February 2015
By Wolfgang Ziegler and Heiko Beier

Image: © Anatoly Maslennikov/ 123rf.com

Prof. Dr. Heiko Beier is a physicist and has been teaching since 2010 in the field of international media communication at the University of Applied Languages at SDI Munich. He is the founder and CEO of moresophy GmbH.


heiko.beier[at]moresophy.de
www.moresophy.de


 


 


Prof. Dr. Wolfgang Ziegler is a physicist and has been teaching Communication and Media Management at the Karlsruhe University of Applied Sciences since 2003. He heads the Institute for Information and Content-Management (I4ICM) of the Steinbeis Transferzentren GmbH at the Karlsruhe University of Applied Sciences.


Wolfgang.ziegler[at]hs-karlsruhe.de
wolfgang.ziegler[at]i4icm.de
www.i4icm.de


 

Content delivery portals: The future of modular content

How will content reach users in the future? Moreover, how can the interests of specific target groups be better considered? Special portals with intelligent search options are promising to solve this task. So what has already become reality and what are realistic visions?

Creating and managing modular information has become a broad topic in technical communication in the past years. [1]. This has led to component content management systems (CCMS) becoming widely used tools, the modular creation of content becoming the tool of many technical writers, and cross media publishing becoming the technical norm. However, what has not changed in many cases is the way published information is planned, created and accessed. The traditional document-based perspective of information often still rules here.

Additional systems and concepts have been emerging for some time now in the area of technical communication, which can be called Content Delivery Portals (CDP) [2]. These types of portals could shed new light on the modularity of content that has been propagated since decades. CDP are being promoted by different streams:

  • The product complexity that has increased with time and the corresponding flood of information in the field of  pre- and after-Sales information
  • The methodical opportunities provided by content management systems to generate variant-specific tasks
  • The increasing process integration and overall data integration in the organization
  • The slow move away from print-oriented documentation formats expected by many
  • The familiarization to the (more and more exclusive) availability of online product information in the consumer field
  • The proliferation of mobile devices with the responsive display forms of web applications and apps
  • Online research, tagging and facet search in the consumer area of web-shops
  • Influence of social media and use of micro-content

The last two points should be considered as drivers from the private and medial area, which are carried over to the professional area as can often be seen. We can therefore expect applications that establish themselves as portals or similar for technical product information.

Content-Delivery-Portals can generally be defined as follows: CDP offer web-based provision of modular, aggregated or document-based information for access through different target groups with the help of content-related search mechanisms.

Therefore, these systems have a delivery component as well as a retrieval/search functionality technically and methodically. “Delivery” can have two meanings here. It indicates the handover or takeover of content from the source systems to a portal, or the usage-side output of content to the user or even the transfer of portals to third parties.

Horizontal and vertical systems

In the corporate context, portal systems with different areas of application have existed for ages outside of technical communication. The market can be divided into horizontal and vertical portals according to Gartner studies [3]. Horizontal portals are applications that integrate information comprehensively and from different sources within the organization: cross enterprise applications. Vertical portals focus on special cases of content types, sources and usage, such as industry-related product portals (Electronic marketplaces) or company-specific product portals. Examples of both types exist for products from the most varied fields such as journalism and media, medical technology or electronics and building supplies.

If the portal systems are transferred to technical communication, it is possible to define similar comprehensive (horizontal) and specialized (vertical) portals. However, the number and type of information sources and usage should be considered as separate dimensions for more accurate presentation: The source can be the component content management system for technical documentation with all its modular content and media for instance, as well as the documents created from it. A multiple number of different source systems and file repositories can also be considered for documents from other areas of product information, such as training, service or sales.

Scenarios for portals

The application can be oriented according to a specific use case of content research or include several usage groups with different requirements related to information and search strategies.

The portal categories derived from those are compared in figure 1 with a look at the sources and purposes, where the boundaries are blurred. Four general scenarios can be derived [2]:

Scenario 1 – Content portals serve the purpose of exclusive search of CCMS content for users who are directly or indirectly involved in creating and/or using CCMS content. This can be processes in the area of research and reception, as well as preceding tasks such as proofreading, approval and translation.

Scenario 2 – Cross portals additionally include different user groups interested in the contents of the source CCMS, but who want to use it across departments for instance. Use cases include the training or service departments wishing to use texts and media even outside a CCMS for presentation purposes. Since a CCMS is usually also in a position to manage delivery documentation and other documents, cross portals can deliver structured access to supplementary or background information.

Scenario 3 – Multi-source portals get content from different sources. Next to the CCMS, these can be different file systems in the simplest form. Databases and the variety of specialized management systems should be accessible as extensions, e.g. spare parts catalogues, product and catalogue data, standards, development information, reports related to diagnosis, testing, service or errors. These portal forms have a limited user group with clearly defined usage scenarios such as search and evidence related to customer and service information to create documents that can be validated. Other usage scenarios could be service information systems that bundle all sources relevant to service and make them accessible online to technicians.

Scenario 4 – Enterprise-Content portals allow internal and external user groups differentiated access to content and documents from different sources. Functionally this also demands more elaborate rights concepts even for larger user volumes and highly specialized options for use such as content aggregation and publishing, subscription functions and user (group) specific search mechanisms. Such a solution should also offer interfaces or services for third party systems as “universal content source”.

Enterprise content portals compete with document management systems or their extensions to some extent through ECM systems (Enterprise content management) with almost the same name. The difference with respect to these competing special systems is simply the starting point of the content. The CDP described here have the CCMS data as the “core” and use the treasure of metadata, the modularization logic and the information depth or semantics, depending on the system and the implementation. Other data and documents are supplementary and not primary.

Figure 1: Classification of portal categories with different scenarios for the sources and usage of information
Source: Wolfgang Ziegler

 

Options for access

Another special feature of CDP applications are the IT and access architecture possibilities for portals shown in figure 2: The nature of web (server)-based systems allows global usage on the net or restricted access to groups and individual users. Information about installation, repair or service could count as examples for this on modular or document basis, as well as download portals for operating instructions. Intranet portals reduce access to specific user groups for internal corporate information, as described for the multi-source portals.

Figure 2: Scalable types of CDP architectures of portals for local on-site applications on machines to web-based online help.
Source: Wolfgang Ziegler

 

Due to the different software, operating systems and browsers existing in many products, machines and plants, it is a short step that is already being taken in different special applications for local usage of reduced web portals with and on the products. Such systems are then useful as onsite help in complex plants with software controls, in large medical equipment, in on-board systems of vehicles or even household appliances such as refrigerators or heating systems.

Onsite and Online help

Many modern products from IT and communication such as telephone systems, printers or routers are equipped with web servers. Onsite help for these is the extension for the search for required explanatory and instructive information in the context of the product. The respective question or even the situation of use is many a times a “known” situation for the machine or the product and can be used as search context. Even the complex configurative as well as (software) version dependent problems of help texts and documentation can be approached methodically with the modular structure of the information from the CCMS and a corresponding variant and metadata management.

In the reduced form, the CDP can also be used “stand alone” and portably on PCs as online help depending on the system architecture and provider. Content can be searched directly in an application on the machine without local installation and can be updated over the Internet if required. The last named functions are already known from other web applications and apps. The extraordinary feature here is however, that the content delivery portals unite the described options and can scale them to different sizes of the CDP-architecture.

Search and find

Modular information forms the basis of how content management systems work. The types of information can be classified based on metadata for instance, as they are relevant for specific information usage. Thus, service technicians have a clear fount of necessary information types such as change information, repair descriptions, overviews and security measures [4]. The methodical classification of modules by product component and information types in the CCMS therefore delivers one of the helpful search and access options even for CDP [5]. Add to it the search over the product spectrum and existing document types.  In contrast, information must already be reduced to the respective products and configuration of the user in customer portals and onsite help.

Special features and restrictions

The search options in the interface of the portals can differ widely: Navigation and search trees for taxonomical and hierarchical dependencies, breadcrumbs as well as fixed tags or tags to be assigned by the user.  Faceted searches have become popular lately in web portals, like at Amazon and many others. They allow quick, simultaneous access to several search categories [6]. This type of search should also develop into one of the most import access types even in CDP.

However, for this it is necessary for the corresponding metadata to be available in a standardized form, which can be used for faceting. This will seldom be the case with respect to multi-source portals in particular, since the contents are usually of different types, as concerns content, format and media type. They do not even originate from a standard content concept. Existing metadata is therefore probably incompatible syntactically as well as conceptually to a great extent.

The same limitation applies to cross-portals that serve different target groups. It makes sense to optimize access to content for different user groups. Along with a product-related structuring and delivery of content, it is helpful for service-oriented applications to develop metadata on customers or specific purposes of use [7]. The requirement for integrating more data and content is often the result, e.g. customer information from a CRM system or support requests from a ticket system. A cross portal developed further to an enterprise content portal in this way can provide a lot of added value, provided that the relevant contents can be detected well.

A target group specific content delivery can hardly be provided anymore in such cases using manual content structuring and mark-up of content. Extended processes can provide automated structuring of content here. They go beyond the search and navigation methods presented until now. To understand them it is worth taking a look at the areas of application for which the transmission of content and assigned products is a direct business purpose and sales driver: Modern media and e-business portals already apply extended processes for a concept-based enrichment of content, as described in the following sections [8].

Three methods for analysis

Search engines are based on the principle of a complete indexing of all contents. An index in this sense is nothing other than a database containing references to all words coming forth in the indexed documents. However, the deciding factor for the quality and intelligence of a search is that the contents of a document must be understood not merely as character strings, but as concepts carrying meaning. Modern search engines permit enriching search indices for such concept information. The most advanced developed solutions use three different technologies at the same time for this:

Linguistic analysis: The identification of word types or random variants and compounds of words using morphological lexica and rules building on it enable an error tolerant evaluation of the character strings in the documents. Linguistics based search engines find more, because they do not just search for exactly what a user has entered in the search machine, but also automatically identify typing errors or variants of words as equivalent. A search query for “Lever” therefore also finds documents with the word “adjusting lever mechanism”, a search for “development” will also find those with the sentence “to develop a product for…” Linguistic analysis is highly dependent on the applied language and the rules applied differ strongly between different languages, notably English and German.

Statistical analysis: As soon as the search engine does not compare character strings anymore, but words, the statistical analysis of unstructured information delivers considerable additional value. Even main key words of a document or contexts between topics across many documents can be determined with it. E.g., it is possible to find out which problems are mentioned conspicuously often in service queries and that too in relation to which product or which customer. Statistical processes can learn. This allows the automatic classification even of service queries, e.g. by the type of problem.

Semantic analysis: Ideally, the automated classification and metadata enrichment is controlled through semantic models. The knowledge that specific customer-groups or industries have specific requirements can be provided centrally through a model for example, and considered while indexing content. Data of this type is usually present manifold, but is not brought in context with the contents. Semantic processes allow the effective use of structured data while evaluating unstructured portal contents. The use of such a process is application specific and varies in complexity from simple synonym extensions to use of rule-based logics. The level of semantics ranges from simple thesauri, to poly-hierarchical taxonomies to semantic concept models like ontologies. Semantics can be used to integrate field-specific vocabulary or different metadata schemata can be merged effectively. Semantics supplements linguistic-statistical processes with a functional intelligence. As a result, documents can then be enriched in the portal with typified metadata and structured queries can be placed in the portal against a former completely unstructured information basis.

Figure 3 explains what additional knowledge can be developed in this way in search machines and so in principle can also be used in cross- and enterprise-content portals. This first ensures that content from different sources can be interpreted in standard form and automatically enriched with standardized metadata. Not just faceted searches can be offered for different target groups, but also proactively individualized recommendations can be generated for content on this basis [9]. Users are thus automatically assigned to different target groups and supplied with content optimized to the relevance level for this target group over the portal. Even search applications with a knowledge graph optimized to the requirements of technical communication can be developed on this basis. Users then get suggestions for further search questions or completely specific solution instructions for technical problems.

Figure 3: Concept-based indices as basis for flexible search and content delivery applications in portals.
Source: Heiko Beier

 

Status of the systems and architectures

Which of the said systems is already reality? The market for CDP is developing noticeably at present. A number of CCMS providers are coming up with one of the first product generations of CDP, with different focus points and depths of the addressed architectures and functionalities. More providers will definitely follow this in particular because basic functions of content and cross portals are already present in CCMS more frequently in the form of web-based review portals, for web accesses in the framework of translation processes or other web clients.

CDP that develop from the CCMS-field naturally possess the logic of metadata-based modularization, aggregation and variant building. Search and filter strategies can be built directly on that. Multi-source concepts are presently possible only limitedly by default, whereby file repositories and external documents already managed in CCMS are easier to integrate.

Furthermore, products and solutions that can be considered as publication media for technical information in the form of independent portals and/or (help) browsers already exist to some extent with extended search mechanisms. However, the takeover of content from third party systems, i.e. from CCMS and other source systems, with all metadata and linking required for it is one of the processes yet to be standardized technologically for example.

Technological competition exists in the area of cross enterprise content portals with document management /ECM systems or Sharepoint solutions, insofar as these are extended with more complex indexing, search and management functions. Similarly existing Product-Lifecycle-Management-applications (PLM) in larger companies or Enterprise resource applications (ERP) with their portals across the company offer a number of the presented concepts. A specialization in the situational (research) requirements of technical communication is however not provided here off the cuff.

Meaningful potential

There is a very high potential in the area of displayed online help that delivers a clearly defined field of application and a frequently required application for machines and systems as content or multi-source portals. Onsite help is already created in a number of industrial projects as individual solutions, as basis of Opensource for instance.

The presented extended indexing options of linguistic, statistic and semantic type are already a reality in related media areas. To transfer these to portal type of technical communication is meaningful and helpful for the comprehensive search for information from different sources and for different user groups.

 

Links and Reference Literature

[1]    Straub, Daniela; Ziegler, Wolfgang (2014): Effizientes Informationsmanagement durch spezielle Content-Management-Systeme, 3. Auflage. Gesellschaft für Technische Kommunikation – tekom e.V.

[2]    Ziegler, Wolfgang (2013): Alles muss raus! Content-Delivery für Informationsportale. Band zur tekom-Jahrestagung.

[3]    Gartner (2013): Magic Quadrant for Horizontal Portals.

[4]    Drohomirezky, Konrad; Ziegler, Wolfgang: Was Techniker benötigen. In: technische kommunikation, H. 3. S. 15–19.

[5]    Drewer, Petra; Ziegler, Wolfgang (2014): Technische Dokumentation. 2. Auflage. Vogel Verlag.

[6]    Kalbach, James; Lindemann, Karen (2010): Facettierte Navigation – jeder kennt sie, doch wie funktioniert sie eigentlich?

[7]    Beier, Heiko (2012): Zielgruppengerechte Zugänge zu technischen Inhalten. In: Zielgruppen für Technische Kommunikation. Schmidt-Römhild, Lübeck.

[8]    Rosenträger, Stefan (2014): Semantische Meta-Daten: Motor für verkaufsfördernde Empfehlungen in Online-Kanälen. In: DOK Magazin, H. 2.

[9]    Beier, Heiko, Symanek, Stefan (2014): Modellbasiertes Matching von Kandidaten-Profilen und Projektausschreibungen im Online-Recruiting. KnowTech 2014: Zukunft der Wissensarbeit, Kongress für Wissensmangement, Social Media und Collaboration.

[10] Google Knowledge Graph (August 2014)