Future Digital Finance 2024

April 02 - 03, 2024

Hyatt Regency New Orleans, LA

Digital Intelligence

     

In this video, Paolo Giambertone of ING Direct (Italy) covers how empowering the digital consumer can help financial institutions maintain scale and cost advantage by maximizing self service.

img img

Video Transcript

My name is Paolo Giambertone. I am working within the Digital Business in ING Direct since 2004. So a relatively long time if we think about all the transformation that has been impacting the digital environment. So maybe one important point is that working in Direct bank that is native digital bank in place that we are in some way going into a reverse path compared to the usual traditional retail banks. Traditional retail banks are going digital. In some way we are partially and in some selective way going a bit more physical when needed.

In fact just a quick overview about the company in order to explain some points of our business model but also to explain the exemptions of the topics we will be touching later. ING Direct is the direct business belonging to ING Group which is a much bigger retail and commercial banking business. The first pilot of ING Direct was launched in Canada during 1997. It was really a phone only bank even without a website which was launched the year after. So in fact it was a bank built around a call centre. Then after the first pilot was successful, other businesses were launched Spain, Italy, France, Australia, UK, Germany and so on. In our current situation what we know is that we are a fully digital bank in the terms that 95% of our customers as a gross average are really interacting via digital but if we look at different segments of our customer base for example the payment account customers, 98% of them are interactive via web or mobile. So this gives us a lot of insight about the correlation between the interaction on digital and in the same time the interaction of other channels which is our primary aim.

Why did we launch for example a very limited group of branches? It was mainly to leverage to multichannel to get some commercial advantages and in the meantime to continue leveraging on self service and digital in order to build scale and cost advantages. And in fact on the commercial side, this really worked fine because after we launched the branch network, we increased the performance, the commercial performance of digital and we also got some interesting performance on the subset of branches. We had the first one was launched exactly at 10 years and 6 months after the startup.

So the topic of today is mainly how maintaining scale and cost advantage while maximizing self service because there are some maybe some hidden point that are worth sharing with you. I said that okay, during 10 years a lot of transformation has been occurred on our side because we improved our product range. On the very beginning we were a mono-product bank with only a savings account working on two channels, web and call centre. So click and call and stop. Then we increased our product range adding mortgages, payments account and we improved our distribution model and in the meantime mobile was exploding so. The intensity of that come from the contacts of our customer is really growing in some important way. And that's mainly driven from the payment account growth and from the mobile growth and again this is the gross average of 40 contacts per year but nowadays, recent customers really have a daily interaction with us.

So apparently 95% of digital contacts, a fast growth, so really direct banking model that apparently is going in some way. This is apparently the good story, yet there is something to watch for because as I told there are some kinds of hidden pit falls. So I will explain this graph. On the upper line, you see the total number of logins on digital, okay. It's a normalized graph and so in the other market line you see the inbound contact centre calls that are coming from the same people who have logged in one or two hours before not more. So apparently there is a quite strict correlation and if you don't break this correlation that is the more customers the more contacts, the more contacts the more inbound contact centre calls. We are not fulfilling the scale we can get from our direct banking model. Okay.

So the important point is breaking this linear correlation otherwise we have some scenarios we don't like. We have to decrease the service level of the call centre, we have to increase in a dramatic way the resources needed to run the call centre and the customer service which of course we partially can do both but we should manage because otherwise we don't feel scale as I was saying. So a very important KPI for us is of course the channel deflection. I would rather call it the key under performance indicator in some situation. Some other situation, it's fine to drive in a selective way the traffic towards call centre or the branches. It depends. But there are of course a lot of situations you don't want this kind of unwanted calls.

So there are three main things we are doing right now. Of course the first thing is being in control which means having the right, the proper digital intelligence in place, but not really in the traditional way. I will try to explain this. The second point is empowering the digital customer so that it doesn't need some assisted contact. The third thing is to improve the potential of the contact centre that is for example moving into a situation of a more evolved contact centre with more capabilities and more tools and more synergy.

A point about having the proper intelligence in place, 90% agreed to a lot of things that I heard this morning from another banking colleague. In our situation, we are managing this kind of try to digitally centric way. Why? Because of course 90% of contacts are happening on digital. So it was much more comfortable for us to push the digital analytics capabilities to the maximum which is not–which is nothing like implementing Google analytics but having in place a very customized solution with a lot of custom variables in place and having a very deeper feed from the other channel in terms of information. But what we have to study is really the digital behavior on digital channels about having in place some dimensions and relations with what is really happening on the other channels.

So of course just to explain what was our roadmap on this. Digital analytics have been really great importance for us for years to support sales. So to do the usual things that are understanding the funnel, understanding the funnel friction point, checking the life of digital companies and so on. Now the situation we have moved in is understanding the digital friction point but in terms of multichannel behavior. That's why for us is crucially important to understand the specific customer segments and being able to analyze very detailed clusters in terms of digital behavior. So getting rid of the aggregated information becomes really not meaningful because the mortgage customer is behaving in a totally different way from a payment account customer and so on. And also the role of different channels.

So in some way having a unique analytical layer is a solution for us to get rid of the silos because as it was mentioned this morning also even the different channels, mobile, the M site, the app, the website are often build as a standalone silos. So we are putting some connection between each silo using a unique layer in terms of analytics. So this was the most important point in order to manage information and be in control.

The second point is empowering the customer. So it sounds quite obvious to implement the sites or changing is nothing new as something that has the same age of the internet. It's apparently obvious because for example a few days ago, we saw a study from Forrester Research that is telling us that only 25% of US Banks are implementing an in-site search engine within the internet banking where most that's are happening. But also a few of this 25% is properly executed. What do we mean for proper execution? It's all about using the method or language scheme. So not working with full text or keywords where this game is quite clear. You have a set of pages or document. You put some keyword and you will show a set of results that these mainly are irrelevant for the customers but it is all about understanding the context.

I have put a couple of examples from other industries. This is a major consumer electronics retail site selling a lot of smartphones. If you put a search query like high end smartphone, you will receive zero results. Maybe they are still selling a lot and a way but this is a typical result of the keyword based approach. On the other end if you try to serve this commercial liners website, you see that there is a tool that is able to deal with language ambiguity. Understanding the context and maybe if needed giving some options to the customers. As we didn't have yet the study from Forrester we were performing an internal survey and we were just benchmarking other banking sites. And the result was that okay. When in semantic search tool was used, 70% of the results were relevant. When the full text approach was used, around 20% of results were displayed. So the take out of this is if you perform the low cost solution, maybe its better not to have it because the customer won't be finding what it needs. And it will be in some way disturbed of the situation and of course it will act on the cost driver for example we want to avoid that is the inbound for call.

What about our own result? We are quite on the beginning on this kind of experience but they sound quite encouraging because okay, 23% of our customers were using this tool during the first six months. We noticed on the segment interesting uplift in terms of first buying plus 12% in sales and we also had the visible result on cost efficiency because we were generating minus 6% total inbound calls which it becomes quite relevant when you have maybe million of customers.

Last topic is about quite well known and advertised the tool that is the web chapter. But in fact here execution matters because you know that you can have two different approaches on this. The proactive chat of course is displayed the proposing the web chat when you notice some kind of behavior that people, that show that people are in some difficulty. They are not able to perform the action they would like to perform and in this case you can decide to proactively show this kind of tool. On the other end you can have a persistent presentation of the chat option on every page of the website or in some specific page of the website.

Personally yeah I'm sponsoring the 100% proactive approach also because okay, it's true that a skilled contact centre agent is able to handle three or four chat sessions at once. That's max, but the minutes he has to spend with the customers are limited reserves and every chat session is lower than a call. So it's really not granted that the voice channel is more expensive than the chat channel. It depends always from the context. In fact we have been developing much more the proactive chat and on the selected set of contacts, we really find that this is granting us some interesting result because the conversation rate of this kind of visits, people engage with the chat compared to the total is exactly double. So of course it's all about being able to constantly develop this kind of proactive intelligence in order to work on the rules, refine them, measure the result with the proper analytics in place and then going and refine. So it's something we can do while applying the traditional digital marketing optimization approach but putting this in the more broader multi-channel context involving of course the resources of assisted the channels, contact centre, and so on.