Why are CFOs so eager to translate IT capital expenditures into operating
Modern business is built on the architecture of microservices. Very much simplifying, there are still some areas that are still procedural and written in good old C. It works - do not touch. And there are modern commercial structures that increasingly go to distributed architectures. Approximately for the same reasons as in the development, it is much faster to make large projects.
The consequence is slightly different vectors for CFO and CIO. This is expressed in the fact that the CIO often wants to bring everything “home” and put up a rack in the office, and the financier does not see any sense in this . And it just so happened that the CIO indirectly submits precisely to the CFO, so everything ends with a move to the cloud.
As written by kaichouin the commentary: “and then, a couple of years later, it turns out that as much money is spent on a cloud a year as the entire own server park cost”. And this situation is quite normal for the financial director. Now I will try to explain how he thinks in such a situation.
Financial director manages the flow of money in the company. His task is to understand where the money is most needed, get it somewhere and serve it there. He in this story is similar to the router with the functions of the balancer.
One of the features of his work is that he must quantize the expenditure of funds by months and years (most often there may be other periods). That is, for example, he wants to know how much it costs to own his own server node. For this, he takes:
- The cost of iron;
- The cost of support-repair-maintenance;
- The cost of money, ie, interest on the loan;
- Payroll admins (salary, contributions to the funds, medical insurance, etc.);
- Support costs (cleaning, office space for admins, their computers, etc.);
- The cost of electricity, etc., all sorts of loading and delivery.
He tries to calculate cycles (usually the park is updated once every 4 years by servers and once every 5-6 years by network devices) and calculates the maintenance of all this in a normal form. It turns out the cost of ownership.
For the director, this cost of ownership is needed in order to lay it in the financial model. In the end, the cost of maintaining your own server one will go into the price paid by the client for the services of his company or for his goods.
Naturally, he wants full predictability. From his point of view, we need a rather unambiguous connection between business expansion or its contraction and growth or reduction of the cost of ownership of this IT part. From his point of view (I am exaggerating), the CIO gives him some random, unpredictable numbers every time a specific situation occurs. And he says something in Elvish about changes in regulations, new technological trends and the need to buy something at once in a bunch.
Why do you need predictability?
Because if you do not plan the allocation of money for IT, no one will provide them. They still need to take somewhere. That is, you need to know in half a year how much will be spent. More precisely, to know where this money will come from and where it will go.
Note: this is in the CAPEX model, i.e. capital expenditure. When it is paid for something, so that it is “at home”, and not for a service or subscription (I now again simplify very much).
If someone says to him that “we spend on IT 5% of revenue per month,” he will begin to moan softly from pleasure. The connection is obvious and very easy to budget: I got the revenue, I took 5% from it, I gave it to IT specialists. No need to predict anything, no need to quantify anything. This month they did not earn anything - they did not give anything away. They earned twice as much - they gave twice as much. It’s clear where the money comes from.
In practice, of course, this is not the case. You can know exactly the cost per month without reference to revenue. It can be reduced or increased, but can only be abandoned with the closure of the business. So if they didn’t earn anything, they still paid ...
If you have to pay for iron once in 3 years, then this means a very complex model of accumulating this money somewhere inside the company, and according to incomprehensible laws.
Why incomprehensible, because there are still many probabilities. Business in itself is unpredictable: somewhere it can grow faster than the plan, somewhere slower, somewhere everything will not work at all. Predicting IT costs turns into a sea of forks and probabilities, which means having to reserve extra money and take risks.
Do you understand why the OPEX-model with payment, as for SaaS, they really like?
And it is not even about convenience, but about a direct benefit. The point is not only in reserving money, but also in their value. Different money at different times cost differently.
Cost of money
Imagine the situation: there is a need to upgrade the server, but there is no money for it here and now. You can take a loan (within a group of companies or in a bank, it does not matter), and then gradually pay it off. Or you can go to the cloud and pay about the same amount there monthly.
From the point of view of CFO, the first case is more complicated because it imposes an obligation on the company. And the money is about the same, because even if the cloud is more expensive than bare metal for this overpayment on the loan, it is still more profitable.
Usually, the situation is further aggravated by the fact that there is a cost of money in the company: 1000 rubles at the beginning of the year can turn into 1090 rubles at a rate of 9% per annum in a bank or 1,200 rubles if the company's efficiency in processing cash flow is 120%. Usually this is the case, because otherwise, instead of a business, you could put money in a bank. That is, IT always argues on economic efficiency with the investment of money in the purchase of goods (which can be wrapped) or with investments, for example, in production.
The third important feature is flexibility. If the plans change, the server one cannot be dismantled or paid for half the admin. With the OPEX-model, everything is simpler: if the service is provided outside or still somehow by subscription, how much it consumed, paid that much. This is important in the case of seasonal business (in a recession it is very difficult to get money and they are expensive). Therefore, to pay a lot at the peak and a little at a downturn, but at the same time, overpaying a certain amount for the whole year may turn out to be several times more profitable than paying once a year or even paying evenly over the months.
Flexibility is also an opportunity to scale. If the business has suddenly grown, then you have four special effects:
- The purchase of a relatively small batch of iron (which is more expensive than usual);
- The effect of lag on this purchase (up to 3 months);
- The increase in the level of "zoo" (most often it is inhomogeneous iron);
- The need to creatively reshape the budget.
As a result, projects go slower, the company becomes inflexible, costs grow, well, and then you know ... Therefore, all CFOs like expensive tools with instant feedback on scaling - they are cheaper from a global perspective.
The CFO wants predictability and sees the big picture from the level of the company as a whole. CIO wants to save and control, and often in a global perspective, this is not the most economically sound strategy (in the case of frequent changes in the market). CFOs or CEOs are trying to load CIOs with the need to have a good understanding of finances just to solve issues with “different vectors”. The result is that CIOs are beginning to be considered in more detail, taking into account the value of money, proper depreciation and all associated costs. The modern result is a “microservice” business organization, where each individual piece of functionality can be replaced, scaled or disabled according to needs. This is still the best practices for fast-changing markets (and in the last 20 years, almost all of these markets are due to IT progress).
I hope now I have answered the question about why CFOs often do not, as seen from an IT point of view.
This is where my colleagues talked about who should not go into the clouds , and here about the myths of migration . Therefore, there are still questions with a banal fear and unwillingness to adopt new technologies.
And all this above in no way excludes banal errors, the human factor, etc.