LONG READ. What is the "digital twin" of the network and why do we need geoservices for working with offices? Complex things in simple words… In early 2019, Alfa-Bank launched a large-scale project to transform its network of branches in Russia. The new network should take into account the changed needs of customers, new formats of banking products and services, and, of course, be convenient and comfortable for everyone. We would like to tell you a little about how our team at Marketing Logic, an analytical company, is involved in the transformation of one of the country's leading banks. This may be interesting, because the project has implemented many innovative solutions that meet modern challenges and trends.
„Bank branches are something that hasn't changed much in the last 10 years. And it is not only about the interior design and the materials used, the concept of service in the department remained unchanged. During this time, digital services have come a long way, but people have not stopped going to the department — even a very good digital service can not always replace a confidential conversation in a cozy atmosphere. So, we need to modernize the branches so that they provide a modern customer experience that is not inferior to the experience in digital channels. Whoever gets there first-wins this market.” Elena Tyatenkova, Head of the Network Development and Transformation Directorate
A look into the past and future Modern financial services are complex, multi-component solutions, and offices as an offline showcase and platform for these services must meet a variety of requirements at once: from security and convenience to a good location where it will be convenient for customers to come. The amount of facts that need to be taken into account when implementing the concept of "optimal network" has long been beyond the scope of expert assessment.
Our company specializes in solutions in the field of advanced analytics and modern technological solutions. Therefore, we were able to offer one of our solutions in the field of geomarketing and big data analysis for the project. Using these tools, the bank was able to create scenarios for the transformation of the current network into the optimal one, varying a wide range of variables from the metrics of the load on the branches to the balance between investments in the branches and the economic return on them.
One of the most important factors in such a system is taking into account the geo-parameters of the network: in which city, district, with what number of population, traffic, transport, environment, plus more than a dozen variables, including internal banking. If you multiply these questions by 400, i.e. the number of offices, and take into account the size of the country, you will understand the scale of the task.
Usually, when talking about geomarketing, people keep a certain map in their head, where the best locations for opening are highlighted. Paradoxically, but in this project, the current geodata went to the second or even third plan. The bank really wanted to take into account the experience of the past. For example, previous activities on your own network. At the level of each department, see all the stages of development, and how this development was influenced by macro - and micro-indicators in dynamics. Take a quick look at the networks of competitors in the past and, perhaps, learn from their mistakes.
For such a task, you need a large data set describing the locations around the branches and the general situation in the city/country in the past. We needed the geodata from 2015. It turns out that the past is necessary in order to look into the future.
Why not rely only on the current situation and why is the long-term forecast so important? The transformation project of a large network can take several years in the face of dynamically changing external factors. They can be both predictable and non-predictable "black swans", which is confirmed, for example, by the situation with the COVID-19 pandemic in 2020. In order to minimize the risks, the system included a forecast of the attractiveness of locations for the next 5 years, highlighting the most stable zones in terms of quality. We also added the ability to dynamically change the targeted development program based on incoming unscheduled factors. This approach raised the overall stability of the development program and prioritized attention on areas with increased development potential throughout the entire duration of the transformation plan.
In total, the formation of the optimal scenario for "network restructuring" required the analysis of a ten-year period: both prospective and retrospective data.
The success of a network is a combination of many factors: office characteristics, employees, business processes, products, locations - and the transformation must take them all into account. When we were conducting our part of the project, our colleagues from the parallel stream showed us a full-size "cardboard" prototype of the department, on which they optimized client flows and business processes. A very clear and logical approach is to conduct testing and optimization in advance, with minimal costs, and get the product as ready for implementation as possible. For our part, we used a similar approach in the project, but without paper-creating a digital twin of branches and a network. "Digital twin" is a modern high-tech trend of working with data. It creates a virtual environment in which the simulated objects are presented with maximum confidence. This implementation allows you to perform an unlimited amount of scenario calculations, comparing them with each other and working out any hypotheses of experts. The common denominator in these calculations is the same digital data format for all branches and the digital model of the network with all its characteristics, parameters and financial indicators.
As a technological platform for creating such a twin, we used our boxed solution-the Geonet service based on the Atlas geoinformation system of our own development. This service allowed us to quickly aggregate all the data that the joint team of the bank and Marketing Logic wanted to combine. Together, we built a "virtual" network of Alfa-Bank and looked at how each change will affect the business, its prospects and financial forecasts.
In other words, the system allows you to create several development options (or different future scenarios) and choose the best one based on a detailed comparison of key performance indicators. The system created development scenarios automatically, marking the best ones and passing them on to the experts.
In the bank's team, the project was led by highly professional employees. And it would be wrong to give them "raw" solutions that require a lot of time to adjust. Therefore, the criterion of "realism" of the transformation scenario is very important. For example, if we built an ideal network with ideal office addresses, but there would be no or almost no premises for rent. Or simply there would be a small probability that the room will appear in the window allocated in the business process for the selection of the room in the location. The solution to this problem also occurred with the help of the concept of a "digital twin". We model the network not in a "vacuum", but in a real urban environment. The system analyzes retrodata on the real estate market and determines the probability of the appearance of a suitable room in the location in the planned opening period. This approach allows you to realistically assess the time required to search for premises in different areas, to form a targeted development plan with minimal changes and to go as accurately as possible through all stages of the project.
«The forecast is built from the bottom up, from the smallest components to the overall economy of the network and the profitability of the bank. First, all calculations are made, then they are converted into money, which makes it possible to evaluate each component and stage from a financial point of view: both investments and returns. Immediately entering into the system the requirements and checklists for the premises, we get a realistic address plan for the development of the network. All the justifications for each action are there initially, and the system gives realistic suggestions. Taking into account the scale of the financial organization allows you to save several months of working time of the bank's specialists» Dmitry Galkin, Managing Partner of Marketing Logic
A modular approach to the formation of an optimal branch in an optimal network Often, the solutions that geomarketing provides are limited to the answer "good location/bad location". We used a modular approach to determining the format of offices, taking into account the needs of each business area and the requirements of the bank. It makes no sense to make a cash register in a branch where no one will use it, or a service area for legal persons in an area where there are few such customers. With the help of geoservices and data analysis, we identified the volume and quantitative needs of each potential location for the office and started from them. Such a system should include not only the options "open office/close office", but also many others: reformat, add functions or remove some of them, increase the staff, reduce it or change its profile or motivation. A separate and important point is the long-term planning of the network, taking into account the latest digital channels of interaction with customers. Such channels change business processes and network capacity and format requirements.
The Geonet system perceives the office as a set of modules combined into a single financial model. Hence, the recommendations are not just for branches, but for individual modules, employees, areas and other parameters, with the details that the bank wants to see. The transformation of the network takes into account the digital ecosystem, ATMs and new office formats as ways of support and service. Simply put, we go from a product and a set of functionality to filling the office within the framework of the phygital concept adopted by the bank.
„Marketing Logic has been developing geoservices and big data analytics for 5 years, but even taking into account our experience and complex portfolio cases, the tasks set by Alfa-Bank are, frankly, non-trivial experience. Phygital's latest concept, extensive network of offices and the bank's desire to stay ahead of the market is a definite challenge for any team. This strategy determines the scale and complexity of the challenges facing us, the project, and the collaboration. Taking into account the technological trends, the development of geo-and AI-solutions, we can confidently say that this approach is the future, and not only in the financial sphere, and we are pleased to be at its origins in Russia” Dmitry Galkin, Managing Partner of Marketing Logic
AI and Big Data in network management – is it the new standard? We have tried to describe as simply as possible the complex process of transformation of the branch network with the involvement of new tools, solutions and approaches. We have already mentioned that there are so many variables and data to analyze that no one in the world is able to process them all together and assess their impact on each other. To improve the accuracy of forecasts, we upload the maximum possible amount of data – these are petabytes of information. The system developed for the bank has 18 predictive models, each in its own area of responsibility. In forming the optimal development footprint, we helped the bank calculate more than 21 million alternative transformation scenarios. In this project, our many years of experience working with artificial intelligence systems and elements and the experience of embedding our solutions in real business processes were very useful to us. Even an ideal solution that is not adapted to the client's long-standing processes creates a conflict between the machine logic and the expert. And the decision made taking into account the careful study of previous experience, the existing business processes of the company (and respect for them, which is important), on the contrary, creates an environment for their joint work. This gives an additional synergistic effect and increases accuracy. We are aware that in 5 years such projects will be easier to do, moreover, it will become a new "traditional way", but if you want to be the first, it is better not to wait.