The problem we had to solve was: "People hate lines. Find all the ways of improving their experience when exiting the store during check out."

Basic research included observing the behavior of the people who are passing through check out lines and fast 5-minute talks with random customers at the store exits, cashiers and security guys. This kind of fast research gave us a chance to find out the direction of holding more deep and specific research and create a list of questions for in-depth interviews.

Team of 6 designers was interviewing respondents for a week and ended up with 20 in-depth interviews overall. Our target audience included users of regular check-out points and self-service points as well. We decided to finish the research after 20 talks because we stopped getting any new information and that meant that we got an answer for our main question: "How people go shopping?"

We created the board with insights using a methodology Jobs-to-be-done to systematize all information we got and then picked out three main types of personas based on their needs.

First type is a Bustler. Time is most valuable thing for him, he is in a hurry. This person can be motivated by different things, like: his ice cream is melting, he is hungry, or he just wants to get home as soon as possible.

Second type is a Geek, and his comfort is everything for him. People like geeks are ready to check all new technologies that are available and spend some time on it. They will get new experience and satisfaction as a result. Geeks can pick the jamon for hours, and time doesn't matter.

Third (and huge) group of people that are separated are Normal People. This group is the intersection of two previous groups, who care about both time and comfort. Basically that means that person cares about time, but 10 minutes difference will not bother him if add some value to this waiting. At the same time person cares about service during check-out, he doesn't want to have other people breathing in the neck and he doesn't want to be stressed when packing his groceries.

After creating personas we started to brainstorm the pain-killers and gain-creators for all of them.

When we got empathy maps for all three types of personas, we summarized all the ideas and evaluated every idea in two-dimensional plane — by complexity of improving and by importance for user. The easiest and most important ideas got 20 points, and the most complicated and not important — 0 points.

Here you can find all the solutions we made. We decided to present a couple of huge solutions like implementation of RFID system (recently Amazon presented some kind of alike system and some stores in Japan already have it, but we improved all existing solutions, united them and believe that it is more successful and simple, as well as its marketing attractiveness is indisputable), and improved self check-out system. Also we got some nice small solutions which can be easily implemented and are important to customers as well. For example, researches showed that people don't care about the time of waiting if they have what to do during waiting — that's why we proposed to install the charging stations for cell phones and WiFi points. The indicators of working check-out points and indicators of working payment terminals can significantly help to choose the cash desk, as well as two-way monitors with information about grocery list can decrease the degree of anxiety for the customers, who used to watch after the check-out.

Testing the complex solution with RFID-labels on the goods showed pretty unexpected results. We were really afraid that people will not understand what to do, and check-out without a need to take everything off the cart and scan every single piece can be confusing. Anyways, appropriate announces in the store closed all the issues and all the test persons did a great job. Also we found some bugs with the interface of the cash desk (for example, it is easier to put a cart to the right sight of you, not to the left). The fast blinking messages are confusing, because person things that he is missing something, and for that reason we had to hold the message "Please, wait for goods to be scanned" for the seconds, no matter that scanning is immediate process. Those bugs were fixed on the second iteration of testing and we got excellent results overall.