Methodology for calculating the optimal level of costs for corporate IT

How to stop being afraid of technology and start applying it effectively

Luddites destroyed machines in 1812. Around the same time, company executives and technical specialists are coming in to optimize IT costs.

How many specialists do you need to service 268 computers? 14 servers for 450 users - is it a lot or a little? Our admins bought such and such a program - is this the right decision? - These are the questions that company leaders ask me when they try to figure out whether their company's costs for corporate IT are optimal or not. Personally, I have no monosyllabic answers to these questions. But even more - I believe that monosyllabic answers to such questions will only harm the company.

The thing is that corporate IT is not just a set of equipment, programs and unshaven people in sweaters with deers. It is also a working tool on which the effectiveness of the company depends. For this reason, talking about the optimality of IT costs, not understanding how these same costs affect the business, to put it mildly, is short-sighted.

Below I want to share my approach to optimizing corporate IT costs. To do this, I will try to clearly explain what benefit the business receives from information technology, how to calculate this benefit and correlate it with costs, and I will propose one of the methods for calculating the optimal level of IT costs. I hope that my article will help some company executives and technical experts to reconsider their views on both corporate IT in general and optimization of costs associated with it in particular. This concludes the introduction and proceeds to the point:

I divided the entire presentation into five consecutive questions:

• When are IT costs at companies optimal, and what is the optimal cost point?
• How to consider the business benefits of information technology and their costs?
• How are IT costs and business benefits related?
• How to calculate the current value of optimal IT costs?
• How to, how not to, and why generally optimize IT costs?

So, in order:

When are the IT costs in companies optimal, and what is the optimal cost point?

No one ever forbade companies to work “the old way” - to draw up contracts by hand, go to the bank with payments, make copies of documents “carbon copy”. No one ever required companies to buy computers and hire admins. Why then are computers everywhere in business?

The fact is that information technology has changed the paradigm of office work in the same way that machines used to change the paradigm of production. Computers allowed to reduce the staff of estimators, typists, accountants, accountants and reduced professional requirements for office professions, reducing them, in fact, to PC operators. Where previously several qualified specialists were required, now one user with a computer is enough.

Companies can still work "the old fashioned way" and contain a huge staff of office employees, but at the current level of technology development, such costs for office work will turn out to be disproportionate or suboptimal. Companies can also try to replace all office employees with robots with artificial intelligence, but at the current level of technology development, such office costs will also be disproportionate or non-optimal.

Information technologies increase the productivity of office workers, but so far are not able to completely replace them. Therefore, the economic effect of the use of information technology depends on the balance between the cost of office staff and the cost of information technology. As always, when choosing between two extremes, there is a middle ground, which I call the optimal cost point : The optimal IT cost

point is the level of IT cost at which the total cost of office work, including IT costs, reaches the lowest possible value.

It is worth noting that behind any point of optimal cost there is well-defined equipment, programs and technical solutions. But technological progress, like business, does not stand still, and the point of optimal costs changes over time, while once purchased equipment and programs remain unchanged. And if the transition to a new point of optimal costs is associated with the abandonment of previously made investments, then this only worsens the economic effect of the use of information technology. For this reason, when assessing the optimality of IT costs, I use the following criterion:

Current IT costs are optimal when they bring the company closer to reaching the optimal cost point and allow it to stay at that point in the long term.

Here I risk completely confusing you, so let's start small - by calculating the optimal cost point for the company at the current time. And for this, you will have to learn to consider the business benefits created by information technology and correctly correlate it with the costs of IT. Therefore, we move on to the next question:

How to consider the benefits to the business of information technology and their costs?

Information technology improves the productivity of office workers by simplifying and automating everyday office routines. Through the use of information technology, the company reduces labor costs for office clerical work and saves office staff time. That is why I prefer to consider the benefits of information technology for business in the work time saved by the use of office employees from their use.

To compare the benefits of information technology with the cost of them, both of these values ​​must be brought to a single system of change. To do this, you can convert the saved time of office employees into money, but such absolute figures differ from one company to another and are not very informative. For this reason, I prefer to measure IT costs and employee time saved in relative terms - as a percentage of the average monthly salary fund for office employees (payroll).

Since IT costs are divided into operational and capital, for the correct comparison of business benefits with the costs of information technology, I adhere to the following principles for calculating these values:

When calculating IT costs, it is necessary to normalize all IT costs for their useful lives. If the server is purchased for 5 years, then it costs the company 1/60 (5 * 12 months) of the cost of its acquisition on a monthly basis. If the project is done for 3 years, then it costs the company 1/36 of its cost every month. The resulting average monthly costs must be correlated with the average monthly payroll of office employees.

When calculating the business benefits of IT, it is necessary to measure how much time the office staff saves IT daily / weekly / monthly and correlate the figure with the total time the employees work for these periods. If employees work on average 8-9 hours a day or about 500 minutes, then every 5 minutes of daily routine costs companies 1% of payroll. If information technology reduces daily routine by at least an hour a day, they save companies 12% of the payroll of office employees.

Further in the article, under the costs of IT and the benefits of it for business, I will mean the values ​​calculated according to the principles above.

For companies that have been operating for more than a year, IT costs can be roughly estimated by calculating all expenses for equipment, programs, admins, communication channels and cartridges for 4-5 years and correlating them with the payroll fund for office employees for the same period. In my experience, for most companies, IT costs with this model of calculation are in the range from 5 to 15% of payroll. Moreover, this ratio depends mainly on the average wage in the company, and not on the coolness of internal IT.

In the case of calculating the absolute benefits of IT, everything is somewhat more complicated. Over the past years, few people already remember how much time it takes to draw up contracts on a typewriter, correspondence with customers through the Russian post office, sketching on purses and keeping inventory in a notebook. The absolute benefits of IT have long been depreciated. But if you try to imagine life without computers, it will become clear that even the simplest computer and printer (no more than 4% of payroll in the long term) saves at least a couple of hours of work per day (every hour of work is 12% from PHOT).

If my task now was to prove the profitability of corporate IT, then on this note it would be possible to end the article. But my task is to find the level of IT costs that will bring the company the maximum economic effect. And for this you will have to deal with the following question:

How are the costs of IT and the benefits to the business?

All parameters of information systems that affect the productivity of office employees can be divided into five general categories:

1. Functionality of information systems
   
   Functionality is a set of actions that users can perform in the information system. It is the functionality of information systems (and the ability to use them) that determines which office tasks can be simplified or even automated, and which ones will have to be done “the old fashioned way”.
   
   
2. The simplicity of the work environment The
   
   simplicity of the work environment is the number of actions taken to achieve the desired result. The need to constantly search for documents on the “file wash”, enter dozens of usernames / passwords for access to different systems, go to the menu of a certain degree of nesting, or run to the printer at the far end of the corridor only complicates the work environment and takes up the working time of employees.
   
   
3. Ergonomics of user workstations
   
   Ergonomics is the correspondence of a workplace to physiological characteristics of a person. If employees are uncomfortable sitting at a computer, if their eyes hurt from the monitor or fingers from the keyboard, then they are more often distracted and work less. No motivation system can resist the pain in the eyes from a cheap monitor or the pain in the back due to an uncomfortable chair.
   
   
4. The speed of information systems
   
   With the speed of information systems, everything is simple - you have to wait for the system to realize (it will start / count / issue the necessary information) or not. The higher the speed of the systems, the less time office workers lose in anticipation of a response from it.
   
   
5. Reliability of information systems and the speed of technical support The
   
   reliability of information systems and the speed of technical support depends on how much time office employees have to spend talking with IT people instead of working. True, in addition to communicating with admins, in some cases, office workers also have to redo work that was lost as a result of a failure, or apologize to clients for missing deadlines.
   

Each of the categories listed above may limit the beneficial effect of other categories. If the information systems in the company either constantly break down, or wildly slow down, or do not have the necessary functionality for work, or if the working environment of office employees is too complicated or inconvenient, it does not matter how well the rest is done - the productivity of office employees will be low.

If we turn to analogies, all of the above criteria affect the productivity of office workers as well as five valves on one pipe affect the amount of fluid passed through it. If one of the valves is fully screwed in, it does not matter how wide the other four are open - the volume of the liquid passed through will be low.

But if in the case of a pipe it is enough to untwist all five gates, then in the case of information systems, untwisting all five quality categories to the fullest is not affordable for most companies. As a result, each of the categories remains “half-open”, and the consequence of such “half-openness” is the presence of distortions, in which one of the categories limits the performance of the others. Such a bias means that the money invested in IT does not give full return, or, in another way, the money invested in IT does not work efficiently. And here we come to the following important definition:

Effective configuration of information systems is such a set of technical solutions, as well as their settings and maintenance, which for the current level of IT costs provide the highest possible productivity of office workers.

If we now draw up the dependence of the benefits for business on IT costs, in which each level of costs corresponds to the most efficient configuration of information systems, we will get some optimal dependence, which I will denote in the future as P (x):

Optimum dependence of the benefits for business (P) on information technology costs (x) - this is such a relationship in which each level of IT costs corresponds to the most effective configuration of information systems for this level of costs .

If you sketch the dependence of business benefits on IT costs, the optimal dependence P (x) on the graph will look something like this (red dotted line):

Figure 1. Sketch of the optimal dependence of business benefits on information technology costs (P (x) - red dotted line on the graph).

A few clarifications on the sketch above:

1. Since I measure the business benefit in the saved salary fund for office employees, the maximum increase in the business benefit from the use of information technologies relative to the current value cannot exceed the current labor costs for office work. Or else, the function P (x) has a limit value. True, in order to reach this limit, you will have to replace all office employees with robots and neural networks.
   
   
2. There is a level of IT costs at which the costs and business benefits of information technology are equal. To mark this point on the sketch, I drew the line P = x (blue dotted line). At the current level of technology development, the replacement of office employees with robots and neural networks is confidently moving to the right side of the schedule.
   
   
3. The x = xcur line and the Pcur point represent the current level of IT costs and the business benefits of using information technology. Typically, in companies, the benefits of IT are less than their possible effective value - I previously explained why this happens.
   
   
4. Point P0, x0 - this is the point of optimal IT costs, in which the costs of office work, including IT costs, reach the lowest possible value. After point x0, any additional investment in IT does not pay off.
   

Now, to find the point of optimal costs, it is enough for us to build the optimal dependence P (x) and find on it such a minimum value of the point x0 for which for any x> x0 the ratio (P (x) -P (x0)) / (x-x0 ) <1. It's time to move on to mathematics:

How to calculate the current value of optimal IT costs?

Here I have to make a reservation that building the optimal dependence P (x) for an abstract company is not an easy task, if at all solved. When building the optimal dependence P (x), it’s easier for me personally to build on the current state of affairs in the company. In this case, the dependence P (x) can be represented as follows:

P (x) = Pcur + P (x1, x2, x3, x4, x5)

where:

P (x) is the total benefit of IT for business,
Pcur is the current benefit of IT for business,
x = xcur + x1 + x2 + x3 + x4 + x5 is the total cost of IT,
xcur is the current cost of IT,
P (x1, x2, x3, x4, x5) - the function of saving staff time depending on the improvement of one of the five categories of quality of information systems (functionality, simplicity, speed, ergonomics, reliability of information systems and the speed of technical support),
x1, x2, x3 , x4, x5 - additional costs for each of the quality categories.

Since Pcur is a constant in the framework of the equation above, to find the optimal cost point, it suffices to construct the optimal dependence of the function P (x1, x2, x3, x4, x5) on the additional IT costs (x1 + x2 + x3 + x4 + x5).

All further decisions are based on the dynamic programming method, which is used both in the theory of computer systems and in such an exciting discipline as “operations research” (on Wikipedia there is a history of the emergence of this discipline - I recommend reading it).

To calculate the optimal form of the function of five variables P (x1, x2, x3, x4, x5), you must first calculate the optimal form of the five functions of one variable: P (x1,0,0,0,0), P (0, x2,0, 0,0), P (0,0, x3,0,0), P (0,0,0, x4,0), P (0,0,0,0,0, x5). To do this, you need to figure out what the company is currently receiving for each of the quality categories, what these losses are associated with and how much it costs to fix them.

To measure current losses for each quality category, you need to find answers to five questions:

1. How much time do employees lose due to lack of functionality?
2. How much time does the unnecessary complexity of systems take from employees?
3. How much time do employees spend on “waiting for a system response”?
4. How much time do employees rest from the computer during the day?
5. How much time do employees stand idle due to a system failure or the need to communicate with technical support?

To answer these questions, you can use monitoring systems, application systems, user surveys, comparative testing and analysis of work processes. Determining the current loss of working time, it is necessary to find the sources of these losses and analyze the cost of eliminating them.

Having compiled a complete list of losses and the costs of eliminating them, as well as reducing these values ​​to a single measurement system (in thousandths of the payroll function), we can proceed to the next step - to compiling the optimal numerical sequences of five functions of one variable. For this, it is necessary to sort out with a uniform step the possible size of investments in each of the quality categories and for each size of investment choose those technical solutions that promise the maximum economic effect. The numerical functions obtained using these principles will look something like this (the numbers below are invented and are used to demonstrate further calculations):

xi	P (xi, 0,0,0,0)	P (0, xi, 0,0,0)	P (0,0, xi, 0,0)	P (0,0,0, xi, 0)	P (0,0,0,0, xi)
1	0.5	0.5	3	5	1
2	0.5	1	3	6	4
3	9	1	3	8	6
4	10	14	8	8	10
5	14	18	9	8	17
6	fifteen	20	fifteen	8	19
7	fifteen	21	fifteen	8	20
8	fifteen	22	fifteen	8	21

Table 1. Illustration of the numerical dependence of each of the quality categories of information systems on additional investments in them. The numbers in the table are used to illustrate the calculations and are the invention of the author.

A small clarification - xi - is the amount of additional investment in each of the quality categories, relative to the current size of the costs. It must be defined as the difference between the cost of the solution used and the new one. That is, xi is not the cost of a new cool piece of iron or a program, but the difference in the cost of a new cool and old simpler (taking into account their useful life).

If we assume that all six quality categories do not affect each other and are additive (in practice, this is not entirely true), then the function P (x1, x2, x3, x4, x5) can be represented as follows:

P (x1, x2, x3, x4, x5) = P (x1,0,0,0,0) + P (0, x2,0,0,0) + P (0,0, x3,0,0 ) + P (0,0,0,0, x4,0) + P (0,0,0,0,0, x5)

Based on this assumption and knowing the numerical dependences of each of the five functions of one variable, we can compose the optimal dependence of P (x1, x2, x3, x4, x5) on Δx = x1 + x2 + x3 + x4 + x5, choosing for each investment size such a configuration of functions of one variable that give the maximum economic effect:

Δx	The space of maximum solutions	Maximum value of P (x1, x2, x3, x4, x5)	Right Derivative dP / dx
1	P (0,0,0,1,0)	5	3
2	P (0,0,1,1,0)	8	1
3	P (3,0,0,0,0,0), P (0,0,1,2,0), P (0,0,0,1,2)	9	5
4	P (0.4.0.0.0), P (3.0.0.1.0)	14	5
5	P (0.4.0.1.0)	19	4
6	P (0,5,0,1,0)	23	3
7	P (0,5,1,1,0)	26
8
....
31	P (5,7,6,3,8), P (5,8,6,3,7), P (6,7,6,3,7)	79	1
thirty	P (5,8,6,3,8), P (6,7,6,3,8), P (6,8,6,3,7)	80	1
31	P (6,8,6,3,8)	81

Table 2: Illustration of the methodology for finding the optimal dependence of business benefits on IT depending on costs. All values ​​are in thousandths of payroll.

I built the table above on the assumption that all five quality categories are independent and additive. In practice, this is not entirely true, therefore, the resulting space of optimal solutions must be additionally checked for additivity. According to the results of the test, some of the maximum solutions may decrease due to the effect of dyssernergy, while others, on the contrary, may grow due to the effect of synergy.

Having checked all the intermediate solutions for synergy / dissynergy, we can proceed to calculate the optimal cost point. To do this, moving in the direction of increasing costs, for each point you need to check the value of the right derivative dP / dx. Points at which the value of the right derivative is less than or equal to unity should be checked for the relation (P (x) -P (x0)) / (x-x0) <1, for all x> x0, where x0 is the current point. The first point at which this ratio is satisfied will be the point of optimal cost.

In the sketch below, I note how the calculated area will look:

Figure 2. Illustration of the optimal dependence P (x) according to the calculation results (blue line on the graph).

Now it remains only to tell what to do next with this point of optimal costs:

How to, how not to, and why generally optimize IT costs?

As I said earlier, the point of optimal costs changes over time due to a change in business or a change in technology. Only the previously purchased equipment, programs and technical solutions remain unchanged. Therefore, in order to achieve maximum economic effect from corporate IT, it is necessary to take into account the possible changes in the point of optimal costs in the future. In practice, this means that:

• Capital expenditures on IT should not be lower than the current value of the optimal cost point and take into account its possible change in the future, if such a change can lead to premature cancellation of technical solutions related to capital expenditures.
  
  
• Operating expenses for IT should provide the maximum long-term economic effect for the company, taking into account both a possible change in the value of the point of optimal costs in the future, and additional costs that may be required for such a change.

But often, companies do not care about the economic effect of the use of information technology, and the optimization of costs for information technology implies only one thing - reducing IT costs. The principles of such optimization are slightly different:

• Invest in IT only in case of emergency.
• Acquire technical solutions in sizes sufficient to solve the problems, but no more.
• Hard price qualification - choose the cheapest solutions from all possible alternatives.
• If possible, capital expenditures can be transferred to operating expenses by leasing equipment and programs, and subscribing to cloud services.
• Customize and develop information systems exclusively by full-time specialists, without involving external contractors.

The problem is that such optimization does not give the desired savings - equipment and programs purchased according to these principles have to be written off ahead of time, renting in the long term eats up more money, and staff specialists do not have enough strength to realize all the functionality laid down in information systems. In the long run, the company does not save anything, but it reduces its effectiveness significantly.

Today, many are confident that the ability to save on IT creates competitive advantages, but in practice, such savings only exacerbate the financial situation in the company. Competitive advantages today create the ability to squeeze out of the information technology the maximum benefit for the business minus the cost of them. The gap between trial and error lies only between theoretical knowledge of how to do this and practical implementation.

I have been dealing with the topic of finding a balance between business benefits and IT costs since 2007, and for those who decide to do this in their company, I want to advise only one thing - do not be afraid to make mistakes. No matter how weighed you make decisions, you cannot avoid mistakes, but you need to be afraid not of mistakes, but of the fact that you will not see them and cannot learn from them. The whole theory above is a high-level interpretation of the experience I have received, and perhaps there are errors in this theory as well. It is for this reason that I share it in order to find out about the mistakes I made as early as possible, and a little earlier to bring my clients closer to reaching the point of optimal IT costs.

Ivan Kormachev
IT Department Company
www.depit.ru