Re: [TaxoCoP] Taxonomy Disaster Stories
- Nice observation Gary - we often stress the business strategy as a driver for taxonomy development and take the content strategy for granted. Would you like to say a little more about how you think the two strategies should interact?POn Mar 14, 2013, at 11:18 AM, Gary Carlson wrote:
This is an excellent argument for viewing taxonomy work in the context of content strategy or at least understanding the relationship between the two. The SEO hit was not a result of a poor taxonomy, it was the result of not having a content strategy to build out the necessary content.On Mar 13, 2013, at 7:52 PM, Patrick Lambe wrote:Another anonymised contribution, this time looking at externally facing taxonomy work. What looks like a productive and rational approach doesn't always work out with the technology or the maintenance implications!I was tasked to expand a marketing site's taxonomy to increase their SEO footprint, allow for more refinement, and maintain the specificity of inbound mappings. The taxonomy grew from a few dozen categories to over 500.We are now in the process of removing many of these categories from the site for a variety of reasons but primarily because of the SEO hit the site has taken since Google's Panda update last year. Google's Panda update penalizes thin content pages and we never had a strategy or the resources to populate these category pages with relevant content.The lesson learned for me is to really really understand whether the organization can support a large taxonomy because there are 'costs' associated. We all love the challenge and understand the benefits of growing a taxonomy to better organize information but there are costs associated in terms of maintenance and optimization (remapping content, keeping the categories populated with fresh content, training of staff to be aware of these new categories, etc.).I've witnessed that the benefits of growing the taxonomy to the depth I did (from 2 levels deep and a few dozen categories to 6 levels deep and over 500 categories) didn't outweigh the costs.On Mar 8, 2 013, at 12:21 PM, Patrick Lambe wrote:Here's another anonymised contribution - some resonances with Chloe's challenge a week or so back:When working in the Asian office of a global company in the late 1990s, I built up a traditional hierarchical 3-level taxonomy applied to file shares that was reasonably successful. It was based on file-plans that we had been building since 1990. We centralized the creation of all folders in 9 countries against the tax onomy with web-page request forms, local and central approval and actual folder creation done by technology people in Australia. The vast majority of end-users were positive about this approach. They were from all parts of the world, about 1,000 users.I transferred in in the early 2000s to the European HQ as the global head for Records and Information Management. They had no RIM program at the global HQ with 2,000+ employees. I tried to convince a group of senior functional heads (Finance, IT, Operations, Logistics) that they needed a taxonomy. They hated the idea. They said it would never work with Europeans, only Asians would be so "obedient". After 2 years I was exhausted. I had sourced a RM system for paper records and installed it with a classification system I was careful to never call 'taxonomy'. It wasn't very accurate because I wasn't allowed to do the requirements interviews, analysis of existing classification schemes, validation tests with small groups of users that had all together worked well in Asia. At the HQ it was all by stealth and in the end I was just guessing what would work.Taxonomy, metadata, facets, classification, controlled vocabulary all speak to control (I know some of these are synonyms). Executives know there will be lots of pushback so they don't want to support these approaches. They like the IT vendor who comes in and says it can all be done transparently with auto-classification and search. As taxonomy professionals we know these only work well in conjunction with taxon omy, metadata, facets, classification and controlled vocabulary but the system is purchased and the IT vendor is long gone before someone like us is allowed to make this point to the executives.
Now that company has a new global head of RIM, the third since I left. He recently told me he is trying to convince the management teams to build a global taxonomy for file shares and Sharepoint. How do you sell the idea of a taxonomy to executives who won't listen?
- Here's another anonymous contribution - a lesson in the importance of preserving team continuity on a large and complex projectTaxonomy in Xanadu
The project was vast in scale, ambitious, with many stakeholders who
had many grand visions. Taxonomy was only a small part of the overall
plan. I joined the project ten years after the project was first
conceived, which should have served as a warning, but because much of
those ten years had been spent in proof-of-concept experiments and
requirements gatherings, and given the vast size of the organisation
and its democratic culture prone to committees and lengthy debating,
this did not seem entirely surprising. However, rather like the Greeks
at Troy, the ten years of planning and minor skirmishes had begun to
take a political toll. One major attempt to deliver had been scaled
back at the last minute, many of the original project team had left,
and I joined at the point of a re-grouping, re-budgeting, and
re-prioritisation exercise. My role was to find a way of integrating
existing taxonomies and thesauruses with the new enterprise
architecture, but I had no direct control over the project budgets and
no direct voice in any of the overall architectural decisions.
Nevertheless, at first, all seemed to go well as there was much
enthusiasm amongst the technical team and the user interface design
team to create a state of the art search and navigation system. The
lead technical architect understood taxonomies and taxonomy management
software and was experienced in delivering intergrated search
solutions. We worked together on software procurement and data
migration planning. All went well, the software was acquired and the
data migrated successfully into the new taxonomy management
application. All that remained was to link it up to the new search
engine. However, by this point, other areas of the project were not
doing so well and the total time and money spent on the overall
project had become of increasing concern to the key stakeholders, who
began to question the original vision. Then the technical architect
left - strike one.
It was decided that in order to save money, the technical architect
would not be replaced. The role was passed like a hot potato amongst
various people who had other full time roles and little experience of
delivering search systems and eventually was simply left unfilled.
Meanwhile, a major crisis in an unrelated area of the project had come
to light and so all resources were diverted away from the search part
of the system, in order to fight that particular fire. The decision by
the key stakeholders that search was a lower priority than other
aspects of the project became politically impossible to reverse -
The interruption to the work on search led to huge knowledge loss. In
a downward spiral of diminishing resources and stakeholder pressure to
cut costs, the data analysts and most of the business analysts who had
overseen the data migration into the search and taxonomy system left.
Almost all of the UI design team left. Most of the technical team,
including the programmers, left. Only once the other aspects of the
project had been brought under control were we able to return to the
completion of the search system. However, by this time the loss of
staff had been so great that not only was there no-one on the
technical team who knew what they needed to do to deliver a complete
search system, but no-one was left who could understand the code that
had already been written nor the scant documentation that had been
left behind. There was no money remaining in the budget to re-hire any
of the original technical team or bring in any new search and taxonomy
technicians - strike three.
I hope that we can find the money and the technical knowledge to
complete the project, as so much of the work that has already been
done was sound. However, I fear that once the “Person from Porlock”
breaks the flow of such a project, the vision and the knowledge
disappear like a dream forever.