As digital content becomes the norm, rather than the exception, the expectations of what can be done with it, and hence the processes to create it become more complex. Publishers and information owners need guidance on implementing digital publishing (to websites, ebooks, even to print!). The requirements are different for professional and academic publishers, for trade publishers, and for B2C information providers. What is described here is not rocket science. The headings below divide the digital publishing process into twelve distinct but linked stages.
Talking to users and stakeholders reveals details of the current system, but equally enables users to state what they think could be improved. By interviewing users across the system, an overall picture can be obtained. The goals of the interviews are:
- To understand how current software tools are used
- To understand how the current production processes work in practice.
- To identify possible areas for improvements.
Deliverable: A summary document that explains in more detail the current publishing infrastructure, in sufficient detail to be able to recommend specific improvements.
All too often the needs of users are assumed. Yet the process of technology change has a far-reaching effect on the entire information industry. Information was paid, now it is free, and in some cases (such as the Financial Times) paid again, or a mixture of the two. Do the users find what they want? Do we know what they are looking for? Surprisingly, some publishers are not very clear how and where they could reuse information they have already compiled. Other publishers find surprises in the way their information is accessed. A study of 20 pensions lawyers all using the same software revealed surprising differences in information retrieval and in understanding of the product – surprising, that is, to the publisher.
The content audit follows the information gathering stage and forms a corollary to it. Although it seems obvious, a check of where and how content is held can reveal problems with the current system, as well as being an essential pre-requisite for any change to the system. The content audit counts the current content totals, as well as capturing samples of existing content from it. This exercise reveals:
- Where and how content is held
- An accurate estimate of repository size requirements and likely growth.
- Representative samples of content.
- Details of content structure that the content owners may not be aware of.
- Where content has multiple uses.
- List of inputs and outputs (e.g. to third-party delivery platforms, e-book platforms, customer interfaces).
- Where metadata originates or is amended.
The content audit makes possible the task of making content searchable across different types, for example books as well as journals. Deliverable: a spreadsheet and explanatory document.
It is rare for any publisher to have processes that are entirely unique. Typically publishers share similar issues, and potentially can share similar solutions. A remarkable feature of the publishing industry is the extent to which peers are happy to share best practice information and to demonstrate existing systems. A report that looks at four or five comparable organisations and identifies best practice can prove highly cost-effective for any organisation contemplating a change to their process. Before committing any substantial resources, then, an organisation can have an informed understanding of the real cost, and the benefits and drawbacks of different approaches. Most importantly, this methodology addresses the problem that potential stakeholders in any new processes may not be aware of the capabilities of a new system without having seen it working, so are not in a position to see the benefits of changes to the existing processes. The resulting report identifies where the centres of excellence are, together with contacts for further exchange of information.
What do we do well? How can we improve? Where are we wasting money? These are the fundamental questions for any publisher. Typically based on the information gathering stages above, I create a set of recommendations for an organisation’s information creation and delivery. The proposals comprise:
- Suggested content structure
- Benefits and drawbacks of potential structures.
- Implications for workflow
- Extent to which standard solutions or customised solutions need be found.
- Suggestions for phased implementation
- Skills required to manage the new structure
- Outline budget to carry out changes
It is possible to create proposals without first gathering information or auditing the content, but this process is inevitably less precise.
It could be said that providing links to other content it what made the Web grow in the first place, but even today many of the promises of linking information are only just starting to be realised. It could be said that digital content today has the greatest value insofar as it can be linked to other, related content. Nonetheless, creating links can be an expensive business. Should it be done manually or automatically? Does it need subject experts at all? How much will it cost and can I add more links later? This stage examines the current state of the art for creating and manipulating linked date from publishable content. It comprises:
- Suggested tools for validating content.
- Ways of linking content internally.
- How to linking data to the outside world.
- Reusing content in different formats.
- Outline costs and idea of phased development.
- A suggested proof of concept.
Traditionally, a publisher looking for new functionality will create a requirements specification and will then invite (or request) responses. Lean development has replaced much of this process, but some elements of the traditional requirements specification remain, albeit informed by the new approach.
One welcome change to the requirements specification has been to move more to use cases than specific requirements, since a requirements statement frequently assumes how the problem will be solved. If I am looking for a solution, I shouldn’t presume to know how best to fix it. A use case or problem statement will state the desired output, but leave to the supplier to specify how to achieve it. Better still, there are other, less confrontational ways of creating new functionality.
Deliverable: Requirements document (Request for Proposal) for a vendor selection process.
When a new system is envisaged, whether a new website or a new underlying content repository and workflow, the most reliable process for identifying a supplier is via a requirements specification: a formal description of the requirements, expressed as a series of use cases and items. Several procurement methods are possible, but typically a long list of suitable vendors is identified, from which six or seven vendors are invited to respond to the Requirements Specification. The suppliers’ responses are then evaluated in a tabular form to create a shortlist. Up to three suppliers would then be invited to present their solution in person. So, whether your system costs £5,000 or £5m, the principle is the same; except that larger tenders have to follow external procurement rules where appropriate. The tasks below are for a closed bidding process. This stage covers:
- Identifying suitable suppliers that have the capability to meet the requirements.
- Taking up customer references and reference site visits.
- Scoring vendor proposals and presentations to an agreed set of criteria.
- Ensuring a level playing field for all vendors: clarification of questions and running an FAQ list to ensure no vendor receives privileged information exclusively.
- Face-to-face presentations by proposed suppliers, who will be asked to carry out an agreed list of tasks.
- Risk analysis of suppliers.
This phase ends with an agreed supplier who will deliver the solution that best meets the requirements, to an agreed cost and timescale.
Once a vendor has been chosen, it is important to ensure the relationship is clearly defined. This stage involves:
- Checking the vendor and system integrator credit status.
- Negotiating the contract with the vendor to protect the publisher.
- Reviewing statements of work from the system supplier.
- Ensuring any custom work that is commissioned is ring-fenced to ensure the deliverables are clearly stated and demonstrable.
Deliverable: agreed contract and statement of work for signature.
More information projects fail at the implementation stage than at any other stage. While it may seem simple to trust the developer to implement the system specified, the devil is in the detail. The implementation of a new system throws up many questions and decisions, not least about prioritisation. A wrong decision during the implementation phase can add several months to the project timescale. Hence an independent project manager that represents the publisher’s interests is required. The tasks involved here comprise:
- Identifying appropriate tests to demonstrate fitness for purpose.
- Managing the transfer of technical skills to the in-house users.
- Ensuring that the organisation understands what the system can do, by providing demonstrations and workshops.
- Ensuring any training by the system integrator or vendor meets user needs.
- Ensuring hosting is set up in a cost-effective way that is commensurate with the system requirements.
- Identifying and recommending external expertise as required (e.g. if changes to a schema or DTD are envisaged).
It is highly recommended that an external project manager with publishing knowledge acts for the publisher. This is because:
- The project manager should have sufficient technical knowledge to understand what is being discussed in the project communication.
- The project manager is familiar with the software development process.
- The project manager is external to the organisation and so can provide objectivity: he is outside the existing business processes.
- The project manager provides a single point of contact with the system integrator.
Any project management methodology can of course be aligned to the publisher’s existing systems: project style can be waterfall or Agile, but whichever methodology is implemented the overall goal is the same: to ensure that requirements are captured accurately, costed clearly, and implemented in a clear order of priority. Agile does not mean less specification; it may mean more. Whichever methodology is adopted, I provide the following methodology:
- A way of ensuring senior management know what is happening.
- No use of technical terminology that obfuscates understanding.
- Single-page reports to the project board every two weeks, including budget updates.
- Genuine and regularly updated risk assessment and mitigation.
- And, at the end, a set of lessons learned.
A site in development requires a clearly collated set of criteria by which it is to be judged. In addition, any site requires a check of basic functionality, including different browsers, different devices (laptops, tablets, desktops, smartphones, and so on). Equally, it’s no use the system providing results if those results don’t appear quickly enough on the user’s screen. A clear set of system tests is provided and the functionality thoroughly tested. Usability testing can effectively be carried out on new systems in development. This can involve both quantitative and qualitative tests, as well as observing the software in use. This involves the observation of users without them necessarily being aware they are being tested. In this way the performance and effectiveness of websites can be evaluated.
The questions many analytics suites answer may be useful but not always relevant. All too often site owners start investigating solutions for problems that are not appropriate to their own system. No analytics package can tell you if the user has found what they were looking for. Hence analytics needs to be interpreted and used alongside other tools for evaluation. Equally, not all websites need to appear high in the search engine rankings. Making a successful website is not the same as ensuring it appears high in the search rankings. Such a listing is appropriate for some, but not for all websites. Better is to ensure that the users you want to come to the site can find it and, most of all, use it – not quite the same question.