Hi carders! We recently redesigned the interface of our ATMs. It’s not that the old one worked poorly, it’s just that changes mean development, and development is always good. We tell you what was changed and how, how we came to this, what came of it and why even update something that already works.
Our ATMs had a pretty good interface. “Good enough” in this case means that it allowed the client to perform all the necessary actions - withdraw money, deposit money, transfer money, etc. But it was somewhat outdated. We also asked ourselves: is there anything we could improve on it? Something that our customers would love, but that we don't see because we're on “this” side of the ATM and don't look at it through the same lenses that users do? How simple, clear and fast is our interface?
We turned to colleagues from the Customer Experience Development Department for answers. Together with them, we planned three stages of research.
First, we decided to ask our customers what they like and don't like about our interface. At this stage, we have identified elements that definitely need updating.
Secondly, we compared our interface with competitors to identify how and why we differ. Based on the results of this stage, we identified a list of potential improvements that would allow us to maintain competitive parity or even surpass competitors in the convenience and speed of interaction with an ATM.
Thirdly, we conducted a “reality check”: we found out whether all the potential advantages that our competitors have and that we may have are really needed and liked by our customers.
We also noticed “extra” screens. For example, previously, when a user withdrew money, he had to confirm the amount twice. This was done to inform the user what the commission would be. But most clients withdraw money without commission, and the extra screen is more of an irritant than a source of useful information.
Another such screen turned out to be a seemingly standard window offering a check. We implemented this function as a separate stage that requires action from the user. The study showed that this also annoys many people - even if a person does not need a receipt, they still have to press the button.
In addition, it turned out that the “Bill Set” function needs to be seriously improved. It allows the client to collect an amount from bills of various denominations, but previously the process was organized so confusingly that even experienced users got lost in it. At the same time, the function was used often.
Finally, our ATM has eight buttons on the start screen. On the one hand, this gives quick access to a larger number of functions, on the other hand, it “blurs” the key ones among them - depositing and withdrawing. Namely, these are what clients are looking for first.
Reworking the start screen allowed us to draw it more clearly and noticeably.
The effect of this simple solution: paper consumption was reduced by three times only because customers stopped accepting “spontaneous” checks.
For example, we had the idea to simplify the process of depositing cash as much as possible, reducing user interaction and the number of screens to a minimum: put money in the ATM - done. Sounds great.
We studied the market and found out that the simplest scenario in most cases looks like this: the client puts money into the bill acceptor, it is credited to the account, and bills that do not pass the check are returned back. Everything seems simple, but there is a nuance: the client learns about the amount of the credited amount after completing the procedure, and in order to replace the “rejected” bills, the script must be completed again.
Research has shown that many people don't like this detail. People want to know how much money will be credited to their account before a soulless machine chews it up. In addition, they showed us a scenario where a client simply decides to get rid of “unnecessary” cash and wants to immediately see how much money he put into the bill acceptor.
As a result, we did not make an accelerated deposit. Our ATMs now immediately show how much money the client has put into them, and if some bills do not pass the check, the person can immediately replace them, before they are sent to the account. A good example of a “reality check” of promising functions that only at first glance look like they are.
According to the results of measurements, the cash withdrawal operation began to take on average 3.3 seconds less than in the old interface, and the cash deposit operation accelerated by about one and a half seconds.
In our opinion, these are significant changes, which is confirmed by our comparative studies of competitor ATMs. Now we are in the top two. And this is provided that we deliberately left some stages of the customer journey that other leaders do not have. Interacting with an ATM is always a cash transaction, which is primarily carried out in a public place. In other words, for the client, this is almost always a situation accompanied by a certain level of stress. The less time he spends in this state, the better his user experience.
We also collected more than 10 different defects in the software that appeared when processing various hardware or user errors. All of these mistakes ultimately led to cash being seized and a claim being raised.
For example, the client did not confirm the deposited amount because he had already left. But we know how much he contributed and why not credit this money to the account? Now it will be so. In general, failures occur extremely rarely and their number is measured in thousandths of a percent, but we have provided a function so that even in these rare cases no undesirable consequences occur for clients.
Of course, the main function of an ATM is to issue or accept money. For it to do this, it is not necessary to conduct scrupulous research and tests; you can simply make “some kind of” interface. However, it is also true that the ATM interface is a miniature bank. The level of his desire to continue using the bank’s services depends on how good the experience the client receives. Therefore, our ATMs will become more and more convenient.
How to understand when it's time to update the UI
In short, the ideal ATM interface should be reasonably simple, understandable and as fast as possible.Our ATMs had a pretty good interface. “Good enough” in this case means that it allowed the client to perform all the necessary actions - withdraw money, deposit money, transfer money, etc. But it was somewhat outdated. We also asked ourselves: is there anything we could improve on it? Something that our customers would love, but that we don't see because we're on “this” side of the ATM and don't look at it through the same lenses that users do? How simple, clear and fast is our interface?
We turned to colleagues from the Customer Experience Development Department for answers. Together with them, we planned three stages of research.
First, we decided to ask our customers what they like and don't like about our interface. At this stage, we have identified elements that definitely need updating.
Secondly, we compared our interface with competitors to identify how and why we differ. Based on the results of this stage, we identified a list of potential improvements that would allow us to maintain competitive parity or even surpass competitors in the convenience and speed of interaction with an ATM.
Thirdly, we conducted a “reality check”: we found out whether all the potential advantages that our competitors have and that we may have are really needed and liked by our customers.
What did the research reveal?
The results showed that we have room for improvement. To begin with, to withdraw cash from our ATMs you had to make three taps, while there were solutions on the market that allowed you to do the same in two or even one.We also noticed “extra” screens. For example, previously, when a user withdrew money, he had to confirm the amount twice. This was done to inform the user what the commission would be. But most clients withdraw money without commission, and the extra screen is more of an irritant than a source of useful information.
Another such screen turned out to be a seemingly standard window offering a check. We implemented this function as a separate stage that requires action from the user. The study showed that this also annoys many people - even if a person does not need a receipt, they still have to press the button.
In addition, it turned out that the “Bill Set” function needs to be seriously improved. It allows the client to collect an amount from bills of various denominations, but previously the process was organized so confusingly that even experienced users got lost in it. At the same time, the function was used often.
Finally, our ATM has eight buttons on the start screen. On the one hand, this gives quick access to a larger number of functions, on the other hand, it “blurs” the key ones among them - depositing and withdrawing. Namely, these are what clients are looking for first.
What have we changed
When it became clear that it was in our previous interface that it did not work well enough, the issue of updating it turned into a matter of technology. Here's what we changed.There are fewer buttons on the main screen
We removed some of the functions in the second-level menu, gave the free space to two “main” buttons: “Withdraw cash” and “Deposit cash”, and managed to reduce the number of touches to complete operations from three to two. In addition, during the study, we found out that not all clients quickly find the end session button. Properly ending a session, especially in a contactless transaction, is a very important feature. After all, if, due to forgetfulness, the client simply leaves the ATM without completing the session, the person who was next in line may continue to work in it. Therefore, we paid attention to the corresponding button.Reworking the start screen allowed us to draw it more clearly and noticeably.
An additional screen with commission information has become context-sensitive
Now we show a screen with the amount and size of the commission only to those clients whose commission is stipulated by the terms of the contract. Everyone else just gets money.Offering a check is no longer intrusive
The screen where you can order a check to be printed now appears automatically while the ATM is counting bills before issuing or crediting, rather than as a separate dialog. This procedure lasts at least 7 seconds, and according to our measurements, this is enough for the client to decide whether he needs a check. If yes, then the client will simply touch a special checkbox. If not, no action is required - the operation will complete automatically.The effect of this simple solution: paper consumption was reduced by three times only because customers stopped accepting “spontaneous” checks.
The “Bill Set” function has become more flexible
We abandoned the inconvenient system of fixed “steps” with the number of bills. Now you can select the desired amount from any number of any denominations that are available at the ATM. The process has become much simpler and more intuitive. This is our pride - almost no one else has this on the market.What we changed with caution
We approached the research stage with a set of hypotheses, which we formed based on market analysis and trends in the field of UI. However, not all of them were selected, and we introduced some functions with some conscious restrictions.For example, we had the idea to simplify the process of depositing cash as much as possible, reducing user interaction and the number of screens to a minimum: put money in the ATM - done. Sounds great.
We studied the market and found out that the simplest scenario in most cases looks like this: the client puts money into the bill acceptor, it is credited to the account, and bills that do not pass the check are returned back. Everything seems simple, but there is a nuance: the client learns about the amount of the credited amount after completing the procedure, and in order to replace the “rejected” bills, the script must be completed again.
Research has shown that many people don't like this detail. People want to know how much money will be credited to their account before a soulless machine chews it up. In addition, they showed us a scenario where a client simply decides to get rid of “unnecessary” cash and wants to immediately see how much money he put into the bill acceptor.
As a result, we did not make an accelerated deposit. Our ATMs now immediately show how much money the client has put into them, and if some bills do not pass the check, the person can immediately replace them, before they are sent to the account. A good example of a “reality check” of promising functions that only at first glance look like they are.
How the changes affected the user experience
Of course, you shouldn’t limit yourself to just introducing new functions. Even those innovations that have been “tested” by the market and customers before implementation can show unexpected results. Therefore, after the update, it is important to find out the audience’s reaction. At a minimum, this allows you to determine the effectiveness of innovations and identify problems that need to be eliminated. As a maximum, understand what other functions can be added or reworked qualitatively.According to the results of measurements, the cash withdrawal operation began to take on average 3.3 seconds less than in the old interface, and the cash deposit operation accelerated by about one and a half seconds.
In our opinion, these are significant changes, which is confirmed by our comparative studies of competitor ATMs. Now we are in the top two. And this is provided that we deliberately left some stages of the customer journey that other leaders do not have. Interacting with an ATM is always a cash transaction, which is primarily carried out in a public place. In other words, for the client, this is almost always a situation accompanied by a certain level of stress. The less time he spends in this state, the better his user experience.
What will we do next
After the stage of eliminating minor shortcomings, we will definitely enter a “new circle” of research with new functions. We are currently working on introducing context-sensitive confirmation of transactions with contactless cards. There are quite dangerous scenarios in which customers begin an ATM session by swiping their card, but do not complete it, but simply leave, leaving the session open for the next user. To prevent accidental use of such situations, we have already clearly highlighted the session end button and are ready to replicate the additional confirmation option. For example, if more than a certain amount of time passes between customer actions, the ATM will ask him to swipe his card again to make sure that the cardholder is still in front of him.We also collected more than 10 different defects in the software that appeared when processing various hardware or user errors. All of these mistakes ultimately led to cash being seized and a claim being raised.
For example, the client did not confirm the deposited amount because he had already left. But we know how much he contributed and why not credit this money to the account? Now it will be so. In general, failures occur extremely rarely and their number is measured in thousandths of a percent, but we have provided a function so that even in these rare cases no undesirable consequences occur for clients.
Of course, the main function of an ATM is to issue or accept money. For it to do this, it is not necessary to conduct scrupulous research and tests; you can simply make “some kind of” interface. However, it is also true that the ATM interface is a miniature bank. The level of his desire to continue using the bank’s services depends on how good the experience the client receives. Therefore, our ATMs will become more and more convenient.
