Contented Management laughing buddha logo

Contented Management

Contented Management

My CMS vendor just got acquired; should I panic?

It’s all the rage for the CMS community; OpenText is acquiring Vignette.

What does this mean for clients of the two companies?

RedDot has been the web content management offering from OpenText for the last few years. It’s a pretty basic tool compared to Vignette, but this has distinct advantages: friendly user interface, quicker to implement, generally cheaper to develop basic functionality. I expect that RedDot will continue to be sold, but that there will be minimal product development. It will probably serve as a cheaper basic WCM in the same way as Alterian market Immediacy as a cheaper alternative to Morello.

The big challenge for the new company will be how to consolidate and exploit LiveLink and Vignette’s core content management offering, VCM. The offering that OpenText should be providing is end-to-end content management from documents and business process to web, but it’s going to be a substantial task to provide this through two pieces of software that are so established. LiveLink does the trick with documents and VCM does it with complex web content. But this certainly doesn’t mean that the two fit together neatly.

A significant benefit for OpenText is the acquisition of Vignette portal (VAP). This will enable OpenText to market web applications rather than just content-driven websites. Again, there will probably have to be some significant work done on the API level to LiveLink to turn this into a fully SOA-enabled platform. Nevertheless, if you’re doing business via the web — and surely everyone is these days — then a portal offering is a necessity for any enterprise content management vendor.

OpenText will be able to offer a product suite to match any of its competitors. But it will be a suite, not an integrated platform. Indeed the company has a poor track record in integrating its product suite: Gauss and ObTree anyone? Even RedDot stands pretty much alone from LiveLink. Oracle, despite its many acquisitions, has a far smoother integration of document and web content management, as does Interwoven.

So what does this mean for you if you’re about to buy? You still need to be wary of LiveLink’s web credentials; this is unlikely to improve for some time as the company attempts to make the various products work together smoothly. And if you’re about to buy RedDot, bargain hard, because I think the prices are likely to come down.

A few other thoughts on the acquisition:

Philippe Parker on , , , | 7 May 2009 | Tweet this |

Contented Management

Stand and deliver

Adriaan Bloem points to the downtime on Oracle’s website following its acquistion of Sun as an indicator that dynamic delivery is generally unsatisfactory. Certainly that website looked like it needed a decent application server on Monday. I wonder where they could get one.

The static vs. dynamic delivery debate is specific to the kind of content you’re producing. Generally, dynamic delivery is well-suited to:

  1. Time-critical content, such as news or user-generated content.
  2. Content that requires lots of automatic relationships and “see also” type links, including product catalogues. This is easier to generate through dynamic queries than it is through a complex relational static publishing model. This is a particular strength of Vignette.
  3. Segmentation and personalistion of content; a strength of FatWire. It can be really hard to deliver content aimed at specific users if you deploy static publishing, although SDL-Tridion has a go by implementing custom server tags into the application server layer.

More generally, you need to weigh up whether you need a dynamic model because you’re publishing lots of content that needs to be up to date, or a static model because you want to guarantee that the content will continue to be served. Either way, you need to be sure that you can change or remove content quickly as well as publish it, and you should follow best practice for delivery performance:

  1. Ensure caching is in place and decent parameters are set; use a CDN if you have a global audience or lots of very large media files.
  2. Optimise your images, CSS and JavaScript and try not to have too many of these being called from a single page.
  3. Use compression techniques, such as GZIP delivery.
  4. Ensure your CMS gives you a tool to purge your cache when you need to.

If Oracle are really concerned about their own website performance, expect them to buy Akamai some time soon.

Philippe Parker on , , , , | 22 April 2009 | Tweet this |

Contented Management

More enterprise myths

It’s true to say that enterprise is a loaded word: it means a lot more to some than to others. We have enterprise content management (ECM), enterprise search, enterprise portals, enterprise resource planning… People like Nicholas Carr have been railing against these all-encompassing applications for years now, questioning how applications that cost so much to install and configure can deliver tangible business benefit, particularly compared to smaller, more targeted systems. The Gilbane Group on the other hand dislike the term enterprise because they believe it’s pure marketing spiel, particularly in the case of content management where few vendors offer the full range of content management products.

It is of course possible to go to a single supplier and get the full WCM, DM, RM, DAM, JCR and IDM gamut of acronyms. The leaders are IBM and Oracle, but Day, Vignette and Open Text all have products covering the main functionality. You have to take care of course that just because the products are owned by the same company and are labelled as a single product family, this does not mean that they can actually talk to each other. Many is the client persuaded to implement a product portfolio from off the shelf, only to spend months and hundreds of thousands on systems integration.

Leaving aside the truth that vendors relate and the more palpable realities their clients are faced with, ask yourself this: why would you need an enterprise application for content management anyway?

Enterprise means not simply across your whole organisation but unique to your organisation. Your ECM will be different to someone else’s, with different security privileges, workflow, storage and retrieval requirements.

Except it’s not.

What you’re trying to do in your organisation is being attempted in every other organisation of a similar scale or vertical. All your competitors, all your partners, all your suppliers and clients will need to control their information and distribute it to the right people. And they want to do it in similar ways, which is why all these vendors are able to sell their content management technologies to so many clients. The thing is, if your requirements aren’t unique, do you need a system that’s unique?

Of course you don’t.

People like Andrew McAfee and JP Rangaswami have been using and writing about disruptive technologies for years. Technologies like wikis, blogs and tagging are disruptive because they upset standard business models and processes where you procure a single technology and then tell everyone how to use it. Under the disruptive model, you let people use a set of tools the way they find most productive. You can add anything to a wiki without it going through workflow, you use blogs instead of email, you use tags instead of a taxonomy. Depending on where you look, these technologies have been more or less successful.

But for me the issue is that it’s not blogs and wikis that are disruptive, it’s the enterprise technologies themselves. Why do organisations feel the need to procure these tools that few people know how to implement and even fewer know how to use? Why not just pick a few technologies that are out there already? The procurement and implementation of these systems actually disrupts the things your organisation is good at, often having a greater negative impact than the business benefits the system will eventually entail. Yes, an enterprise system gives you one butt to kick, but you still have to do some butt-kicking.
For example, why set up a massive LDAP directory that a bunch of systems administrators need to maintain, when you can use OpenID? If you used this to authenticate people, they can use the same username and password for their social life as their daily business. Isn’t this simpler for everyone? Why set up project team servers? Just let each project team set up a blogger account with a new blog for each project and restrict who can view it. They can use the same email address for their email, calendar, and even documents. And those documents could be shared as a wiki. Some of these technologies will work better than others, and there are of course security implications.

Your organisation does not need to control technology, it needs to exploit it. So before you procure a new CMS ask yourself:

  • Am I trying to do something that is already being done by some of my staff using existing tools?
  • Why can’t I extend those tools to support my business?
  • Do I really want to manage a new supplier, a new project and on-going support?

Isn’t it easier to view web technologies as a facility your enterprise makes use of, like roads or a rail system? Let your employees make their own way to work, don’t go out and buy a bus to round them all up in.

Philippe Parker on , , , , | 20 March 2008 | Tweet this |

Contented Management

Managing content management software

InfoWorld publishes annual awards for data management applications. It selected Oracle UCM (formerly Stellent) as the best enterprise content management application and SDL-Tridion as the best web CMS. Both of these products have their strengths: UCM enables you to share documents stored natively in Microsoft Word as web pages relatively easily, while Tridion boasts impressive workflow capabilities likely to meet any web publishing need. Nevertheless, these products have a common shortcoming in a significant area.

While content management software is designed to manage content — words and media are processed, stored and delivered in a controlled fashion — this management itself needs to be managed: workflows need to be created, users added and withdrawn, permissions changed. The most important aspect of this management is changing how content is defined. This generally takes two forms: management of a taxonomy or classification system, and management of the “content types”, or schema that defines whether you’re looking at a press release, event or personal profile. It is extremely unlikely that you’ll get all these definitions right when you first set up your CMS, so the ability to alter and extend the systems is paramount.

Tridion and Oracle both struggle in this area. As a static publishing system, if you want to update a component (content type) in Tridion, you need to republish your schema, add content definitions in each place you’ve made a schema change and then republish the content. In UCM meanwhile, if you can navigate to the relevant administration area to actually make the change and fathom out the unnecessarily ugly interface, you still need to republish the schema and all the affected content even though it’s dynamically driven! You even need to do this if you make a change to the metadata schema, because this isn’t stored in the database as an index, but as a string in a field. This is something you’d expect Oracle to correct pronto with the 11g release due Q4 2008.

You might have thought this kind of thing would be pretty high up most vendors’ agenda, but this is a problem for many CMS. Vignette works ok unless you customise the schema, which many users do. You probably need to look at Mediasurface Morello if you think your content types are likely to change. The product allows you to rename and add fields as required. You need to change the presentation template if you want to see the field you’ve added displayed on a web page, and you run into trouble if your new field needs to be mandatory since your existing content won’t include this information, but it’s considerably easier to manage the content management this way than to republish an entire site.

So caveat emptor. Updating your schema should be an essential part of any CMS product selection, irrespective of awards.

Philippe Parker on , , , , | 17 March 2008 | Tweet this |

Contented Management

Oracle’s challenge: know your product

Last week I attended the Oracle User Group UK conference, with warm enthusiasm and a heavy cold. User groups can be a great way for clients to share implementation experiences, as well as an opportunity to collar suppliers and get a less sanitised view on product roadmaps. I have heard that the Stellent user community wasn’t particularly active, but Oracle are well used to running user groups for the rest of their product range, so this was part of a very large event.

One speaker (I’ll preserve his anonymity) who seemed to strike a chord with delegates raised the point that his organisation’s implementation partner seemed relatively uninformed about Stellent, and that poor decisions around customisation and bespoke development had led to a poor reputation for the product. We’ve already discussed the product vs. implementation issue in a previous post, but the fact that lots of Stellent clients seemed to have the same problem suggests two things to me.

Firstly, the product may be difficult to implement well. Customisations tend to be required for content entry, so perhaps Stellent didn’t know its audience as well as it should have done. This view is perhaps corroborated by the latest release of version 10gR3 which is now bundled with the Ephox rich text editor (already supplied with IBM content manager and Vignette). This attempts to address some of Site Studio’s issues with cross-platform compatibility and accessibility.

Secondly, there’s a problem with product understanding, not just among implementation partners but within Oracle itself. The Stellent partner base in the UK has traditionally been relatively small. Small systems integrators have focussed on the product’s document management capabilities, with web publishing seen as something of a bonus feature rather than an end in itself. The partners are not web specialists, while the real web specialists — design and build media agencies — haven’t really invested in the product because they see it as more than just web, potentially stretching their capabilities. This is exacerbated by the need to train developers in a proprietary scripting language, IDOC.

Now the limited numbers of the core Stellent team are being swelled by Oracle’s professional services arm. But these aren’t content management specialists, and that’s obvious to many clients who may balk at paying Oracle’s day rates in return for staff on a steep learning curve.

So the user group is turning out to be a really useful forum for all involved. Clients can avoid repeating each other’s implementation errors, while the supplier gets to grips with the common business challenges their client base is trying to address. It’s a bit of role reversal, but hopefully this form of social networking will lead to ECM 2.0.

Philippe Parker on , , | 10 December 2007 | Tweet this |

Contented Management

Enterprise too? ECM’s long tail

Over the years, the content management market has seen a great deal of consolidation through acquisition, creating vendors with more extensive product ranges that they tout as enterprise almost by default. If you have web, document, digital asset and records management then you must be enterprise.

There are a number of problems with this consolidation that are well-documented, notably the maturity of product integration; just because you buy Oracle UCM (Stellent) doesn’t mean it works out of the box with Oracle WebCenter. But there’s another issue too: not all the clients are enterprise. Once you’ve sold your massive projects into big corporate clients, how do you tap into the long tail?

Increasingly we’re seeing the larger vendors buy up smaller companies not just to become more enterprise, but to reach a broader market that can’t afford enterprise licence fees. We see this in SDL’s acquisition of Tridion, very much a mid-tier WCM. It’s also been in evidence with the RedDot / Open Text product offering, with the RedDot WCM being able to offer trendier features aimed a less “enterprise” installations, such as User Generated Content plug-ins.

Perhaps the most obvious example of this non-enterprise approach is Mediasurface. Even though Mediasurface is a WCM rather than ECM offering, it has many clients who would struggle to pay for the core product licence, Morello client, database licence, and Solaris servers. Yet it has many small clients who it does well from, particularly in UK local government. To increase its stake in this market, Mediasurface has acquired both Immediacy and the SilverBullet hosted CMS offering, rebranded as Pepperio. This enables the company to dip more easily into the long tail.

So why is this important for you, the customer?

On the positive side, it means that if you’re a small customer you can still get a product from an established vendor rather than a high-risk niche supplier. If you’re a large customer, it enables the vendor to leverage the features of the products in its portfolio to provide you with a more comprehensive system, potentially in a more agile way.

On the negative side, if you do go for a small product from one of these companies, you have to ensure that you don’t turn yourself into a small fish in a big pond. If you go for a small WCM package and you need something quick, you’re more likely to get it when you represent 10% of the supplier’s revenue than if you represent less than 1%. And if you are the big fish, don’t expect the small pond to be anything other than a set of nearly joined up puddles.

Philippe Parker on , , , | 3 October 2007 | Tweet this |