During our e-commerce research studies, particularly on checkout usability, we’ve found that tab-style and inline accordion form layouts can inadvertently confuse users, or even flat out violate their expectations.
The issue arises when users can’t figure out which form fields will be submitted – whether it is only the fields in the currently active inline accordion or tab “sheet”, or whether the collapsed “sheets” will be submitted as well.
In this article we’ll dive into the observations from our test sessions, highlight the issues and design affordances that both inline accordion forms and tabbed form layouts produce, and touch upon some design alternatives.
When designing user-friendly form layouts it’s key to ensure that the user always instinctively knows which form fields will be submitted. Now, typically this is a non-issue since users assume all form fields will be submitted, and on most sites they are.
However, tab-style and inline accordion form layouts can muddy this relationship, making it unclear if the fields in each “sheet” are mutually exclusive or if switching between them simply toggles their visibility but not the actual form. This can make users uncertain as to which fields in the form will actually be submitted, which is highly problematic as this leads to a sense of unpredictability and the fear of potential data loss (obviously very undesirable sensations to invoke in a user filling out your forms, e.g. during the checkout process).
Common questions that tab-style and inline accordion form designs invoke in users include:
Now, sometimes the options will logically imply whether they are mutually exclusive or not. For example, with payment options it is typically clear that they are mutually exclusive: either you pay with credit card or PayPal, but it’s important to note that in these cases it’s the content and not the design that’s driving the affordance. Therefore, whenever the content could potentially be interpreted either way, it will be, and confusion and uncertainty is bound to arise. In fact, even the aforementioned “payment options” example could potentially be confusing since some sites actually do allow multiple payment types to be combined.
Good usability is about predictability – does the user understand what will happen when they take XYZ action – i.e. can they accurately predict the consequences of a given UI interaction? Tab-style and inline accordion form designs tend to fail this test. It’s the same reason multi-column form designs can be troublesome – users may become in doubt if all columns will be submitted or only one of them. With inline accordions and tab-style forms, users often can’t tell if the “sheets” are mutually exclusive or if they are simply toggling their visibility.
This is why accordion checkout designs tend to work just fine, while inlined non-flow accordion designs cause trouble. When the accordion design is sequential and progresses the user from page to page (e.g. an accordion checkout process), it’s clear that only the current section will be submitted, while previous sections display summaries of the entered data letting the user know they have already been saved, while the future sections won’t be clickable at all as the user hasn’t yet progressed to them. It’s a linear flow that the user works their way through, step by step. (Indeed, our research finds that users perceive accordion steps as separate pages which is why it’s so important to adhere to the user’s back-button expectations when adopting this checkout design.) This stands in contrast to inline accordion forms where the user can open and close sections ad-hoc and where there may even be other form fields above and/or below the accordion element. Because of the indiscriminate section access, it isn’t obvious what has been saved and what will be saved. Tabs of course suffer much the same fate.
If an inline accordion or tab-style form layout must be used, there are some subtle design affordances to be aware of. It’s key that the design affordance of these form layouts align with their actual technical behavior and implementation.
For accordion designs, it is possible to visually have multiple “sheets” be open at once (i.e. sheets open and close independently of each other – see left-hand example), or the implementation can restrict the accordion to only have a single sheet open at a time (i.e. opening a new sheet automatically collapses the currently active one – see right-hand example).
This creates an affordance as to whether all sheets or only the currently active one will be submitted. Allowing multiple open sheets suggests that all sheets will be submitted, although the user may be unsure if only the open ones will be submitted. If only allowing one sheet to be open at a time, most users will assume that only the active sheet will be submitted since the other fields auto-close whenever a new section is selected (i.e. the interface is mutually exclusive, therefore the options are perceived so too).
Tab-style form designs are a little more tricky because visually only one tab can be “open” at any given time. It doesn’t have the same affordance available as the accordion-style design, making it more difficult for the user to figure out whether all tabs or only the open one will be submitted – the design can’t imply this through its interface / behavior.
Accordion- and tab-style forms do however share an affordance when it comes to the primary button (i.e. ‘Save’ button). When the button is placed within the active sheet, it suggests only that sheet will be saved when it’s clicked. However, if the button is placed outside the form element, things get a little more complicated. For tab designs, it suggests that all tabs will be saved. Whereas for inline accordion designs that allow multiple sheets to be open at time, it still remains unclear to the user whether the collapsed sections will be saved or not – the design only allows them to feel certain that the currently open sheet(s) will be saved.
So given these UI weaknesses of tab-style and inline accordion form designs, what are the appropriate alternatives?
If the content isn’t mutually exclusive, consider simply listing all of the fields on the page or splitting them out across multiple form steps. Our research shows that good checkout usability isn’t about the number steps but rather how user-friendly those steps are. Whether you have a 2 or 4 checkout steps doesn’t matter that much, what matters is what users have to do at each step, and how they are asked to do it (i.e. are users able to seamlessly progress through the steps).
Alternatively, standard form elements like radio buttons and checkboxes can be used or integrated into the design. Since these standard UI elements have at least a decade’s worth of consistent web usage, the vast majority of users will be able to instantly recognize them and intuitively grasp the form’s behavior because of the strong usage connotations tied to those standardized UI elements.
For example, if the content is mutually exclusive, then use the standard UI element designed to indicate this: radio buttons. These can be integrated into a tab-like design, but it’s important that the radio buttons are an integrated part of that as they clearly set user expectations. (Just make sure to watch out for potential radio button proximity issues, as described in Checkout Design: Payment Method Selection.) Meanwhile if the sections are optional and multiple may be submitted at once (i.e. ad-hoc behavior), checkboxes would be the way to go.
It all boils down to that core usability principle of predictability: when seeing and interacting with the form, will all users instinctively know what will happen? Is it clear what data has been saved and what will (and won’t) be saved? Tapping into standard UI elements such as radio buttons and checkboxes can help clarify such form behavior – although a perhaps safer strategy is to avoid designs which provoke such concerns in the first place.
Join 22,000+ readers and get Baymard’s research articles by RSS feed or
Topics include user experience, web design, and e-commerce
Articles are always delivered ad-free and in their full length
1-click unsubscribe at any time
Great article !
Fascinating read! How widely used are the inline accordion/tab-style forms in the e-commerce world?
We don’t have a number for non-flow accordion forms, but 14% of the top 100 e-commerce use an accordion style checkout flow: http://baymard.com/checkout-usability/benchmark/top-100/accordion
Note however, that accordion style checkouts ( http://baymard.com/blog/accordion-style-checkout ) are for a linear flow and therefor doesn’t necessarily cause the issues addressed in this article.
Great information, But from a usability perspective, this article was hard to follow.
I had to read it three times to finally get it. The readability of this article would have been easier if you better defined the differences between inline accordion form, and a sequential accordion checkout design earlier on, and made clear and consistent use of the labeling. (I finally found your definitions buried in a photo caption)
Specifically where I kept stumbling, was where you were pointing out the issues with “Accordion Forms” in the first 7-8 paragraphs then in paragraph 9 you switched to praising accordion checkout designs saying they “tend to work just fine”
I was like…what?
Thanks for the comment Alfonzo. I fully agree with you, it is a rather subtle difference in script but with a major impact on the semantics of the text. We’ll be sure to highlight such distinctions better (and earlier) in future articles.
@Alfonzo Zo you are right, the author speaks about usability but his article is very difficult to follow.
Good information tho…
Does this conflict with the recommended article following this one: “Accordion Style Checkouts – the Holy Grail of Checkout Usability?”
Not at all, and for two reasons:
1) As detailed in the article on this page, accordion-style checkouts are linear flow-based processes and therefore do not suffer from the usability problems of non-flow inline accordions with ad-hoc access. However, as is evident from the above comments too, we didn’t do a good job at clarifying this distinction early on in the article. We promise to be more mindful of this in future articles.
2) As detailed in the recommended article, accordion-style checkouts aren’t the holy grail of checkout usability. They can be good, even great, and they can be bad, even horrible. Great checkout usability isn’t about “One-page Checkout” vs “Accordion Checkout” vs “Multi-step Checkout”, rather it is about adhering to a number of fundamental design principles and then getting a whole host of design details right. We outline a lot of these fundamental design principles and details in our Checkout Usability articles: http://baymard.com/research/checkout-usability
Kindly write articles related to the b2b also :)
A more logical and very popular method is to control the users journey through the sections, using a simple NEXT button. This at least will ensure that the user understands there is only one SAVE event and provides the added benefit of ensuring any mandatory fields are completed for each section.
Disagree with choosing BestBuy as the superior payment. Why segment out all modes of payment with Paypal? Use all payments, and eliminate the tabs altogether, and allow the screen to dynamically adjust for ANY payment option is superior. We’ve had users who can’t even see that Paypal tab.
the only differentiating point on accordions is open-stay-open vs open-and-close else. Not much of a difference when dealing with mobile media as that requires someone to visually scroll back up.
I think the author needs to rethink and get more of a large-scale user base than to base assumptions on what confuses, whom and when.
UI Designers and UX pros – take this with a grain of salt.
The one thing is that yes, tabs are horrible for mobile, unless done properly and make the buttons size with at least the size of a finger, as that can be much superior than use of horizontal scrolling I’ve seen done (T-Mobile, etc).
The author also forgets about anchoring end-users and keeping them aware of where they are in any process flow…
The author is a junior to mid-level designer… not an architect..
© 2021 Baymard Institute US: +1 (415) 315-9567 EU: +45 3696 9567 firstname.lastname@example.org