InfoArk

Information Architects

Projects

We then advised the IFC specialists on using this controlled vocabulary to create search rules to discover relevant documents. These curated documents could in turn be used to refine search results using machine-learning algorithms, a hybrid approach producing better results than rules or AI could do alone. 

Refinitiv (2021 to 2023)

Refinitiv is the financial data and analysis service spun out of Reuters and now owned by the London Stock Exchange Group.

Whether its customers are foreign exchange traders or global asset managers, the delivery of key market information must be efficient, simple and familiar. 

We designed a logical, intuitive and consistent information structure for website navigation and search, while retaining the detailed news and data filters on which clients rely.

Which? (2019 to 2020)

Wolters Kluwer (2018 to 2021) 

 

We worked with an internal information architect and editorial subject experts to transform their navigational taxonomy into a full descriptive ontology by adding synonyms and related terms. We then created a rules-based classification system to tag and deliver content based on the frequency, prominence and proximity of these richly described terms. Iterative testing and development achieved accuracy scores of more than 90 percent, matching or exceeding their traditional manual process.

 

Baker McKenzie (2018) 

 

We structured the topics as a taxonomy to improve navigation in a SharePoint-based intranet, but then extended these concepts with synonyms to improve search. Local keywords were organized under these categories to support multiple geographic jurisdictions. We also devised information management principles, policies and governance guidelines to enable strategic development. 

 

The international agribusiness company wanted to speed up the creation and distribution of research studies, reports and submissions to regulatory agencies.  

 

InfoArk organised an enterprise-wide information model for automatically tagging content to improve collaboration and search, and for connecting linked data held in multiple repositories, both internal and external.

We tested these capabilities for tagging unstructured and structured information in a Proof of Concept using MarkLogic’s metadata solution, and GATE’s natural language processing software. 

 

The project also identified key elements of digital content held within documents, such as summaries and data tables, for efficient reassembly and publication.    

 

Browne Jacobson (2016) 

In addition, this standard vocabulary would form the core of an eventual multi-faceted taxonomy for all its information management initiatives. 
 

The firm also required assistance in structuring this relevant metadata to drive effective information retrieval and delivery.

In the short term, we focused on improving training capabilities.  However, the taxonomy will also streamline the information available to the firm’s lawyers and professional specialists, to support their partnership with clients.

Medicines and Healthcare Products Regulatory Agency (2015)

The information architecture needed to support numerous tactical projects to improve retrieval of structured and unstructured content, notably in the migration from a poorly organised shared drive to a SharePoint platform.

Within a comprehensive information management program, InfoArk devised a multi-faceted taxonomy reflecting the range of enterprise activities, and proposed complementary ways to employ it as SharePoint team sites, content folders, term set tags and a document naming convention.

We also demonstrated how the taxonomy can be extended with equivalent and associated terms to improve content classification, navigation, search and delivery.

Clifford Chance (2015)

We also ensured that this taxonomy would align easily with the business entities, attributes and values defined by the data and solutions architects, to support systems using structured, semi-structured and unstructured information.

In addition, we advised on integrating the taxonomy into a SharePoint content management solution, and proposed a representative governance structure and transparent process for managing the controlled terms.

Cambridge University Information Services (2015)

InfoArk organised the services based on key attributes, so the catalogue could present logical and consistent information targeted to the needs of its multiple stakeholders, from students and staff to researchers and technical support teams.

The work involved structuring an intuitive, user-focused hierarchy for a web-based display, and an associated data model and data dictionary to unify the resources required for effective IT support.

Serco (2014 to 2015) 

 

The models, designed as Entity Relationship Diagrams, placed service contracts in the middle of a three-level hierarchy. Related contracts were assigned to a “cluster” above the individual contract, with specific contractual delivery agreements detailed below. Relevant attributes were assigned at the correct level. 

 

This structure provided a clear and consistent definition for documents that had previously all been considered “contracts,” understandably leading to difficulties in counting and managing the company’s outsourcing commitments. The models included consistent treatment of other business entities, such as Serco’s corporate structure and its customers. 

  

 

Unilever (2014 to 2015) 

 

For a new content management platform, InfoArk created a multi-faceted R&D taxonomy. We split substances and commodities from their eventual products, and developed categories for technologies, business and science capabilities, information types, geographic locations and corporate structure.  
 
We streamlined lengthy vocabulary lists by focusing on common attributes within and across multiple product lines. We also customised a taxonomy/ontology management tool to capture and display equivalent and associated terms, to improve navigation and search. For example, the product benefit of “cholesterol lowering” was linked to products defined as “low fat.”  

 

Finally, we helped design an efficient workflow for classifying documents, and a governance process for updating the taxonomy.

Institute of Chartered Accountants in England and Wales (2012 to 2014)

The taxonomy we devised reflects the multiple facets of the Institute’s content, with descriptive terms drawn from geographic locations, organisation types, industry sectors, financial and legal vocabularies, document types, key events and user roles. As the catalyst for a new website design, the taxonomy highlights the Institute’s premium content and will improve user access through streamlined navigation, focused search and role-based delivery.  

In the final phase, we extended the taxonomy with synonyms and related topics in SmartLogic’s Ontology Manager, to support rules for automated content classification. The classification rules accurately and consistently tagged the 20,000 documents on the revamped website, to improve information retrieval.  

O2 -- Telefónica UK (2013 to 2014)

The telecommunications company needed clearer definitions and cleaner structures for two key business entities – customers and products.

Joining the Information Strategy and Management team, our chief information architect Jonathan Engel created consistent data models and metadata to describe business customers and the product catalog across the enterprise. In particular, he supported a Proof of Concept to streamline business processes by unifying the information available in two major technology systems -- Salesforce.com (containing customer contacts and the product catalog), and the billing engine (detailing specific orders and usage).

He also devised a representative governance structure and process for information policies and business data, to ensure the views of major stakeholders were represented in change management. To support increased team collaboration, he designed a new SharePoint site.

Jonathan also used his expertise in customising search queries to support the Business Intelligence team analysing social media and other Big Data sources. In addition, he worked on a month-long pilot project to identify prime living and working locations, based on combining selected Open Data with proprietary O2 information on customer preferences and activities.

Ernst & Young (2012 to 2013)

As Senior Taxonomy Manager, our chief information architect Jonathan Engel had global responsibility for devising consistent metadata to ensure relevant information was easily accessible to the fee-earning staff.

He worked with subject specialists to create a logical, multi-faceted hierarchy, focusing on the firm’s service lines, content types and geographic locations. He extended the taxonomy with synonyms and related topics to improve tagging accuracy and search results for knowledge and content management initiatives, especially a new SharePoint 2013 installation.

His expertise in automated content classification supported a Proof of Concept for the SharePoint initiative and selection of a preferred vendor, Concept Searching. He also delivered a representative information governance structure and process, and convinced key stakeholders to participate.

As part of a knowledge management programme, InfoArk completed a Proof of Concept to improve the information architecture within the NES Knowledge Network, an online resource for health and social care professionals, educators, researchers and service users.

Tasks included devising a high-level, multi-faceted taxonomy to meet stakeholder requirements and form the basis for a new, user-centred interface in collaboration with web usability specialists. We then selected lower-level headings from two controlled vocabularies – the Social Care Institute for Excellence online thesaurus and the MeSH thesaurus for medical terms.

To support specific functions such as evidence-based, clinical decision-making, the taxonomy also defined content types, key sections within documents, and two levels of content specialisation.

We extended the initial taxonomy with synonyms and related topics, and demonstrated how this structure could automatically generate successful, popular search queries in the FAST search syntax.

Royal Botanic Gardens, Kew (2012)

InfoArk devised a consistent information model for Kew Gardens’ physical and digital images, to unify the RBG collections within a new Digital Asset Management System.

We consulted subject specialists, photographers and collections managers and built the model to incorporate multiple attributes of plant specimens and digital content.

The captured attributes range from format and quality specifications to geographic and subject descriptions. The new information model will allow users across the organisation to navigate and search efficiently for relevant digital content.

We built an enterprise-wide taxonomy to support the UK’s leading online shopping company (Littlewoods, Very) and its related financial services products.

The work included a survey of key stakeholders, an analysis of usability preferences, and workshops with in-house information specialists.

The resulting taxonomy focuses equally on describing the product catalog, customers and key sales occasions and the relationships linking all three, e.g. by defining a customer profile, appropriate product types and relevant sales opportunities.

This consistent corporate-wide vocabulary will improve navigation and search on its websites, streamline marketing and order processing, and generate more focused business intelligence.

Barnardo’s (2008-2012)

The improvements in information management followed the establishment of a reliable link between the data on service users and the information on service offerings. This connection allowed an accurate audit and consistent analysis of the charity’s impact.

The corporate vocabulary also introduced consistent terminology into a new tendering pipeline, to help the charity track and analyse its project bids.

In future, this consistent, structured vocabulary will streamline user navigation through the Livelink content management system, support communities of practice and improve enterprise search. 

InfoArk designed a comprehensive news ontology for The Times and The Sunday Times, starting with preferred terms, then adding synonyms, related topics and numerous “contextual keywords” found in relevant documents.

We built these relationships into well-structured FAST search queries for each topic, to identify articles to include as training sets for automatically tagging news stories with accurate metadata via Nstein (now part of OpenText).

This consistently applied metadata is transforming the way editors produce and deliver news – the linked tags suggest related articles for news packages and identify appropriate content and advertisements to match readers’ interests.

Oxfam International (2009)

After analysing user surveys and “hit” rates, we consolidated the core taxonomy, built streamlined navigational links around key topics like Health or Education and devised a new home page featuring more graphical icons and integrated, customised search.

Following a presentation to information and domain specialists, the site map and wire frames were refined. The result was welcomed by senior staff as a cleaner, more effective site for retrieving relevant content.

UK Environment Agency (2008)

We ran a staff workshop to establish relevant top-level headings, then reviewed thousands of terms for inclusion in the scheme, including synonyms and associated terms from a recognised standard, the General Multilingual Environmental Thesaurus (GEMET).

We then built a hybrid classification system that combined hierarchical structure, synonyms and relationships. We also showed how this corporate vocabulary improved navigation, search and retrieval, and could be used to push content to staff and citizens according to profile attributes or personas.

Wiley/Blackwell publishers (2008)

 

Using an innovative method to place classification terms within a pre-tested search syntax, we generated automatic, well-structured rules and relevant training documents, which applied the content tags consistently within Inxight’s Categorizer Workbench.

 

We then tested the categorisation results for accuracy against a manually-tagged sample, refining the classification structure, rules and training documents where necessary.

 

UK Healthcare Commission/Care Quality Commission (2006 to 2008)

This health and social care regulator needed to support its risk assessment and analysis teams by making relevant information easier to find.

We began with an information audit, identifying internal and external information sources and structuring the related metadata into a high-level information architecture. We worked with subject specialists to fill this hierarchy with preferred terms from three standard medical vocabularies: MeSH, SNOMED and the Map of Medicine. We then extended this taxonomy with synonyms, related topics and contextual keywords.

We also consolidated health and social care metadata ahead of the agency’s merger with the Commission for Social Care Inspection. We helped with vendor selection for an enterprise search solution, by defining requirements and running a Proof of Concept.

We then integrated the information architecture and selected search solution into a SharePoint portal and Siebel/Oracle business intelligence package.

This “information cabinet” unifies access to separate document repositories, supports collaboration, and features taxonomy-led navigation, customized search queries, business analytics, user alerts and links to subject specialists.  To customise the search and alerting functions, we also created investigative, semantic tools for data analysts.

Eidos (2007)

Elements of this hierarchy would also support the marketing department’s digital archive for product promotion.

We met with art directors, corporate librarians and marketing staff to consolidate and refine their proposed taxonomy. We kept the hierarchy lean and logical by introducing a separate metadata category for shared attributes. Users could then find related assets from different areas of the taxonomy.

Telstra (2006)

•  Studied background documents to review current and legacy IT initiatives

•  Identified sources of classification headings -- internal and external
•  Interviewed information managers

•  Held workshops combining bottom-up/top-down analysis of potential taxonomy categories

•  Highlighted gaps and areas of overlap

•  Recommended relevant content classification standards
•  Clustered and defined 300 taxonomy headings, along with synonyms, related topics and properties, to create a draft classification system

•  Recommended project phases for introducing and integrating this content classification

UK Department for International Development (2004 to 2006)

We began with a content classification review that identified and analysed core category headings, then built them into a high-level hierarchy. We supplemented this taxonomy with a thesaurus of synonyms plus related terms to support successful search.

We reviewed vendors providing taxonomy maintenance, automated classification and search. When Verity K2 was selected for the project, we organised the automation of the content-tagging process by supplying consistently structured and weighted metadata rules for each taxonomy heading and testing them by searching for sample documents.

Dow Jones (2005 to 2008)

We delivered a comprehensive content structure that retained the best elements of the existing system while incorporating industry standards.

We also assessed the accuracy of current content tagging and ran a successful Proof of Concept for expanding the automated application of this metadata. Using Inxight’s Categorizer Workbench (now an SAP product), we demonstrated how the classification database could generate consistent tagging rules and relevant training documents to support more efficient search.

Syngenta (2004)

 

We discussed usability requirements with information managers, presented four proposed portal structures and then ran a workshop to evaluate the corresponding wire frame models and choose the most effective one.

TNT Express (2004)

We worked with OpenText to integrate the content headings as metadata tags, then extended the classification structure to include focused vocabularies from the global company’s IT unit and head office functions.

Reuters (2000 to 2003)

He also defined the key requirements for a new content management system and Web-based news delivery. He then managed the implementation of an automated categorisation system, using Inxight’s Categorizer Workbench alongside the Documentum-based content management solution.

Using templates and pick lists, he applied the same metadata to multimedia content, for example when he customised a prototype Digital Asset Management system, supervised an on-demand business video service and produced a strategic plan for integrating TV, newspaper and Web graphics production.

                                                                    Copyright © Jonathan Engel 2024.