Wednesday, August 24, 2005

What's in a name?

Oh dear, the importance of a name!

Whatever we choose will only be useful to some people! Anyway, here's some of the words we've been juggling.

There are 4 sections:

  • How the company relates to its staff.
    • Contains: Employment policies (i.e. legal obligations), induction, benefits and recruitment procedures.
  • How the staff relate to the company and each other.
    • Contains: Social stuff, individual development stuff, money stuff (i.e. making expenses claims), personal details and 'travel' which we are not certain about
  • Resources for doing your job.
    • Contains: The company premises, equipment and how to use it, communications (phone, email web), reference material.
  • Our work.
    • Contains: Anything specifically job related like a project for a customer for example.
We went from abstract reminder tags to wordy description sentences and now we need succinct navigation headings.

No problem! ;o/

New hierarchy - not having fun!

The new hierarchy that we're trying to test is suffering from a nomenclature problem.

The information groups feel right but finding titles for them is proving impossible. It's making it very difficult to test the hierarchy as people can't decide which group to look in 1st as the names don't mean anything to them. (They are currently placeholder labels for us...because we can't find names!).

I'm going to have to think of some way to find better names for these groups! There isn't time to spend ages doing this and there is definitely no time to make a new hierarchy (unless this one is fundamentally flawed...).

It better not be flawed!

Monday, August 22, 2005

Current status

The posts so far have been describing things that have already happened. So I thought it would be a good idea to say where we are right now.
  • We are about to start testing the new hierarchy which incorporates the staff handbook and breaks the departmental structure.
  • We have re-assessed the tasks in the project plan.
  • We have changed our intentions for how to gather updates on company events.
    • We will introduce the ad-hoc generation of stories but still publish them in the current newsletter website until December.
    • We will form a group to gather stories - the 'Contribution Group'.
    • We will not introduce any software to automate this until the process is working on its own.
Now, I must go and write some scenarios for the new hierarchy...

Return and uncertainty #2

After returning to the project following an almost 3 month break we needed to re-assess what we would do with the newsletter.

A major change has been caused by the editor leaving to go to a new job and the replacement editor withdrawing their acceptance of the job. Myself, my manager and the Marketing manager have produced the last 2 editions of the newsletter (including the changes mentioned already).

This has given us the opportunity to think about our original intentions.

Our original plan was to have a writer and editor in each department who would produce that department's stories and send them to the editor-in-chief who would publish them on the Intranet home page.

The intention is still to integrate the newsletter and the Intranet but the manner of implementation has changed.

We now plan to form a group of people who can look for things to report and either inform the 'editorial group' (currently me, my manager and the Marketing manager) that a story is available.

The wider group can also generate the story, either by interviewing the person/people who will be involved in the event or by asking the participant(s) to write a report on it. All the editing will be done centrally and not in the department. There will still be a two-way review.

We also plan to start implementing this by applying the process without any automation software.

We have asked certain people to join the 'story gathering group' and identified events to write about for the next newsletter edition. The Marketing manager will advise the department heads that we are going to approach people to write stories and then we will do just that.

There will be support for the people asked to write and if they really don't want to write at all we will interview them to get the story.

The publication of an 'edition' will be gradually phased out and the newsletter and Intranet will be integrated by December.

Return and uncertainty

So here we were, nearly 3 months since we'd looked at the hierarchy, trying to remember how we got here and why we'd put things where we had....

I felt quite uncomfortable looking at the hierarchy we'd ended up with. And even more uncomfortable when I thought back 3 months and remembered it feeling slightly more comfortable than it did now....

I decided to represent the hierarchy as a diagram, as I suspected something. My suspicions were confirmed. We had initially decided that the new Intranet would not reflect departmental structure. We had deliberately not included the department-specific content in the card sorting exercise or the hierarchy section. Unfortunately our hierarchy showed a very strong deparmental structure... Oh dear. The more I looked at it the less comfortable I felt.

Then I looked at the contents of the staff handbook. I remembered thinking before we left to work on the 'something else' that we would need to integrate the two sets of content. This was just impossible in the hierarchy we had at that point.

What to do?

Even though we had put more than 4 months work into this project already and a lot of research had produced this hierarchy it didn't feel right.

We knew that arranging things according to the departmental structure was a deeply ingrained habit throughout the company. We also knew that this was not helpful and needed to be changed. All the advice, columns, websites etc that we'd read also said exactly that.

The more I thought about it the more the Intranet seemed to be the best vehicle for the Staff Handbook. As the handbook contains, literally, everything, it's very difficult to maintain as a printed handbook in a folder. Updates to policy and procedure have to be circulated on paper.

If the Intranet was to, essentially, be the handbook we would need to make some radical changes.

So, we did. As we completely destroyed the departmental organisation of the handbook and the hierarchy it was very uncomfortable. Now, adding further content to the hierarchy is surprisingly easy.

To get back on top of the project we:
  • Re-assessed our plan, checking tasks and timescales
  • Met the HR manager to talk about changing the presentation of the staff handbook
The HR manager feels good about the proposed hierarchy and handbook delivery so we are now ready to test our new hierarchy.

But we now have to face the prospect of delivering something which may, initially, be uncomfortable.


La, la, la, tinny, cheesy hold music for 2 months....

As we reached hierarchy revision no 3 we had to stop to work on 'something else'.

As these things go the 'something else' was expected to be 2 weeks work and ended up being more like 2 months work (then it was holiday time...).

Results: Hierarchy Testing

Following the technique described here: cardbased classification evaluation

We used a paper based hierarchy. All of the sections were numbered and printed out onto paper and a folder made up with all of the printed pages.

As the article describes we numbered the sections like this example:

1. Departments
1.1 Human Resources
1.1.1 Policies

And so on for all of the content. Then we could note down by number what path people followed through the hieararchy as they were following the scenarios we presented.

We amended the hierarchy as we were testing it. So if something didn't work, a document wasn't found for instance, we either, changed its name, changed the name of the section it was in or moved the document to another section.

This was a very flexible technique and showed exactly when something was buried or when a title of a section was confusing.

Results: Structure the information using focus group results and information needs analysis

After putting together the categories chosen by the focus group and our own categories we had a hierarchy we could test.

This task was entered in the plan to mark the point at which we had done the card sorting; the information needs analysis and created the hierarchies.

This meant we were now ready for the hierarchy testing stage.

Results: Create tasks for hierarchy and interface testing

We looked through the contents of each category and wrote an instruction to find some information in each of the categories, for example: “Read instructions on booking a meeting room using Microsoft Outlook”.

Results: Amend the content to reflect information needs

After the interviews we reviewed the content and removed from the list the things that people would be able to get directly from the internet.

For example a section called ‘useful links’ was not kept. We also decided not to include a discussion forum for the 1st phase release.

We decided we needed to concentrate on the things which were absolutely necessary for people to do their jobs and to leave out the things which ‘would be nice to have’.

Results: Information Needs Analysis

Before testing the hierarchy to see if people would be able to find anything we wanted to check what sort of things people might look for. We also wanted to check that the list of content still reflected people’s needs.

This took the form of an interview on what people did. I did ask what they would like on the Intranet as a final question but not as the main focus.

This followed the recommendation in the Step Two paper regarding stakeholder interviews ( The people in the focus groups were surveyed and the senior managers.

We found that the information we had in the list was still relevant but that a lot of it was duplicated by the information in the Staff Handbook.

The staff handbook had not been available on the Intranet, it was a series of polices printed and handed out to staff in a folder when they started. Because of this it hadn’t featured in the card sorting exercise. Some of the policies were listed on the Intranet but not the majority of the handbook.

Results: Focus Group Summary

Once all the focus group members had done the card sorting exercise we used the following technique, again from Boxes and Arrows: Analyzing Card Sort Results with a Spreadsheet Template.

We did find clear categories in the names but clear disagreement on the items to put into the categories.

The names and the placement of the items were used to amend our original categories.

The categories and their names would form the basis of the Intranet navigation so we started calling the categories a hierarchy at this point.

Once we had analysed all the focus group results and finished amending our original categories (keeping a backup, naturally) we moved onto Information Needs Analysis.

Results: Category Assessment in Focus Groups

We randomly chose a member from each department to be part of a ‘focus group’ involved in the re-development of the Intranet throughout the project.

The 1st task for the focus group was card sorting (

We printed all the categories we had onto card and cut them into individual items (there were too many to write them out, and my writing is illegible to all but me). We gave each card a number and then mixed the cards up.

We gave each person a set of cards and asked them to group things that they felt were related.

The participants were given 2 hours to do this and 12 people took part.

Results: Category Assessment

Following on from the Information analysis stage we re-organised the content listed into more sensible categories, in consultation with the content owners.

This re-organisation took a long time to do.

We also included items that were not on the existing Intranet but that would be useful to have.

We had decided that the new Intranet would be built and maintained in a centralised manner. Look and feel would be consistent throughout and would be part of a central build.

Departments wishing to publish content would have access to do that but not to change the design or structure.

With this in mind we set out to create a new set of categories for the Intranet that would have all the information that was relevant to more than one area in a central category.

Once we had reached a satisfactory grouping we used this in the next task.

Results: Intranet Survey

Looking at the out of date content and the poor usability of the site it would have been logical to guess that no-one used the Intranet at all. However we couldn’t guess, we did an online survey to see how often the Intranet was used.

We had over 50% response rate and of this 50% approximately half said they used the Intranet daily and half weekly.

We couldn’t remove the current Intranet as those people who were using it daily had come to rely on it.

People’s comments requested referred to:

  • The difficulty finding information (you could usually only get something if you already knew where it was).
  • Inconsistent style.
  • Out of date information.
We asked people what they would like on the Intranet and the most popular request was for a staff directory containing photographs. The next popular was a search facility.

Results: Information Analysis

I started by making a content inventory of all the items on the Intranet.

Then we contacted the content owners and asked what could be removed and what needed to be kept. We also suggested a potential move of the remaining material to a more central location, if appropriate.

This was generally well received by the content owners. They were happy to see that something was happening on developing a new Intranet.

The content list created during this task was then used in the category assessment task.

Results: Usability Review

OK, now I've desribed the steps I planned to take, what actually happened?

The first thing I did was the Usability Review

I followed all the links in the site and documented the outcome. I found, amongst other things:
  • Multiple broken links including the link to the Intranet home page being broken on all pages.
  • No unified style guide – each department’s area was different in layout, fonts and colours (they had been designed and built by different people independently).
  • No policy governed opening new browser windows so this happened randomly and without warning.
  • As there was no style guide a number of dept areas had used a default serif font of Times New Roman.
  • Blinking text and scrolling banners were used on some pages.
  • Colour choices made text difficult to read (e.g. bright pink text on a gradient pale-blue to white background).
  • A certain section's page layout was so bad if you didn’t re-size the browser window part of the navigation was obscured by the logo.
  • On most pages the company logo linked to the company website on the internet except where it was broken and one in dept area where linked it to the guidelines on reproducing the company logo (a PDF). These guidelines were not available from anywhere else.
  • Content was massively out of date.
  • Content was duplicated.
  • Content was found in unexpected areas.
  • It was sometimes impossible to increase text size.
  • There were no hover states on some links.
  • We could not see any way to find something if you didn’t already know where it was.
  • There was no logic to the arrangement of the content.
The above situation developed because there was no central control over the look and feel or the development of the Intranet.

There was no control over the publishing and each department had complete freedom to design, build and publish.

Then a publishing freeze was called to start re-developing the Intranet and then other business took priority, this caused the out of date content.

Wednesday, August 17, 2005

The newsletter - part 2

A new publishing process

The current editor and I changed the appearance of the PDF to make it easier to read from screen and then made a website to replace it.

The intention now is to introduce a publishing process which involves:

  • The person directly involved in doing something writing about it (not the dept manager or the editor).
  • The article being reviewed by another member of the originating department who checks for accuracy, commercial sensitivity and adherence to the company style guide.
  • The dept reviewer sending the piece to the editor for a final check who then publishes it if there are no changes to make.
  • If the editor makes a change it is reviewed by the dept reviewer or the author in a cycle which continues until both parties are happy with the article.
There will also be the following changes:

  • Replacement of a ‘magazine’ collecting the updates at the end of a period with articles published individually, when relevant.
  • The news reading and content generation will be actively supported by the Executive Directors.
  • People will be actively encouraged to write the news.
  • People will be actively encouraged to read the newsletter during company time.
  • The news items will appear on the home page of the Intranet (not in a separate website or PDF) and will be optimised for reading from screen.
  • The editor will interview people to generate content.
  • The editor will request articles on specific events from relevant people.
In the meantime we will continue to develop the current newsletter website, making changes until we are close to an interface we can integrate with the Intranet.

The initial newsletter website design was used for 4 editions of the newsletter (4 months) and I have just done a re-design of the newsletter website following a user survey.

The initial design reflected the layout of the PDF too closely which caused some confusion. The new version has been simplified considerably by only presenting one article on the home page and categorising the content.

The problems with this website will not all be solved by this re-design. Some of the practices will help people to get used to reading online (such as using the title attribute whenever possible) however the categorisation will never work because the content just does not suit it.

We have forced the content of this edition into certain groups to avoid using the departmental categories because these are not useful for people. Arranging things departmentally is a deeply ingrained habit that the company is trying to move away from.

There are too many crossovers between the chosen categories. The categories are not well defined and may or may not contain items in the next edition. The content will be different next edition. The items are mostly independent, only the staff, personal and social items can be easily grouped.

This is bad now, for the usability of the current website, but it is good for the future. The plan is to publish these items uncategorised and as they happen, separately to each other. The fact that they can’t be grouped into categories helps to justify this intention.

The progress on the newsletter so far has taken approximately 6 months and has been done in parallel with the Intranet project stages I'll describe in the next entries.

The next stage in the newsletter re-development is to choose a department to introduce the new publishing process to. This will be done without any automation (e.g. Contribute or blogging software) and then once the process has been refined and software reviewed we will automate the process and then involve further departments one at a time.

Progress updates on this will follow after the rest of the project has been described.

The newsletter - part 1

After discussion with Marketing we came to the agreement that the newsletter and the Intranet should be combined.

In order to do this we would need to:

  • Establish a new publishing process.
  • Change the current culture surrounding the newsletter including the way content was generated and edited plus encouraging people to read it at work.

  • The newsletter was a print document delivered over the Intranet in PDF format.
  • There was a plan to convert this PDF into a website but other business took priority.
  • As an interim solution the design of the PDF was changed from being black and white and laid out according to department entries to a ‘magazine style’ layout. Bright colours were used to try and make the publication livelier and the content was placed randomly throughout the PDF instead of according to department. Unfortunately this was done according to print layout rules and was not optimised for screen reading, this forced people to continue to print the newsletter.
The editing and submissions process changed too. Instead of sending edited items back to authors for a final check things were amended and then published directly.

Unfortunately this became more than temporary (it has been in use for over a year) and caused the following to develop:

  • The enthusiasm amongst the company for making an online newsletter died down.
  • The processes for publishing to the newsletter were not public, transparent or documented and they contained no review stages.
  • The transition to online delivery was never properly made.
  • People stopped sending in content to be published although submissions by department heads were mandatory. The mandatory nature was enforced.
However the items sent for publication in the newsletter were sometimes so brief they didn’t say anything. The problem was people were sending in content because they had to not because they had something important to tell the company. The content that was received was not sufficient to publish straight away so the editor had to write articles based on the items received.

Part 2 of the newsletter will be in the next post.