MKB is the first Hungarian bank, and among the first across the globe, that sends personalised messages to its customers based on their spending behaviour and life events. And they’re about to take it up a notch. The bank will soon be offering real-time, personalised banking products, too. They do so with the help of W.UP’s sales and engagement tool, Sales.UP that takes different types of customer data, think transactions or geolocation, and translates them into profiles and life events, aka insights.
We’ve sat down with Attila Kezdődy, MKB’s Head of Data and CRM, to talk about the implementation challenges and the benefits they expect to achieve with the rollout. A mathematician-cum-economist, he started out as an analyst and programmer, then moved on to client relations and client management. In the last 10 years, he has worked as head of CRM at three different banks in Hungary and joined MKB Bank in the summer of 2016.
Why has MKB Bank decided to implement a real-time sales and engagement tool? What were the challenges you wanted to overcome?
When I first started at MKB Bank, vice president Márk Hetényi and I made it our mission to take customer experience to a new level. Márk had several ideas about which direction we should take, and I had a working knowledge of what can and cannot be done. Are you able to respond to changes in customers’ lives within a day? Great! But you can do better.
We looked at new technologies and the rapid evolution of artificial intelligence and said, ‘Hey, that’s doable’. No one else has come this far in the Hungarian financial services sector. Our cooperation with W.UP entered into the conceptualisation stage at the end of 2016 and we signed a contract in the spring of 2017. Due to the strategic partnership and joint development, it took us a year to get into the implementation phase. But the finished product is ready to be deployed within six months.
Digital sales made easy
What vendors and technologies did you consider during the planning process?
Our first idea was to do the development in-house. But we knew that it would be very challenging. We could sense that we are about to embark on something really exciting, just looking at the number of companies who wanted to work with us as vendors. But all of them told us straight that it would be a learning process for them, too. No one had a ready-to-use solution. They brought several ideas on how to collaborate with us. The bigger ones basically offered to give us Lego bricks and it would have been our responsibility to make them work. I’ve been there, done that too many times. I knew that it would have been very complicated. Not to mention that it would have required virtually endless resources from our side.
This is what drew us to W.UP’s solution. They offered a dedicated team for software development and a separate team of data scientists who bring new ideas on how to tweak the system, which is far ahead of everyone else in the industry. W.UP had started developing its product long before we started negotiations but they also needed a partner for rolling out a full-blown project.
With any new development project, there’s usually an advisory firm whose task is to figure out what to do and how to do it in the form of a functional specification. And then there’s a software development company that delivers the technology. In W.UP’s case, these two teams work together within one company. That’s a huge advantage because there are no disruptions or blockages in the communication flow. The other thing is, and this is something that we started appreciating mid-project, W.UP develops our mobile application, too. This means that whenever we talk about how things can be incorporated in the mobile app, we don’t need to involve a third party to make it happen. Everything is in one place. Plus there’s a geolocation module in Sales.UP to be built into the mobile app that we’ve also rolled out and started using. This is a win-win for everyone: it’s useful to us but also to W.UP because they learn a lot about how the system works in practice.
There’s an engine behind Sales.UP but other components work independently, irrespective of what technology we use. And there are elements, such as campaign management, that we’ve added ourselves. Sales.UP has a built-in campaign management module but MKB Bank also has its own and they can’t run in parallel. In the end, we decided to deploy Sales.UP in a way that it draws data from a central offer database, and the same database is used for running the bank’s own campaigns. Whenever we need to check if a campaign is good to go, both systems ping the central database to make sure that there are no discrepancies in real time.
Alfréd, your digital smart assistant
What areas within your organisation need to work hand in hand on the implementation project?
Content creation has been a joint responsibility for our sales, product, digital and CRM units. The roll-out was more IT development and maintenance. And of course, our marketing team was in charge of coming up with engaging, easy-to-grasp offers.
With digital solutions, it’s always a question of what people will use them for. You can’t just do a high-level technological implementation. If you want to sell something and want people to love it, you must get the implementation right. This is why we’ve come up with Alfréd, a digital smart assistant talking to clients instead of an app.
What technology-related challenges have you faced during the implementation?
We’ve run into a couple of IT issues during the process. We needed a test system for receiving real-time data. Our own IT experts were surprised to see where the data came from and how. How exactly do you get card transaction data, for instance? We first thought that all we needed to do was copy the data of transactions between systems and take it from there. But eventually, we realised that we needed entirely new fields, both from the data warehouse and the real-time systems.
My idea, the same W.UP has come up with, is working. Customers need to be profiled and their data should be updated continuously. But it’s enough to update the profiles on a daily or weekly basis. No need to redraw profiles every single time a new transaction has been made.
The implementation is still challenging, though, because the software must be easy to containerise and scale, it must be fast and must process queries in-memory. This required a whole new set of test systems, not just building one on top of the other. It doesn’t work like that.
We’ve realised that no one else in Hungary has done a similar implementation. Campaign management coordination through the offer database was new to us, and it was something that we had to deal with mid-process. But that’s all behind us now. The last step was to fine-tune the geolocation feature.
The initial population was a tough one. It’s part of the profiling process, which basically means observing daily changes. For that, you need a stage zero, when you upload every piece of information you already have about every single customer into Sales.UP. This took several hours, even optimised. To make sure that everything was running smoothly, we had to reset everything three or four times, including the operating system and communication software, and successfully complete the initial population and launch. Debugging was next, then we could start the whole process all over again. For eight weeks following the soft launch, only employees could use the new interface. Once everything seemed stable, customers could download and start using the app, too.
Messages that matter
What life stages require personalised messages or offers?
I’d never thought it would be so complicated to fill the software with actual content. Developers and maintenance experts help with technical, testing and operational matters. But you really have to think through when to contact customers so it doesn’t feel pushy or irrelevant, how often and with what message. We’d put a lot of effort into this but altogether we could only gather seven service- and sales-related insights to launch with that drew on transactional and geolocation data.
Let me give you an example of the life situations we started out with. You walk up to an ATM and insert your card to make a cash withdrawal. But the ATM doesn’t give you the money. When this happens, the ATM screen usually says that there was a problem and the transaction could not be completed. Now Alfréd sends you a message, saying that you’ve reached your daily limit for cash withdrawals, and if you need more money, you should just raise your limit from HUF 100,000 to HUF 200,000, and that’s it. Problem solved.
Alfréd, the digital smart assistant of MKB Bank
We have a long backlog and this is definitely something we’ll need to address. Channel usage is one of them, and it’s actually the next component that we’d like to incorporate. It basically monitors how customers use the mobile app and the online banking platform. If a customer can’t complete a savings account transaction, for example, we’ll be able to help them by sending messages or logging in.
How do you make sure that the cooperation runs smoothly?
To make any implementation project a success, you need a core team that takes charge of it and keeps the whole thing moving. Now we’re doing this as part of a project but no matter how passionate we are, we will run out of ideas one day. This is where our cooperation will need to proceed into the next stage. What I like about W.UP is that the company is building an insight hub. They offer newer and newer insights and this is how you can ensure continuous development. You don’t just get technical support. You get more and more ideas.
This is crucial because it’s quite hard to come up with new things all the time. Just think about what you’ve done today: how you took your regular route, got in a cab or stopped for coffee. What can your bank offer you that’s different from the usual offers? Bringing in new ideas is extremely challenging and I can’t guarantee that there will be a new message every day. Based on our estimation, maybe 1-2% of customers will receive these offers. But that’s OK.
Next up is figuring out how to take our offers to a new level. In this respect, we’ll probably think beyond banking products and consider other products and services that we can sell through digital channels. This will go way beyond banking but that’s how we’ll get to the stage where we can cater to all our clients’ needs, using our capabilities, financial savvy and other resources.
As far as I know, no one else in the Hungarian banking scene is planning to roll out something like this. Usually, there are 1-2 players in each country who have the courage to take this step, even though digital transformation in banking has been going on for ten years now.
Return on digital sales
When will you start offering digital products?
When we started the project a year ago, our bank didn’t have any products to sell through digital channels. Even though we’d been running event-driven campaigns for years. By event-driven campaign, I mean that we’d analysed customers’ daily account activity and approached them the next day with a corresponding offer. These campaigns are much more personalised and accurate than regular ones that contact customers over something that had happened months ago.
The problem was that we couldn’t move the sales process forward digitally, only through call centres and in branches. Not to mention that next-day response times might have been seen quick a few years ago but today it just doesn’t do the trick anymore. You don’t necessarily expect your bank to respond within a matter of hours or one day, but if they do, that usually has the wow factor.
The lack of digital offers was so glaring that our bank has actually launched a new project to create fully digital loan products. This is critical because otherwise, this whole system would be nothing but a shiny new messaging tool that wouldn’t help sales one bit because of low-quality leads. Soon we’ll be launching our digital private loan so finally everything will fall into place.
What benefits do you expect from the implementation?
It’s not very easy to build a business case for service-related messages because customer satisfaction might increase but making more customers use the app – how do you even measure that? Our key performance indicators include the number of mobile app downloads, the number of messages read and the number of clicks. And of course, higher customer satisfaction.
Currently, there are 40,000 customers who actively use the mobile app, and several thousand other customers have downloaded it but have yet to use it. This is still only a small percentage of netbank users, who account for less than half of our customer base. To reach all these people, we’ll also need our 500-strong branch network because they meet the people who bank through other channels.
We need to raise awareness and get them excited about the tool so we can continue development. One of the main barriers was that less than 50% of our employees had used the mobile app. So we’ve launched internal campaigns to promote the application and educate our colleagues about the latest version. If you want to become a digital bank, you must sell your application to your employees first so they will use it and promote it to customers.
Until you have products to go with the system, you’ll only see indirect benefits. What we have to do now is raise awareness and in the next quarter, then we can continue with sales. That’s when we can start talking about a business case. Fortunately, if customer experience and customer satisfaction are among a bank’s core objectives, soft ROI and hard ROI are equally important. Our customers will be the ultimate judge.