Workshop #13: Web Content Management Systems: Architectures & Products

KM World and Intranets 2005
Workshop #13: Web Content Management Systems: Architectures & Products
Presenter: Tony Byrne, creator of CMS Watch and CMS Report

Web content management = WCM or web cms

1800 different web cms vendors. How do you choose?

Speaker: Tony Byrne founded CMS Watch – http://www.cmswatch.com/

Enterprise Content Management (ECM)

- how does WCM fall into this?

- imaging – first area developed
- Knowledge management
- Digital assets management
- Product data management
- Portals
- Collaboration
- Records management
- And others…

Often vendors will call themselves “ECM” but not necessarily for all these areas; hope that you will use their product across various rolls.

Have in common:
- processing content;
- content transformation;

Web Content Management

Structured document – usually follows a document model, may have tagging, follows a template of some kind

Some systems deal better with unstructured content; some deal better with structured

- moves content into forms usable by end-user in web formats – PDF, web pages, for PDA, XML (syndication)

Are you working with content management, or website management?

Is technology meant to be “out of the box” (but limited) or are you creating a platform on top of which you can create other applications.

Need for content reusability puts a lot of emphasis on the editorial system and processes.

Two Phases in the Web Content Management Life Cycle

- 1st Phase: employee/architectural – author facing
- 2nd Phase: client/delivery phase – client facing – publishing & distribution

Library services – control over versions of a document; who has control over versioning

Authoring & Transformation

- have to absolutely get this right
- a lot of systems make incorrect assumptions regarding the needs of content authors/owners
- consider content managers, authors, power users, average users – they will all have different needs
- DHTML and Javascript have issues from user perspective (e.g. does not support various browsers); however, a lighter solution (however, thick? Have to download the programs) Not customizable.
- XHTML – more standardized than some other HTMLs
- Copy and pasting content formatting from Word. Not a problem with Rich Text Editor, but with Word; may start looking different – HTML gets cleaned up underneath especially if people aren’t using Styles. Vendor demos like to show cutting and pasting – need to understand what is happening and try with your own documents

Types of solutions:

- Thick clients - Have to instal the programs. Not customizable.
- Morello – client interface; very slick. Not customizable; need to be able to use out of the box; a thick client that has to be installed
- Popular on intranets: WebDAV compliant: drag content into directories; something may happen when it is put into folder. Has to be simple structure.
- Word to XML or Word to database interface e.g. Fatwire. Demos very well, but users rarely use it. Works well if using Word Styles; can map the styles to database fields – may be good for law firm using Styles
- Outlook as a client – Outlook plug-in – approve/reject from there; however can’t see the underlying content in HTML out of the box; however, may be able to configure it that way
- In context editing – browse to part of the site, click on page component and edit it right there (Red Dot innovated this and excels at it; however, others have adopted). Minimal training; good when lots of people are updating. Good for lots of “casual” content The big issue from an enterprise perspective: with heavily reused content, cannot see how the one change affects across other places across the intranet/website. Can quickly deteriorate.
- Retrieval: page finder template
- Pre-publish – create content and release at different times

Usability?

- create a persona and create interface for that person. You may think they will hate an interface, but they may love it – test with actual users at different levels
- best interface may be one that can be easily modified for different content authors, rather than using something straight out of the box

Tagging

- limited vocabularies are a must
- can you change the taxonomies
- can you change the underlying tagging fields easily?
- Tagging Interfaces
o Hierarchical – tree-based interface; check off for multiple classification to one item

Workflow

- almost everyone asks for it, hardly anyone uses it
- e.g. repetitive task tracked and content pulled out automatically
- not used if not already in their workflow process
- Models
o Newspaper production is good analogy for the roles (writer, editor, copyeditor, managing editor)
o Best not to create parallel workflow; allow only one person at a time to work with something
- Very difficult for a consultant to be able to model the workflow of an organization
- “workflows” and “tasks” should ideally be in the same in-box
- Various languages loaded up simultaneously – must live in a world of parallel workflow

Be cautious of all WYSIWIG design tools - experience has been very poor with GUI designs.

Promotion Path

- organizing the steps to “go live” with content
- may want a staging area to pull content and associated template code, images, etc. together and viewing it before it is published
- Interwoven only works in the production area; push it out to another area for content delivery
- Enterprise web content management – easy to run into problems late in the process e.g. pull model may not comply with security requirements; may need to be pushed
- Is the one vendor good at all aspects? Inside firewall – limited number of users; outside firewall (i.e. public website, on extranet) – could get very high traffic and need greater scalability

Keep asking questions with vendors. Do you really need the number of servers they are suggesting? Buying more product than you really need? Greatest problem in industry is people overbuying product, not underbuying. Don’t really need all this gear they try to sell. Most often will just “publish and go”.

Issues in Delivery Phase

- with CMS, portal or other delivery product

Page Generation

- how and where are templates actually stored? May be in CMS with its own tag, but may be proprietary format; others have them outside the CMS – using standard tools developers know; however, templates reside outside the CMS
- are you already a Java shop, or PHP?
- How many presentation templates do you really need? Use minimum you can get away with
- If you make a change to a template, does it change the “parent” or does it create a new template?
- Two extremes as to how pages get generated: baking and frying
- Baking – pre-generating a page out of the database – creates into a static HTML file – advantages: great performance; however, everyone getting the same page (or one of just a few versions). Newspaper model: “bake” a new edition every night and post it – Proctor and Gamble uses
- Frying – content is not assembled until user “clicks” – get streamed down to HTML – good for personalization since you don’t know which users will want that data
- Hybrid “Parbake” – pre-assemble as much as you can in advance, and then the DB manager puts out new content

Caching in CMS

- someone makes a change and the change does not show; in an ideal world, the CMS would automatically flush the cache to show new content


Replication
- done for reliability, performance, or both
- may get better performance because both working
- will need licenses on both servers for “load balancing” model
- varies by vendor – some will give discount or free for “failover” server (backup for emergencies); other will make you pay for both.

Publishing Out to Other Formats

- WCM – not quite at single source of content that can be put out to various locations and forms yet – designed to send out to various locations/sites; designed to send out to various electronic formats; cannot yet send out to print.
- Primarily a content modeling challenge, not a technology challenge

Wireless:
o WML Source code – similar to Apple’s old Hypercard
o For streaming stuff to wireless device, this is good to go; however, not designed for the norms and format for the device. Presupposes you have shredded the model down to the level of detail needed for a PDA.
o Need to start with the output and work back to create distinct elements needed
- If you want to be close to your customer and adaptive, you need to manage it in a more sophisticated way.

What a WCM Product Won’t Do

- it won’t organize your content and navigation
- won’t make the site more usable or improve the presentation
- won’t optimize your content for the wireless environment

At least 1800 web content management products

- more vendors entering the scene and little products getting bigger
- VERY fragmented market
- Most are regional; active in their own areas or a specific industry (usually small, under 10 people on staff)
- Very few standards as to how to publish to the web
- Consultants creating solutions for their clients
- Inexpensive to get into developing in this area
- Young industry

Vendor Notes

Speaker has observed that most organizations reducing the number of content creators rather than increasing.

You are “getting married” to the vendor. Look at their personality, corporate culture. E.g. Stellent – based in Minnesota, very earnest, not slick, very engineering-oriented

Important to read New sGroups and go to User Group meetings (ask if you can observe) before buying into a product – e.g. how would you find that IBM have two different products they are selling under the same name??
Ektron – cheap CMS version of software - $3,000; or XML version $15,000

Important to look at file naming for website – will it maintain the names, especially to maintain your Google ranking.

If you put yourself into the mid-market, look at at least one low cost vendor to see what you will get. Will give you a good idea of what you are going to get from mid-market.

Blog: type of content set up – content comes up in reverse chronological order – single, simple use case – very simple management system to manage

Do you need a CMS right away?

- if you haven’t cleaned up your content, if you are just starting out, probably too early. Need to create your business case and solid requirements. Recommends using a small in-between solution as a stop-gap measure (“starter home” idea).

Pitfalls:

- not budgeting for services; will need to bring in the vendor experts even if you have your own developers in-house
o consulting for getting requirements nailed down; customization; integration; migration; testing
o there is a lot of work. E.g. If CMS is $250,000, it will cost $1.5 million to implement
- it is always a trip to the dentist – go home and “floss” your HTML – comment tags, XHTML compliant, use “HTML Tidy” to clean it up; make sure headers and footers are standard
- premature selection – don’t rush into technology project before you figure out what you really need – this is why people overbuy
o e.g. Describe HOW you need to integrate with document management system
o easy to use – What does this mean? Better to create personas to describe the types of people who will be using the system.
- “love at first sight”

Selection

- Sooner you can get the content authors USING something, the sooner they will see what their requirements are and tell you what they need to work.
- Buy a system after multiple vendors have installed their system and allow you to try it out
- The larger the group, the simpler the tool you need. Always buy simpler than what you think you might need
- Go through a thorough selection process – RFI, RFP
- Watch out for “love at first sight” – don’t just pick a CMS because you like the demo, company, salesperson
- RFI, test period necessary and figure out governance before buying – 9 months setting requirements and preparing content; 9 months for selection process. If you shortcut this, e.g. 3 month selection, you will end up waiting at least 6 months learning to use the system and preparing to get it up and runnin
- Appropriate to have a series of roll-outs e.g. over 3 years, rather than one big roll-out
- you are building a relationship with the vendor which will hopefully last 10 or more years
- less protected from a big vendor than a smaller vendor

Selection Process
o put together an interdisciplinary selection team
o common to see these groups argue through vendor demos; good to agree on terms before you go into a meeting; there is no standardization in the industry. Good to come up with your own glossary.
o The vendor may be more important then the product. Go for the people you have the best rapport with rather than with the ideal solution i.e. put some of your due diligence into investigating the implementation term
o Send some of your team to vendor product training
o Meet the team you will be working with in advance
o Keep your business requirements/objectives squarely in front of you esp. if you are asked to make deadline

Staffing

Allow generalized staff to specialize. E.g. general webmasters have opportunity to go on to more advanced role.