I am part of a large webapp project, but a small UX team (we just went from 1 to 2, a few weeks ago). We were brought in terribly late in the process, to the point where there was literally no chance to do site orientation, product familiarization, competitor analysis, or pre-development user studies (Yes, there's a huge, existing user group out there. Sadly, we have no access to them... we DO have Subject Matter Experts in-house though). I am currently running emergency workshops to begin getting a feel for the past product and what the expectations are for the up-and-coming new iteration, as well of areas of concern. To give you a feel for the scale of this project, the webapp is projected to have over 1200 pages, with an identified 120+ business processes.

Now with the context slightly defined, onto the question!

A very important part of this project is the implementation of a navigation toolbar. To give a point of reference, this application will support easily over 140 independent tasks (better to think of them as business processes).

As discussed in other threads, nested hierarchy is a huge concern, and a point of constant consternation. Some of my workshops focus on card-sorting to see how the relationships between these elements are perceived by the SME’s.

However, besides prototyping and subsequent testing (our timeframe is under 4 months for a deliverable of almost 14 processes), I need a quick method for finding appropriate guidelines for constructing this navigation toolbar. Guidelines that really speak to the users needs. I can do "genius design" and using my experience, create something I think is appropriate. But, then why am I here? The methods I use need to inform not only at this stage, but scalable to later stages as the number of processes increases dramatically.

Can StackExchange identify any quick methods besides rapid prototyping, card sorting, SME interviews? Anything specific to hierarchy sorting/display? Remember, this webapp is MASSIVE, and we don't have time (or even access) to do “proper” user testing.

  • You make me think you could have tipical profiles (best guess, informal card sorting, clickstream top content segmentation) with user customization and org navigation as last resort.
    – jrosell
    Feb 29, 2012 at 18:42
  • To clarify on some of these points, and perhaps provide more context (I could have written an essay! But no one would have read it to bother answering) We have an older version of the product, and it's possible i can get some taskflows and priorities from the technical staff, but the problem is that the roles of the users limit their access to certain functions. And these roles aren't fully defined yet. When I said we came in late on the project, I truly meant it. If I had been here a year ago, this information would have been made a priority.
    – Drew
    Feb 29, 2012 at 20:37

2 Answers 2


It's not the whole picture, but analytics might be really useful (if you can get some) to help you rough sort the giant IA task at hand. Depending on how sophisticated the data is, you might be able to identify notable patterns:

  • Highly-trafficked areas (Are these really key functions? If so, prioritize them.)
  • Tasks users tend to perform in clusters (This is like a statistical card-sort.)
  • "Common flows", where a particular decision branch or drill-in pattern is followed frequently (Can we make sure these flows are easy to follow? Can we remove some clicks?)

You could use the data-crunching to prioritize what areas you will follow-up on manually with rapid prototyping, card sorting, and SME interviews.

  • Specifically, get a prototype out to some testers and gather analytics on them early on while there's still time to pivot.
    – sirtimbly
    Feb 29, 2012 at 20:20
  • These are very good points. I was thinking along these lines as well. We need some solid data to assist in the decision making process. We're going to probably flesh out a few role's that are part of our early deliverable and run some iterations of prototypes to try and gather this data. I'll post results as we achieve them. Thanks for the quick feedback!
    – Drew
    Feb 29, 2012 at 20:41

peteorpeter has given an excellent answer and just to add to his points, an approach we used for a content publishing portal project which had similar time frames was to establish the key buckets or categories and define the taxonomy under them.

Once we had a rough idea of the taxonomy and the key buckets,we sent them to the stakeholders and asked them to take some time to have a look at it and get back to us with their preferred taxonomy. The key thing was that we also included the senior management in this constantly so that the people who could provide us the taxonomy were aware of the weight and the timelines of the project.

Once the taxonomy was roughly defined after inputs from the key stakeholders, we had a few meetings where we presented the combined taxonomy to the stakeholders and led them through a card/sticky note sorting exercise to determine common points of agreement and disagreement. We also clearly defined if there could be sub categories within the taxonomy and how many levels it could go and if the users would be fine with having to do as many 3-5 clicks to get a specific document. Once that was confirmed (in writing), we started to structure the navigation and presented it at regular intervals to key stakeholders just to get their inputs with regards to the usability (of the navigation) and accessibility of content.

The last stage was the visual design part of it and while working on it, we prepared a style guide which explained the color scheme we were using and why (we explained it terms of branding and conformance to existing standards and importance of prominence of key categories) and presented it to the team.

I know this is not the answer you were hoping for, but just some inputs on the process we followed :)

Some key points:

  1. Get any design decision or taxonomy decision which was accepted in writing and freeze it down so that people can't ask you to change it every other day
  2. Keep the management involved of the challenges and the dependencies
  3. Ensure you try to drive everything from the business point of view

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.