Inktrap Team

Self-checkout machines have become a familiar sight at nearly all major supermarkets. By 2020 there is expected to be over 325,000 self-checkout machines worldwide¹ and with this fast and vast amount of growth over the last 30 years, how come so many people love to hate them?

This question has stuck with the team at Inktrap and after many discussions about why we dislike self-checkouts, we came to the conclusion that we should investigate these machines and see if we could create our own improved self-checkout machine interface.

Getting that sweet multi-buy deal, but having to click “finish and pay” then “back” then “finish and pay” again to get it to show up. ?

Though this project is self-initiated, it is still good practice to start off with a brief. A brief allows us to set requirements and limitations, this restriction may sound counter-intuitive, but it actually keeps a project focused and easier to manage.

Our brief was incredibly simple and went along the lines of:

Our client InkMart requires an explorative piece of work that would research and design a new interface for existing third-party self-checkout machines.

With our simple brief in place, we decided to take the Google Sprint approach to our project. If you’re not familiar to the Google Sprint process, it’s a 5 day period that encourages a quick work pace of research, design and testing a prototype.

To start our project from a balanced perspective, we needed to identify what our existing assumptions and preconceptions surrounding self-checkouts were. We did this because we know how assumptions can derail a project due to assumptions not being based on fact or real evidence.

Tackling our assumptions: Why people don’t use self-checkout

There is a general assumption within our studio that people do not like using self-checkout machines. Not only was this something we felt in the studio but also across the internet we found plenty of articles complaining about the systems. We collected our assumptions and discovered that these assumptions are based on several themes:

Moral issues

  • Users feel that the machines are taking people’s jobs though *SPOILER* based on our research the staff did not worry about this and wanted to encourage users to use the self-checkout machines as they are quicker.

Physical issues

  • People with mobility issues generally wouldn’t use the machines as they just aren’t accessible to them.

Practical issues

  • People would rather go to a manned checkout when buying age-restricted products as they know a human has to verify their age.
  • People who had too many items that don’t fit into the sometimes small bagging area of a self-checkout would rather go to a manned checkout.
  • People who have attempted to use self-checkout in the past and struggled, or had a bad experience, would then use a manned checkout.

Accidental issue

  • People would not actively avoid self-checkout but they may have not realised they were available or the saw that others weren’t using them and assumed they were broken — Classic human behaviour of pack mentality and following the crowd.

Due to the lack of existing research easily accessible to us and the need to test our assumptions, we went out to various supermarkets to conduct our own first-hand research; This research took place in our local Tesco, Waitrose, Sainsburys and Co-op.

This machine is cards only. Yet it assumes you’re paying by cash and forces you to select the “card” tab before it will accept your card payment…?

Our research methods included watching customers behaviour within a shop setting that had self-checkouts, interviewing staff who attended the self-checkouts and testing the machines ourselves.

We spent around 2–3 hours conducting our research, primarily we spent only this amount of time researching due to the time restricted nature of the sprint.

Why people use self-checkouts

While carrying out our research we discovered that people do happily use self-checkout machines which came as a surprise to us and this is a prime example of why you must identify your assumptions and test them to see if they hold true. Once again the reasons people used self-checkouts fell into clear categories and those we had already identified during our assumption session:

Practical issue

  • People believed that using self-checkout saved time.
  • People would rather avoid the manned checkout.
  • People were advised by the staff to go to the self-checkout.

Immoral issue

  • People would know that it is easier to steal from a machine than a manned checkout — This may seem like a weird point to raise, but we literally saw people deceiving the machines so they could either get something cheaper or for free.

From these customer-driven insights, we can see that time and friction is a key issue for many people. People are generally looking for the most time effective and seamless method of buying items, and their perceptions and previous experiences with self-checkouts will most likely inform their checkout preference.

While watching users interact with self-checkouts we were able to analyse key and repetitive behaviours and turn them into a list of expectations that users had when buying items through the self-checkout machines.

People expected the machines just to work

This isn’t a super surprising expectation to have, but this is classed as the baseline of what all users expect (a non-negotiable) and if this isn’t met then no number of snazzy features will make the user experience any better.

People expected the machines to be fast

One of the most obvious pain points during our research session was people tapping furiously on the self-checkout screen because they were waiting to move onto the next section or screen. The hardware itself was sometimes slow and this really frustrated the users.

People expected to be guided on what to do next

However, users didn’t necessarily expect the screen to tell them what to do. For example, a user scanned an item then was unable to understand what to do next, even the screen had the text ‘Place in bagging area’ the user was unable to find where the bagging area was.

People expected no interruptions during their checkout process

People were generally surprised when they were unexpectedly interrupted for example the classic ‘unexpected item in bagging area’ frustrated people as the error stopped people in their tracks with no clear information as what to do next other than wait for a member of staff.

People expected to pay when they were ready

With the rise of contactless payments, many of those people we observed wished to pay instantly without having to select their payment method. We’re not sure if this was a learnt behaviour from other self-checkout machines or just an assumption due to the card readers appearing active during the entire purchase process.

As our research was primarily observant we noted a few other interesting insights;

  1. The user groups age range was really broad. Users were between the age of 20–50 years old and some were much older in Waitrose.
  2. Only 2 people out of 24 used cash.
  3. People would physically flag staff members down for help.
  4. People usually had 1 to 3 items when using self-checkout.
  5. Machines not suitable for those with eyesight or mobility issues.
  6. Users didn’t tend to be in a rush though were bothered by delays.

After a solid day of generating a research plan, carrying out the research and considering what we had learnt, the next step was to take a look at user flows.

Check out Part 2 to see how we developed our own flow and wireframes.

If you’d like to keep up-to-date with what we’re up to and our future freebies at Inktrap, follow us on Twitter and sign up to our weekly design newsletter Minimum Viable Publication.

Project and article written by Liz Hamburger and Rachel Brockbank