Similar needs, different results

I recently had the pleasure of concluding two consulting assignments. Though both clients have a lot in common, the outcome was quite different. Why?

Situation and solutions

Both clients are publishers that serve national and international markets. Both offer multilingual products to professional users in PDF and XML-derived formats (such as HTML or ePub). Both need to update their publishing processes to keep up with changing stakeholder demands and are faces with outdated publishing infrastructures.

Image by Sean MacEntee from Flickr

But at one client we advised an Alfresco based solution, whereas the other client we advised a MarkLogic based solution. Being so similar, why two different outcomes?

XML as design choice

Needing editing by countless authors and editors worldwide, client A implicitly chose to set an open and well maintainable storage format as a standard. Generally, (X)HTML will not suffice for publishers, failing at standardisation of contents and the quality of the printed products. So XML it is.

Some excellent Commercial-off-the-Shelf XML products are available (including SDLs Live Content and Alfresco based Componize), but the of-the-shelf part usually means complying with their internal standards: DITA. Although DITA is a great XML format, it is not suitable in all cases, as it was in this case.

Wanting more specialised workflows and interfaces, we concluded that a custom XML based CMS combined with an online XML editor would fit this client’s needs best, with MarkLogic being the best platform to implement this. Binary content (like PDF) will be supported in the system through a layer of XML Metadata.

XML as an option

Client B’s content is mainly created in separate and formalised processes, with lots of content being available only in PDF. Editing being out of this equation, re-use and workflow were of much greater importance.

This client also sees the opportunities for XML and more dynamic publishing, but the hurdles are substantially bigger, and the rewards in editing would not be felt. Rather than going all-in on XML, client B will transform selected publications to XML.

For this client the out-of-the-box functionalities like workflow and flexible content storage tips the scale to an Enterprise Content Management System. Alfresco being the ECMS of choice, this strategic platform will provide the room to manage content and changing customer needs that will characterize enterprise IT for years to come.

Conclusion

It is often said that the devil is in the details. But these details often drive big design choices. Decisive product-ownership provides the roadmap for such strategic choices.

Thesis students

Being close to the Delft Technical University, my employer has tightened our relations with universities. I have the pleasure of mentoring two students performing their thesis study at HintTech.

Krishna Akundi from TU Delft has just completed his thesis study on visualizing large news archives in the temporal perspective. From a more fundamental research of the subject, his study has transformed into the quest for finding the most usable interface to visualize news trends in topics over time.

Marian Szabo from TU Delft has just started on the subjects of finding and fixing errors in ontologies. All ontologies have the potential for errors, and Marian quest is trying to help fixing them by finding, visualizing and showing possibilities to repair the errors.

Both offer interesting challenges in the exploration, understanding and visualization of Big Content sets, especially sets of RFD.

We are grateful to Newz for use of their data and providing substantive direction.

Content Matters Podcast on Semantic Technologies

I am very happy and proud to be featured on this months Content Matters Podcast. Iain Griffin and I talk about Semantic Technologies. We discuss the general concepts and the use of these technologies in recent projects such as Newz and Kennisnet.

If you are active in Content Strategy: Subscribe to the excellent Content Matters Podcast on iTunes or Soundcloud.

Every page is Page One

In my last post I referred to the Lessons from the New York Times innovation report.
One of the conclusions was that they were focusing too much time and energy on Page One. The NY Times number of Home Page Visitors is in a steady decline, whereas Page Views are fairly constant:

In episode 29 of the excellent Content Matters Podcast Mark Baker also talked about this idea and his book: Every Page is Page One.

Marks focus is Topic-based Writing for Technical Communication and the Web, but I think his message can be translated to almost any context (it does for the  NY Times!).

Remember the days that accessing a website used to be typing in the URL or using your bookmarks? (Who still uses bookmarks? As a Chrome user: Chrome knows my bookmarks better than me).

But now, much of the traffic to websites is either through search-engines answering specific queries – so dropping off users deep into your website – or by shared deeplinks to specific pages in your website.

Both will ignore your homepage. And where the page itself will probably look OK (you will have taken care of that one), what impression does this One Page give the visitor – the visitors Page One?

Will the visitor understand you and your story from this page? Will they leave your website it at this page because they don’t understand you? Or will they visit some of your other pages, each one as important as Page One?

Een inspirerende editor

Binnen HintTech organiseren we regelmatig One Day Product sessies. Tijdens deze One Day Product sessies bedenken en bouwen we binnen één dag een werkend product.

Bij de One Day Product sessie van donderdag 22 mei hebben we gewerkt aan beeldverwerving bij artikelen. Klassieke beeldverwerving is vaak een duur en weinig creatief proces, met een erg statisch eindresultaat. Zou het niet fantastisch zijn als je tekstverwerker je inspireert tijdens het schrijven door interessant beeldmateriaal te tonen?

Met dit idee ging het team uit Dayon, HintTech en Tripitch aan de slag. Irina ontwierp een mooi interface, dat door Martin en Sebastiaan werd gebouwd. Tegelijk gingen Jorg, Fedor en Fabian aan de slag met tekstherkenning en het aanboren van beeldbanken.

Na een lange dag van nadenken, ploeteren, discussiëren en bouwen mag het resultaat er zijn: MOOD THIS!

Probeer het uit! Type een tekst in, en als u een paar seconden niet typt gaat MOOD THIS! op zoek naar een moodboard met beelden en kleuren op basis van woorden uit de tekst. Het resultaat zal vaak verrassend zijn (het is afhankelijk van social media tagging) en hopelijk inspirerend, maar vaak ook gewoon erg leuk.


MOOD THIS! is geïnspireerd door Plastr,  een project van Sebastiaan Hoejenbos en Michiel Peters.

Lessons from the New York Times innovation report

A few days ago, the full NY Times innovation report was leaked to the press. The NY Times did a rigorous self-examination, and came to some incredibly important observations on the why and how of true transformation to the digital era.

Reading this document sparked a WOW moment in me. I think it is an absolute must read for anyone dealing with content strategy. Not so much for profound new thoughts, but more for the insight in where NY Times digital performance was good and where it is was poor, affirming that changing a content strategy is hard.
The Nieman Journalism Lab wrote an excellent report on it.

I would like to share a few of the observations and proposals, which I run into in my daily practice:

Create a strategy team

The newsroom is too focused on the daily task to produce the daily newspaper, which leaves insufficient time and focus to assess and implement strategic changes.

I think we all can relate to this problem. It is OK to have a team focused on the daily chores, but somewhere change has to happen. Creating a strategy team is a good idea – as long as it is well-anchored within the standing organisation and has the power to implement change.

Put the Reader first & Grow an audience

The NY Times offers content through all the obvious channels. But by better headlines, better search and better social media skills, NY Times was outperformed, even on their own content. Simple changes such as posting at a time your audience is actually reading, repurposing, repeating, repackaging and personalizing.

The battle for an audience isn’t fought anymore on selling a paper or a website by its front-page: every article has to fight its own battle on the saturated content market.

The report also suggests branching off and seeking audiences in different forms. Hosting events and talking to real customers should be on anybody’s agenda.

Structure to repurpose

The report tells a compelling story on why content needs to be enriched through tagging and structured data.

This reaffirms the choices we made in Newz project: structuring content by location, time (and events), story type and topics (persons and organisations). The NY Times adds to this: timeliness, story threads, story tone and the use of imagery.

I feel that any content-organisation should consider structuring content on some of these base functions.

Overcome the huge cultural change to become ‘Digital First’

The report points out that changing a newsroom from a century-old habit of paper & front-page to ‘Digital First’ is a huge challenge. The cultural change will take new technology savvy people and new managers encouraging cultural change.

The Times created many beautiful digital showcases like “Snow Fall”. But the NY Times built “Snow Fall” and not a “Snow Fall building tool”.

Anyone working difficult markets should be willing to experiment more in presentation formats, accept imperfection, accept failure, measure so you know what succeeds and have the tools ready to expand on success.

Read this!

If you are in any way active in content strategy or news: Please read this report. See that you are not alone in your struggle and gorge on the valuable insights.

Thanks to a Dutch analysis from the Bladendokter that pointed me to the report.

MarkLogic World Amsterdam

May 15th, the MarkLogic World Tour touched down in Amsterdam. Approximately 300 attendees flocked the Amsterdam Arena to be immersed in MarkLogic and it possibilities. As longstanding MarkLogic partner and implementer, my employer Dayon and HintTech sponsored this event. MarkLogic presented interesting new Developments, Case studies and Technical Sessions.

Big Content challenges 

Keynote speaker an MarkLogic CEO Gary Bloom and Joe Pasqua, Senior Vice President of Product Strategy explained MarkLogic choices for addressing Big Data challenges, but also the Big Content challenges (see my earlier blog) such as Heterogenic structured data: supporting Multi Schema XML and changing schema's.

MarkLogic 8

MarkLogic 7 already introduced a lot of interesting new features; MarkLogic 8 raises the bar some more:

• JavaScript & JSON support
  Building Applications on MarkLogic will be available to lots more developers: front-end oriented developers and other developers who prefer {} over <>
• SPARQL 1.1 & Inferencing
  The Triple Store in MarkLogic 7 is a good starting point, but becomes much more complete by adding Inferencing and SPARQL 1.1 support
• Bitemporal support
  A very interesting feature, which I could use immediately: What did we know at a certain point in time?

Cases

I visited the more business oriented tracks, with cases from Standards Norway, Intel, and the Healthcare.gov case, known to us as Obamacare. MarkLogic explained that their parts of the system kept performing throughout the toughest time. If only all components would have been this reliable…
Some of our customers presented as well: Ted van Dongen of Swets told how a great tool needs good people to implement. Thanks for the compliments Ted!
Our (previously) own Michel de Ru contributed for Newz to a panel session on Sharing Data despite the Silos.

Can you have too much of a good thing?

I will admit to anyone that I am biased: I think MarkLogic is a great tool that solves many challenges my customers face. But I had an interesting discussion: with all these new features, will MarkLogic become a too monolithic solution? Can you have too much of a good thing?
My 2 eurocents are: For now, keep piling on the good stuff, MarkLogic!