Optimization of placement of bills in ATMs

Inside the average ATM is from 1 to 3 million rubles . They are arranged in 4 cassettes, each of which contains notes of its own value. Replacing cassettes is similar to replacing cartridges in a printer: the module is removed (and no matter how much money is left - collectors do not even know it according to safety rules), and in its place another module is inserted exactly along the arrow indicating which side and how to do it.

Collection is an expensive and quite risky in terms of security procedure, so any ATM service is quite expensive. The natural desire of the bank is to reduce the amount of collection. It turns out the classic problem of resource allocation: on the one hand, there should always be money for customers at an ATM, on the other hand, ideally, the bank wants the last bill to leave the box exactly before collection.

Banks optimize all cash handling processes: this is both the mentioned modularity of the cassettes (so as not to take out money and not be considered hands), and a special arrangement of bags in the car, exactly corresponding to their order of use along the route, and much more. At the same time, everyone understands that if an ATM does not give money, the client will be upset: this is not a beer stall, where you can choose another variety or just cross the road and get the goods from another place. Not receiving money from the “native” ATM, the client is very upset, indignant and begins to lose confidence in the bank.
')
Accordingly, the main priority is to always have money. At the same time, I would like them to be, if not abutting them, then with a minimum margin - and it is here that we need very good IT technologies, which can cost very good money.

Tasks

In principle, the analyst has two tasks:

1. It is optimal to form a purchase breakdown: which bills, what dignity and how many pieces should be in an ATM.
2. Know the schedule of collection and calculate, depending on the specific situation, such a stock of funds so that it will last until the next time without unplanned departures.

Let's start with a breakdown by bill . As I have already said, there are usually 4 tapes: it’s clear that the two of them immediately go under the smaller ones: under the hundred and five hundred. But there are nuances: for example, 500-ruble notes can not be put at all, but put more than others. Or use two boxes at once for the same denomination. Or do not put five thousand and so on. It all depends on the consumption scenario and the issuance algorithm: if it is a student canteen, then you may need to collect it with full 3 cassettes with finished weave.

The second point is actually how much balance should be. The scenario of using an ATM is important here: for example, there are only payroll ATMs that stand in offices, there are ATMs that stand in shopping centers, and there are ATMs that stand in squares at train stations. It is obvious that payroll will have pronounced "critical days", those in shopping centers have seasonal peaks and lows, corresponding to common peaks and lows of retail, in squares and railway stations much depends on fluctuations in flow density and its character.

For example, there is an ATM in the business center. In those days, when there is a salary in the offices - it should be hammered to the eyeballs with large denominations so that those who rent them can withdraw as much as they need. Then, for example, it is filled to one third with bills of lesser value, and by some holiday or weekend it is filled with more. Based on the analysis of transactions, you can understand the peaks of withdrawals, understand the client profile of this ATM, find out what breakdown should be done.

There are a lot of nuances, but in general, the task largely boils down to solving the problem of providing goods on the shelf for stores . With all this, there are a lot of security features, for example, a ban on systematic visits to an ATM (using the same routes).

To optimize, we need to know

About ATM:

• Number of cassettes
• Algorithm of issue (is it possible to choose large / small?)
• Does he accept money or just give away
• What additional services it provides
• How and when was the money withdrawn (history of withdrawals and balances for Data Mining)

About the fact that around the ATM:

• Where is he standing (in the business center, in the shopping center, in the bank branch)
• The mode of operation of the place where he stands

About the conditions of the problem:

• How can I combine ATMs in groups
• Restrictions on the breakdown of bills (bills of 50 are ready to put? And 100?)
• Multicurrency ATM or not?
• The cost of finding money in an ATM
• Collection fees
• Purpose (i.e. what level of service is needed)

Decision

The first part of the solution is architectural . We need not count on how to carry money, but for a start - design and build an ATM network so that it works in an optimal way. This is done based on heaps of data. Our analyst can plan the network of ATMs in such a way that everything optimally stands taking into account a variety of restrictions - the presence of a number of competitors, street traffic, shopping centers, railway stations, seasonal changes, and so on.

Then we have to decide how much money and in which purchasing breakdown should lie in an ATM. Plus - how often it should be collected.

More ATMs can be replenished with customers, and the money is collected there in a separate cassette, which can also overflow. Money from this cassette is not issued: input-output at an ATM in different processes. From the point of view of IT, it seems logical to use this money for issuance, but according to the legislation, all this must first be carried out through the accounting department, and for this, the tape must be brought to the bank branch. Therefore, it is beneficial to make a cash on immediately and to replace the cassette with money from replenishments (since they have arrived anyway).

Next you need to ensure the correct filling of ATMs . To do this, we take a system that allows us to optimize the route from the point of view of minimizing costs and at the same time put restrictions like “routes should change once a week” (except for the security services, no one knows which route the collection vehicle will take tomorrow).

Returning to logistics, we have several blocks: first, we need to strategically plan it, second, we need to understand who to collect with what amounts and so on, and third, we need to optimize the collection schedule, both for individual ATMs and ATM chains right away. If there are a lot of them, then for each of us we consider how to optimally collect it, and then group them by routes.

For example, if one ATM needs to be collected once every 2 days, and the other once every 3 days, it will be two exits. If you collect both at once, it will be a single exit, which is more profitable and safer. At this level, the solution is in those ATMs that need to be collected less often, put less money in order not to create excessive idle stocks.

It is clear that all these blocks can be done on anything. For example, logistics is Oracle Transportation Management, forecasts for IBM SPSS to make, network placement on something else. Almost every major vendor has his own solution.

Reality

The main obstacle to good optimization is the security requirements. To put it simply, optimization in itself implies a clear system of visiting points, and protection requirements say that such a system will not be long. Further, there are options:

• The security service itself is working on routes . We can simply make recommendations and make monitors how much ATM in which rolls money: the connection between the ATM software and the central office is a common thing. But we are just solving another problem - we don’t reactively want to act: “Aah, we have a red light, the rest is zero, panic!”, We want to either help this red light to light up today and notify us that there is enough money only for tomorrow - either we must make a competent collection plan so that the lighting of this red light bulb is something urgent for us. Roughly speaking, it is much more economical to spend forces on solving the problem at the architectural level than to try to correct the situation on the operational level each time.
• The security service is ready for innovation . In this case, an environment usually unfolds where it is possible to build routes with regard to the optimal traffic pattern.
• The bank is ready for strategic planning . Here we are talking about the miscalculation of the optimal network location, setting optimal routes and alerts on the stock of money.

Little future

Considering the level of work of the predictive software using the example of this reading of the thoughts of the clients of the call-center , you can do the following: when authorizing a card, you can create individual work scenarios and suggestions. For example - “replenish my phone”, “give money for lunch”, “me as usual”, all sorts of “I feel lucky”, in conjunction with the purpose of payments - reminders about the Internet payment (with an instant opportunity to resolve the issue), payment of fines (even dialed 20 minutes ago) and so on.

The next obvious step is to understand which customers to offer. You can respond to actions with an ATM, analyze patterns of behavior. For example, launch personalized marketing companies, make “tasty” offers to customers who can leave, and so on.

There are practically no such complex projects in Russia. At best, individual blocks were worked out. Banks practically do not do this: not everyone has a thought that it is possible to live even better and at the expense of what.

Saving

From the standard scheme “when there are 200 thousand in an ATM, you have to panic and go” to a smart ATM without taking into account space solutions the difference is from 40 to 80 thousand rubles of savings per year to an ATM. Given the number of ATMs of one bank, this difference fully covers the cost of solving the problem and implementing the system.