian Norling is developing an excellent range of books and is the Swedish representative of IntraTeam
  • Mark Morell focuses on intranet governance as an author and consultant
  • Michael Sampson writes books and blogs about all aspects of collaboration and digital workplaces
  • Paul Miller set up the Intranet Benchmarking Forum and transformed it into the Digital Workplace Group
  • Sam Marshall publishes reports of an exceptionally high quality
  • Susan Hanley writes blogs and books on SharePoint with a strong intranet and portal focus
  • Tony Byrne sets the standard in assessing the performance of digital applications
  • Wedge Black and Brian Lamb are the entrepreneurs behind the Intranet Now conference

The Intranetizen team also deserve recognition.

Any list like this runs the risk of missing someone obvious. If you feel you are that person please let me know.

Martin White

Migrating from Google GSA – the technology is not the core issue

I am greatly indebted to Search Technologies for publishing the results of a survey of clients and other contacts about attitudes to Google GSA migration. The cessation of GSA supply and support hit the headlines back in February this year. I was not the only consultant and vendor to sense a market opportunity. After the initial flurry of activity everything went quiet but I suspect that this survey and an associated e-book will do much to alert GSA customers of the need to put migration planning on their 2017 objectives.

It is worth looking at some of the outcomes of the survey. The first is that 62% of implementations were out of the box and a further 38% had some degree of customisation. One of the important attributes of the GSA was that all it needed to up and working was some rack space and an IT manager who knew (or could find out) where to point it. Maintenance requirements were minimal. It therefore does not surprise me that only 22% of respondents have a migration plan and 72% would like to have a plan, but do not know where to start. Of those who had a plan (I assume) only 30% were planning an on-premise replacement, with 23% planning to wait to move to the Google cloud search solution in due course. I was not surprised to see that most respondents were looking at open source options. The associated e-book is a good summary of the technical issues that need to be considered when developing a migration plan, with a good focus on connectors, content preparation and security management.  However I was very surprised to note the recommendation to “Make sure your in-house IT staff has the bandwidth and skill sets needed to conduct a thorough assessment of the elements above and develop a detailed plan for transitioning from the Google Search Appliance to a new search engine.”

If an organisation has a Google GSA the business justification was usually that it did not need any search skills to manage it. The extent to which you could tweak the ranking was pretty limited. So how would the IT team have developed the skills needed to undertake not just a technical audit but the user requirements analysis that is essential to a successful search implementation? In a recent CMSWire post (which had a staggering number of retweets – thank you) I made the point that users want information and not documents, and the range of those information requirements was very wide indeed. Each of these requirements involves developing a good understanding of how users will search and what information they expect to see on the first one or two pages of results. I also suspect that the GSA was primarily an intranet solution but organisations will now be looking not just for a replacement for intranet search but a solution for enterprise-wide search.

The second issue is that even if the technology is pushed into a cloud the team needed to run a good search application is much larger than most organisations are willing to consider. At the recent Enterprise Search conference in Washington Ernst & Young disclosed that it has a search team of 6 people. The statement drew gasps from audience. To support the migration, testing and on-going management of the GSA replacement is going to require a team, even if a virtual one of people doing job-sharing alongside other commitments. For business planning purposes it will be a major challenge to explain why the increase in headcount is needed and an even bigger challenge to find and train the team in time. Good search managers in the UK are commanding very good salaries, almost certainly beyond those of IT managers at the same grade.

All the major search implementation companies are offering advice but if you would like some personal vendor-independent advice then both Agnes Molnar and I are more than willing to help. I don’t know how many organisations are using a GSA for either an intranet or a website but it is certainly in the thousands, so there is plenty of work for all of us!

Martin White

Findwise Findability Survey 2016 – strategy wins out!

The outcomes of the Findwise Findability Survey 2016 were presented by its author, Mattias Ellison, at the Findwise Findability Day in Stockholm last month. In the interests of transparency I have been involved to some extent with the design of the survey and the presentation of the results. The 2016 report can be downloaded from the Findwise site. With all annual surveys the challenge is to keep a balance between questions that relate to the trends in search implementation dating back to 2012 and yet pay attention to topics that deserve special attention at the present time. I think Findwise has just the right balance in the 2016 report.

I’m not going to work through every chart and table in detail as I want to encourage you to download the report and read it for yourself. For me the main interest this year has been the set of questions on how a search strategy has an impact on search performance. There is certainly a welcome trend towards organisations having a search strategy, up from only 20% in 2012 to over 50% this year. The report presents a series of charts which show that having a search strategy has a significant benefit on search performance, mainly because the strategy provides a business case for investment in team resources, metadata and analytics. The chart on the roles participating on a search governance programme shows a higher level of business involvement when there is a strategy in place. Indeed there is no aspect of search management that does not appear to benefit from having a search strategy. Which then makes me ask why still half the organisations in the survey do not have a strategy.

Based on my consulting work I think that the answer to a lack of a strategy is that although at an operational level search managers understand the value of a strategy they cannot find a sponsor or owner for the strategy. This is especially the case where an organisation has multiple search applications acquired and supported from different budgets, and there is no overall ownership of search. Findwise does provide some guidance on strategy development and you can find a list of headings for a search strategy on the website of my Enterprise Search book, the entire focus of which is the need to take a strategic perspective on enterprise search.

This survey is a lonely beam of light on the fairly mysterious world of search management. AIIM did publish a survey on enterprise search in 2014 but now search is not listed as a technology that the organisation sees as important. No comment! Undertaking research on the scale of the Findability Survey is a significant commitment by Findwise, especially in achieving a high level of participation, and the search community should not only be grateful for this commitment but reward it through participating in the 2017 survey. If you want to make a business case for more investment by your organisation in search then the 2016 Survey makes a definitive case for doing so through the development of a search strategy.

Martin White

The complex search process of invention

I’ve lost count of the number of times I have seen business cases for search and for collaboration that cite ‘enhancing innovation’ as a reason for investment. When you dig deeper you realise that this is just a sound bite that has no basis in reality. A significant amount of effort has been expended over the years on measuring innovation rates, for example based on patents and published research papers, but inside the enterprise things are a lot more complex. Innovation and invention take time, and there is no better example of this than the stories behind the award of the 2016 Nobel Prize for Chemistry to Sauvage, Stoddart and Feringa. The award is the culmination of research dating back to the early 1980s an then a process of success and disappointment but above all a commitment to a goal. Another stimulus was the recent acquisition by Nokia of Bell Labs, the epitome of an invention powerhouse. Bell Labs was where Claude Shannon founded the digital age in 1948 with his paper on the Mathematical Theory of Communication.

Perhaps because I have worked alongside Ben Feringa on a Royal Society of Chemistry committee this year’s award has caught my attention more than most. I recalled seeing a research paper some time ago about the the processes behind invention and it took me a while to track it down to the journal Research Policy. It makes interesting reading. I should say at the outset that ‘search’ is being used in the widest of contexts, not the use of a search application. The authors summarise their paper very well.

“Using an extensive archival content analysis of notable inventors we find that the search and discovery process of invention is inherently complex, non-linear, and disjointed. Successful inventors are skilled at managing these complex systems, receptive to feedback, and able to revisit and change course. Our search model includes a stimulus, net casting for information, categorizing that information, linking unrelated ideas, and  discovery. Our findings articulate the search process as a complex progression through a series of simple stages. As such, the study contributes to our understanding of complexity and the complex systems view of the invention process.”

There are parallels here with the way in which business decisions are made. It is not just a process of undertaking a search and then making a decision. There are multiple steps with a significant amount of feedback at each step. Only though understanding the processes of decision making, of which invention is arguably a special case, can we begin to develop search applications that support this process. As an example, being able to maintain a search query history so that a search can be re-run with a small change in query structure on the basis of further consideration of the challenge. One of the implications is that seemed relevant at the time of a search may turn out to be irrelevant only a day or two later. This rather makes a mess of assessing ‘relevance‘ as the basis for search performance.

The Research Policy paper is not open access, and I’m not suggesting that purchasing it is going to transform your search application in the next 24 hours. It does provide an excuse for me to highlight the need to understand user requirements at a very granular, process-specific, level, and not with simplistic surveys about ‘What information do you need to find’ or providing an ability to personalize that is too often an excuse not to look carefully into user requirements.

Martin White



Findability Day 2016 – Stockholm

The 2016 Findability Day event, sponsored by Findwise, took place in Stockholm on 27 October. The diversity of topics for the eight presentations was quite remarkable. After a summary of the 2016 Findability Survey by Mattias Ellison (the subject of a forthcoming blog) the opening speaker was a joint IBM/Findwise presentation on the implications of the General Data Protection Regulation (GDPR) which takes effect on 25 May 2018. I suspect most of the 200 or so delegates were quite surprised by the scope and implications of GDPR and the need for some careful planning. Next up came Misty Weaver with a superb analysis of how to make content findable. The final paper in the morning was given by Henrick Sunnefeldt on the search strategy at SKF. As well as a global enterprise search application there are 16 specialised search applications but all are managed on an integrated basis. I was especially interested in the evaluation processes that Henrick and his colleagues used to track search performance.

The afternoon session started with a rather lacklustre presentation from HPE (aka Hewlett Packard in the past) on Artificial (or Augmented) Intelligence. It was disappointing that in the initial overview there was no reference to Alan Turing or Donald Michie, a colleague of Turing’s and one of the pioneers in AI. Next up came Theresa Regli (Real Story Group) who highlighted that every single one of the vendors covered by the RSG services had some form of embedded search capability, and each needed an appropriate degree of management to get the best from the application. After coffee Kai Wahner (TIBCO)  gave a very good overview of the value of analytics in business decision making which was notable for not being a promo plug for the company. The conference closed with a very thoughtful and enjoyable presentation from Andreas Ekstrom which is beyond summarisation –  looking at his TED talk will explain why.

Adding the presentations with the opportunity to talk search from 10.00 to 16.30 is why for me this event is a must-attend each year. This is, sadly, the only search-specific event aimed at an enterprise audience. There is also the BCS Search Solutions event in London in November each year but the scope is more towards the research community. Obviously this is a Findwise event but the company do not use this event in an Apple/Oracle/Cisco style where every speaker is there to extol the virtues of a vendor product. Lest you might be concerned all the presentations are in English. I have no hesitation in recommending you consider attending the 2017 event. I will list it in my conference diary as soon as the date is confirmed.

Martin White