Making the most of Optimal Treejack

If you want to find out how well your website navigation structure works for your customers, Treejack is a great tool for the job.
If you want to know why certain parts performed poorly, and what to do about it, you’ll need to get inside the head of your customer. The tools for this are your eyes and ears.

Treejack, was developed in New Zealand by Optimal Workshop so has been built with a user-centered approach in mind.

It’s a tool to test the navigation structure of your website. Treejack will pinpoint the most difficult areas or items to find, based on click-trails as survey participants navigate through a prototype of your website’s structure. (the prototype is a simple ‘tree’ of text links generated from a spreadsheet you paste  in… couldn’t be easier)

Treejack is a great tool, saving time and headspace, but it is no silver bullet.

You’ll get summarised and detailed outputs showing where each participant went, how directly and quickly they found set items during the survey. … but it won’t tell you how much sense it made to them, or why the tricky areas were confusing.

To design a website that’s intuitive to navigate it’s essential to understand how your customers will interact with it. There is simply no substitute for observation when it comes to gathering these insights.

Teaming Treejack up with qualitative one-on-one research makes a killer double-act bringing you the best of both worlds.

Some tips for integrating Treejack into user research sessions:

  • Run a warm up exercise on a generic ‘tree’ … Clicking through a bare-bones navigation is quite abstract so this helps participants get used to the interaction style.
  • Encourage participants to ‘think aloud’ while using the prototype. When you notice them pause, they’ll be thinking. Having them vocalise their experience is the closest you’ll get to knowing what’s behind their thoughts and any indecision.
  • Save your questions till after each task. Interrupting the participant mid-flow can make them change their behaviour, skewing the Treejack report. Let them click through naturally then discuss it afterwards. You’ll need to rely on your note taking here.
  • Have a duplicate Treejack survey open in another tab. This way you can ask participants to re-trace their steps without affecting the Treejack results.
  • Ask the participants to ‘rate’ each task for how much sense it made to them etc. Treejack tells where and how they found the item, but doesn’t tell you whether this made sense to them.
  • More participants, fewer tasks. As people develop a familiarity with the ‘tree’ they will start memorising where things are, making your findings less useful.
  • Use your eyes. The old adage, “it’s what they do, not say” is as relevant as ever here.

I’d be interested to hear anyone elses experiences …
Go check it out at

6 thoughts on “Making the most of Optimal Treejack

  1. Harry

    So you used treejack in a moderated, co-present study? That’s interesting since AFAIK it’s designed for unmoderated, remote testing.

    How much value do you think you’d have got out of the study if you’d have gone down the unmoderated, remote route?

  2. Nick

    Yes, I’m pretty sure it wasn’t intended for face to face studies, but it turns out to be a real asset.

    Much of the value of Treejack lies in the time it saves you by building a prototype and capturing journeys, summarising data etc.

    However, the results of an unmoderated treejack report could be misinterpreted without the qualitative component.

    If you’re not present when participants are rumbling through your IA, you’ll be blind to false positives and could easily miss key insights.

    E.g. if you had a dreaded ‘miscellaneous’ category and many participants ‘sucessfully’ found things within it, the treejack report would show that this is a strong part of the navigation.

    Like many of these online tools, they can tell you what happened, but not why, and it’s the why that helps us.

  3. Karl


    I’ve just completed a Treejack study as well, and I had several tasks which, in summary, look very successful.

    However in post session questioning it became apparent that there was a process of either deduction or a click and hope for the best strategy at play.

    I found participants didn’t understand several labels which they clicked on, leading to successful conclusions to their task; but these were chosen because other choices made even less sense.

    The Treejack summary says this is a success, but I would say otherwise 😉

    I used a slightly different method to Nick – having users complete a series of 12 – 15 tasks before stopping and then going back over tasks which I thought required some more investigation. I didn’t want users to hop too much from an investigatory frame of mind, to a reflective mode between every task as I thought that this might lead them to be more and more analytical, or careful with each subsequent task.

  4. Sam Ng :: Optimal Workshop

    Hi guys,

    Great to hear you’ve all had varying levels of success with Treejack.

    Harry – you’re half right – we designed Treejack primarily for unmoderated testing. But, Treejack was born from moderated, in-person testing. The very first prototype was a stack of paper with a human “computer”!

    As mentioned by others, Treejack and in fact all our tools, are always going to be supplementary to in-person testing – if you can afford the luxury of doing so. Think of it as the difference between running a survey or an interview. Interviews will always give you much richer insights, but surveys are a lot quicker and less time consuming.

    Karl – your point about ensuring people don’t hop too much into a reflective mode is a critical one. Ensuring that participants don’t spend more time thinking than in the “real world” is central to the results’ validity. In any type of study, but particularly moderated sessions, “test conditions” are invariably introduced so reducing other forms of bias becomes critical.

    There is also a randomisation feature in Treejack that no only mixes the order but can also present only a certain subset of tasks to present a participant. Use this to reduce the learning effect, as well as fewer tasks per participant.

    Lastly, we are considering running a Treejack Q&A session. Anyone who is interested, just email me.


  5. Deb

    It’s disappointing that “Edit Tree” does not include inserting new lines. You may add or delete or change words in a single line, you can even delete the whole line, but you cannot add a line. This has casued me some hassle. In addition, there’s no way to allow two right answers is one of them has a link beneath it. Maybe a participant would consider the overview enough, and the detail may not be needed. However, if you want to track who stoppped at the overview and who drilled to detail, you can’t, because you cannot make Overview a right answer when there’s something under it.

  6. Nick Post author

    Hi Deb. Sounds like you are giving Treejack a good workout. Perhaps you should get in touch with the nice people at optimalworkshop as they might value your comments.

Reply to Nick

Your email address will not be published. Required fields are marked *