IKEA. The Swedish flat pack furniture giant founded 74 years ago that offers beautiful, functionality driven affordable designs for the home, the office and beyond. I love IKEA and everytime I go in, I come out with more little plastic food bowls and freezer bag ties than I will ever admit to. IKEA is also a mainstay in the team building and relationship counselling activity circuits due to their humorously self-described “husband killer” home assembly processes. The stores themselves are home to complex layouts navigated by blue lined pathways that have just about resulted in missing persons reports being filed and many heated discussions in my household. With all this in mind, I’ve often wondered how the IKEA experience — for better and for worse — might translate to their website. Do people get lost online as much as they do instore? For this week’s website review, I got to find all of this out when I put the findability of product related content on IKEA’s US website to the test using Treejack.
Website: IKEA US
Research scope: Top global navigation bar and child pages below
Not in scope: Inspiration page, footer, help content and general information content
The first thing I noticed about IKEA’s US website information architecture (IA) is it’s as big as you’d expect it to be given the range of products and services on offer, but it’s not very deep. At most it’s 4 levels down and only in some places which at face value, could be a sign of good things to come. Wider, shallower IAs tend to offer a better user experience because they surface more keywords upfront and require less navigation related heavy lifting than narrow, deeper IAs.
Because I wanted to find out if and where people get lost on the IKEA US website, Treejack was a no brainer for my testing tool of choice. Treejack is our online tree testing tool (also known as reverse card sorting) that assesses the findability of content and pinpoints exactly where people are getting lost in the structure. Given the breadth of IKEA’s IA, I decided to include the maximum 10 tasks in this study to provide good coverage of the vast product categories. The tool will let you include more but it’s a good idea to stick with 10 or fewer tasks per study to ensure your study doesn’t take up too much of your participants’ valuable time and to keep that abandonment rate as low as possible.
Also, since I was only interested to learn about the findability of content related to the goal of shopping for specific products, the ‘Inspiration’ page, all help content, general information content and the footer were excluded from this study. The ‘Inspiration’ page got the boot not only because it’s tailored more to general browsing rather than goal-orientated shopping, but also because it has a unique structure that really should be tested separately. It’s quite flat, contains many different content types and is filter based meaning it behaves quite differently to the rest of the website. It’s almost like an IA within an IA and it merits its own research study. Maybe there’s another potential case study here — we shall see. The other excluded items were removed because they have nothing to do with what I’m testing in this case. When creating studies of your own, don’t feel like you have to include the entire website IA. It’s perfectly fine to adjust your research scope based on what you’re trying to learn and achieve. Follow your instincts and tailor each study to suit your needs — there’s no right or wrong here.
Once I removed those pages, I was left with ‘Products’, ‘New’, ‘Offers’, ‘Services’ and ‘New Lower Price’ as the scope for testing (see below image) — still more than 75% of the content on the website. I kept ‘Services’ in scope as there are a number of products living under there including the precious commodity that is IKEA assembly instructions.
Global navigation bar areas tested in this study highlighted in yellow
Setting up the study
My first task was to build out the tree structure of IKEA’s IA into a spreadsheet so it was ready to go into Treejack — IKEA made it really easy for me. When I hover over ‘Products’ I get a peek at the second level of the IA, but when I click on that same label, I get to see everything that lies beneath (see below image).
3 levels of the IA neatly packaged up in one screen view
Pulling the IA from here was quick and easy. I just had to be careful because the order the content is presented in here is very different to the one you see when you hover over ‘Products’ on the live website. The content is exactly the same but in terms of evaluating findability, I needed to ensure my tree matched what the user is presented with when they hover at the second level of the IA.
Choosing tasks for this study
The 10 task based questions I included in this study were designed to cover as much of the product range as possible but to also include some of IKEA’s hidden gems. IKEA has many niche and somewhat unexpected products that add a lot of value but don’t always fit neatly in specific categories and might appear in a few different spots. It’s not all sofas and Billy bookcases. There’s a range of chairs designed to transition a small child from a highchair to a full size dining chair without the need for a pesky booster seat. IKEA has an extensive range of products to help you assemble their furniture (e.g., drills and tape measures). It has a pet furniture range that makes me supremely jealous of my US cousins because it has yet to hit my sandy shores in the land of Oz (Australia). The 10 tasks for this Treejack study were:
- You’ve just moved house and need to reassemble some of your furniture but you’ve lost your assembly instructions! Where would you expect to find them?
- You have a lot of furniture you need to assemble and are short on time. You’re thinking you might need a screwdriver drill to help speed up the process. Where would you expect to find these?
- A friend just purchased an organization system for their kitchen cupboard and you want to check out the range for yourself. Where would you go to find this information?
- You’re renovating your home and need a new stove. Where would you go to find one?
- You’ve got a busy Summer planned and need a container that will allow you to safely transport cold food from your home to the beach. Where would you expect to find something like that?
- You just adopted a cat and she needs a scratching post. You’re short on space and a friend suggested a special type of scratching mat that you can wrap around your existing furniture. Where would you expect to find this?
- Your 3 year old son is too big for his high chair and too short for a regular chair. He hates using the booster seat and you’re wondering if there is a size in between that might put an end to the nightly arguments at the table. Where would you go to find out?
- Your sister has just had a baby and has asked you to help find her a mattress protector for a crib. Where you expect to find one?
- You’re in your second year of college and your study space at home is a mess! Your parents have offered to buy you a file cabinet because they’re sick of cleaning up after you. Where would you expect to find one?
- You live in a large open plan loft and would like to change the design of your space with room dividers. Where would go to find something like that?
Assembling the study in Treejack
Once I had my tree and my tasks ready to go, it was a simple matter of copying and pasting my spreadsheet housed tree into the tool along with my tasks. Treejack guides you through the process of building your study and has you up and running in no time. I set the correct answers to my tasks to give the tool something to structure my results against and I set the participant identifier to anonymous because I honestly don’t need to know my participants’ identities. I also opted to randomize the task order to add more certainty to the results because when tasks are presented to all participants in the same order, it can be hard to distinguish between a poorly performing task and participant laziness- especially if it’s the last one! Unlike your average IKEA piece of furniture, a Treejack study is easy to assemble and probably won’t result in the police being called.
For this study, I used the Optimal Workshop recruitment service which can be accessed via the Recruitment tab in Treejack. This service is reliable and super quick giving me completed participant responses in matter of hours from people who belong to my user group. It allows me to specify age, gender, location and more, providing easy access to research participants that I wouldn’t be able to obtain for myself in the same timeframe and with the same level of certainty around suitability. The recruitment brief for this study was 50 participants with an equal mix of genders all residing in the United States.
My IKEA Treejack study has 58 participants and of that number 51 (87%) completed it and 7 abandoned it. That’s pretty good! Abandonments are recorded when participants close the tab or window of the the study without submitting their responses — a 13% abandonment rate isn’t anything to worry about. Generally speaking an abandonment rate of less than 25% is usually within acceptable limits; it really depends on the study. High abandonment rates might indicate there are too many tasks in the study or it might mean there’s a problem with the IA. The tool will record when and where people check out of the study and this can easily be found in the results section.
What is concerning for this study however, is what I found under the ‘Overview’ tab of the results in Treejack. At a glance, I can see from the below data visualization that IKEA’s US website did not perform well in this study. It’s not what we want to see, and I had high hopes for this one with its wider, shallow IA, but it’s important to remember that a poorly performing Treejack test is a good thing because the tool pinpoints exactly where the improvements need to be made. It takes a lot of the guesswork out of the equation and enables precise iteration work based on evidence. My 50+ participants have spoken and now it’s time to dig in and find out what opportunities for positive change they’ve shared for IKEA!
Task Overview visualization for this Treejack study
My next stop is the Task Results tab in Treejack. This section provides a high level overview for each task and also aggregates the response data for each one into an overall score out of 10. A good score is anything over 7/10. In this study, one task (Task 6) scored an 8/10 with 86% of participants correctly locating a scratching post mat for a cat. That’s great, and an essential item for cat lovers the world over, but I wouldn’t celebrate just yet. The average overall score for this study was just 3/10 with 7 of the remaining tasks recording overall scores of 3/10 or less. Let’s take a look at a few of them in more detail.
Task 1 results
For Task 1, participants were asked to locate assembly instructions after having moved house and lost track of them along the way. It could happen to anyone! In this study, 51% of participants (26 people) were unable to correctly identify the location of the assembly instructions. On the live website, they are located under ‘Services’ and appear alongside another section called ‘Assembly’. Of the 26 participants that failed, 23 (69.7%) nominated ‘Assembly’ as their answer which is actually where you go to pay someone else to do it for you. Now, I don’t think that’s a bad idea, but the task here was to locate the instructions in order to assemble it yourself. The pietree data visualization produced by Treejack for this task (see below image) is quite contained and therefore clearly illustrates that ‘Services’ parent node as a bit of a crossroads for participants in this study.
Task 1 pietree visualization in Treejack (action taken view)
Based on this study, I would suggest that IKEA consider changing ‘Assembly’ to something along the lines of ‘Assembly Services’ to help distinguish it from ‘Assembly Instructions’.
Task 2 results
The results for Task 2 offer a chance to resolve a missed opportunity for IKEA. In this study 75% of participants failed to correctly identify the location of IKEA’s oh so handy range of assembly tools — those drills and tape measures we discussed earlier. I’m honestly not surprised because that content is stored in two really weird spots — ‘Secondary Storage’ and ‘Kitchens’. Now ‘Kitchens’ makes some sense because IKEA sells whole kitchens but what part of ‘Secondary Storage’ screams ‘Tools & Hardware’? Honestly, I have no idea and neither did most of my participants with 2 selecting the option under ‘Kitchens’ and 5 going with ‘Secondary Storage’. The pietree visualization for this one is scattered and big — too big to fit on my screen legibly.
My recommendation to IKEA would be to consider surfacing that content alongside anything that needs assembling or else create a new level 2 category for it in the IA to surface it higher because these products add a lot of value to IKEA’s overall offering and I’d hate to see people miss out on it.
Much like Task 2, Task 5 was designed to test another unusual but incredibly valuable product that supports the IKEA experience. If you’ve ever been to an IKEA store, you know they also sell food — Swedish food mostly. IKEA isn’t just about affordable furniture; a big part of the experience is learning about Swedish culture and trying it on for yourself. My local store alone has a restaurant, a cafe, a sweet shop and a food market selling perishable items. Depending on where you live, you may need a cooler bag to get your Swedish meat, fish and dairy products home and IKEA sells them! As their website points out, they’re also super handy for any event that requires food to be transported to. Unfortunately, in this study 86% of participants couldn’t find these magical bags!
This content can be found under ‘Freezer bags & cookbooks’ which lives under ‘Food’. This task resulted in another messy, scattered pietree visualization and ‘Home’ was clicked 142 times indicating people were lost trying to find this one. Looking at the Paths tab in Treejack for this task, I can see some fairly lengthy navigation pathways (see below).
Participant Paths tab in Treejack
The Paths tab in Treejack shows me a text version of the complete pathway each participant followed for each task — including when they turned around and where they went next! It also allows me to filter the results by success rate or I can filter by participant using the Participants tab. It’s very useful for determining exactly how much trekking around the IA was required for participants to reach their final destination.
For this task in this study, when I scroll down, I can see lengthy pathways were not only taken by participants who were unsuccessful but also by those who were (see below image). Finding out that a participant successfully completed a task isn’t necessarily a win or the end of the story. For any Treejack study it’s important to always check that the path to success wasn’t paved with navigational misery and confusion! If your participants completed their goal but had to traipse all over the IA to get there, your user experience is not where it should be and you’ve got more work to do.
Participant Paths tab in Treejack (page scrolled down)
So, what can IKEA do to improve findability for cooler bags? I would suggest moving it or surfacing it in a few different places. As for where, I have a few ideas. The pietree shows that ‘Eating’ was clicked 25 times with 21 instances of people turning back from there most likely because they couldn’t find what they were looking for — the fact that they even looked there makes it a candidate for a better cooler bag location. Another possibility is ‘Small Storage’ which was visited by participants 31 times during this study.
Task 10 results
Lastly, we’re going to dive into the results of Task 10. This task required participants to nominate where they would expect to find room divider products and no one was able to correctly locate them. In this study 92% of participants failed and the remaining 8% (4 participants) skipped the task. The Paths tab for this task shows me that all 4 of my skippers made a red hot go of it before throwing in the towel and indirectly skipping the task (see below). Direct Skips are recorded when participants skip the task without attempting the task — there was none of that in this case.
Paths tab for Task 10 filtered by Indirect Skips only (no Direct Skips recorded for this task)
For those who don’t know what those are, room dividers are portable screens that can be used to customize and divide a larger space into multiple smaller spaces. They are very useful for lofts, open plan apartments that don’t have a walled in bedroom or bathroom, or large spaces in homes, offices and businesses. They’re an example of another unusual product that adds tremendous value to IKEA’s overall product offering. On the current live website, they are buried under ‘Racks & stands’ which lives under ‘Secondary Storage’ — they don’t even have their own category. It’s also interesting to note that only one of the room dividers available offers storage capacity; the others would potentially lose their portability if people started hanging things on them because they don’t have wheels. The pietree visualization for this one was another big scattered mess with ‘Home’ being clicked a whopping 248 times. People got quite lost trying to locate this piece of furniture. I would suggest that IKEA gives them their own category called ‘Room dividers’ and surfaces that content under additional level 2 IA headings such as ‘Bedroom’ , ‘Living Room’ ‘Office Furniture’ and ‘For Business’.
Other findings from this study:
- 47% of participants were unable to correctly locate a mattress protector for a crib
- 63% of participants were able to locate a file cabinet and the remaining 37% (18 people) were scattered with 8 of them nominating ‘Display cabinets’ as their answer
- 67% of participants were unable to find the correct location of IKEA’s ‘Junior Chairs’ when asked to find a chair for a child too big for a high chair and too small for a full size dining chair without the need for a booster seat. It could be a labeling issue that’s tripping people up
- 41% of participants were unable to correctly identify the location of a stove
- 94% of participants in this study failed to successfully locate a storage system for a kitchen pantry
To answer my original curiosity of ‘Do people get lost online as much as they do instore?’. The results from this study says yes but all identified issues can be easily fixed with some minor tweaks.
Always remember to congratulate yourself when your Treejack study produces some eyebrow raising results, because it means you now have the data you need to refine your IA with surgical precision. This study definitely highlighted some great opportunities for improvement! IKEA has always had an excellent sense of humor and have a history of being generally good sports, so let’s see if they take our suggestions for improvement into consideration!
Key findings and recommendations for IKEA:
- Almost half the participants in this study were tripped up by the label ‘Assembly’ when looking for content that lives under ‘Assembly Instructions’. Consider changing ‘Assembly’ to something along the lines of ‘Assembly Services’ to help distinguish it from ‘Assembly Instructions’.
- Reclaim a potential lost opportunity by surfacing ‘Tools & Hardware’ content in more places that contain products requiring assembly as 75% of participants in this study couldn’t find it.
- Surface ‘Freezer bags & cookbooks’ under ‘Eating’ and ‘Small Storage’ in addition to its current location under ‘Food’ to maximize exposure and support customers.
- Consider revisiting the labeling of the ‘Junior Chairs’ as 67% of participants did not select that as their nominated response when looking for a chair for a child transitioning between a high chair and a full size dining chair potentially indicating a labeling issue.