Automation of IT processes in conditions of low motivation and / or qualifications of performers
The main area of my work for 16 years is the automation of enterprises. Since it all began in 1996, in a small town and in the absence of literature on the programming of personal computers, everything was done by trial and error or the “method of scientific testing”. The times have changed, a lot of techniques have appeared (I don’t use them myself) for automation, implementation and support of software for automation of activities.
But at the majority of enterprises (we will immediately determine, there are small and medium businesses, state organizations, large enterprises in a state of collapse, total: where they can’t / don’t want to hire big teams with a modern automation approach) the “smart” implementation just falls apart when it comes to comes to ordinary performers. And the reason is not so much in the low qualification of middle-level employees and performers, but in the motivation of these employees. Therefore, developed certain "rules" that allow you to:
- speed up the automation process;
- find a common language with the leadership;
- do not spoil your nerves;
Serious implementations start only with the support of management (persons who have the authority to make decisions) of the organization. Management, you must paint what tasks, in what order are going to do. Employees convey the idea that they are fulfilling the requirements of the leadership, and we can fulfill them together, or it will be done without us.
')
No matter how “profitable” the introduction over the years - do not spoil your karma. You need to implement quickly, the faster you solve the problem of one customer, the faster the second third will appear. With the rapid implementation of the customer, rather expand the task (and increase the budget). Otherwise, you can get a minus in reputation and a minus in payment.
Performers whose work you are to automate in most cases are not interested in this, for several reasons:
a) laziness - “it used to be easier”;
b) unwillingness to relearn - “I just remembered (a) what, when, where to press, why relearn?”;
c) motivation - “they won't raise my salary for retraining. They want too much from me for this RFP ";
d) “I am a guru” - “you don’t understand anything, I organized everything here and it’s better not to do it”
a) fear for the workplace - “they will now introduce and reduce”;
b) fear will not learn - “I will not learn this”;
It is easier to work with group 2, because There is no internal confrontation, but there is a fear of "not getting into the clip." It is convenient and necessary to be guided by such employees. Internally, they are ready to help with the implementation, they show problem areas, but you need to be very careful and not “substitute” them in the person of group 1. Consider group 1 and group 2 - this is one team “implementers come and go, fraternity ring ... lecture - forever. "
Arguments for group 2
“A) fear for the workplace -“ they will now introduce and reduce ”
- the management (initiator of the implementation) is now focused on employees who will quickly and efficiently (save on implementation) reorganize processes. And we are interested, quickly and without problems to complete the task. Therefore, a quick and high-quality solution of the problem is beneficial for everyone.
"B) fear will not learn -" I will not learn this "
- it is precisely all the difficulties of new processes that we must overcome with you - so that the manual will not hear about these difficulties.
This is a group of litmus - your work, dragging over to your side - your victory. Group 1 lost credibility in front of group 2 - your victory.
Arguments for group 2 must be technical. And organizational discussed under the auspices of leadership.
"A) laziness -" it used to be easier "
Your proposed solutions should simplify the tasks and separate them into processes.
b) unwillingness to relearn - “I just remembered (a) what, when, where to press, why relearn”;
The task must be so divided that each process can be described by an instruction and / or a scheme that fits on 1-2 sheets. The best option when the interface itself is an instruction
c) motivation - “they won't raise my salary for retraining. For such a RFP they want me to be able to (a) ";
with this problem, the only thing that helps is to ask the executor “we should solve this problem in the implementation mode - the management, or it will be solved (and not to rise more) in the execution mode - the management mode”. In fact, to raise this problem to the level of leadership, it is not worth it, it increases the confrontation.
d) “I am a guru” - “you don’t understand anything, I organized everything here and it’s better not to do it”
ask the “guru” to describe how the processes are currently organized, what are the strengths and weaknesses of the current scheme. At 99%, the process of creating circuits will begin at best.
The most important thing to understand is that the management first of all wants to automate the processes carried out by personnel that are not valued by the EXECUTORS. Management wants to:
- did not require training of performers;
- performers were interchangeable;
- the work of performers does not violate more complex tasks;
- the work scheme was conveyor;
In the actions of the performers need to understand:
- the contractor has no motivation to take responsibility, so all functions should be regulated;
- the performer does not have to make decisions, therefore all variants of events must be regulated;
- the performer, while performing the work, thinks not about her, but about personal worries, therefore the processes that he performs should be focused on simple actions: clicked, chose, entered, the “Ok” button.
And most importantly, “The performers do the work , not work .” Therefore, claims can only be the correctness of the instructions.
1. The solution of the problem should be organized in a pipeline. The simpler each stage will be, the more efficiently the performer will work.
2. The contractor must be put on one working (day, shift, position) cycle, on one function. After a few hours of execution, he will perform it on the machine.
3. Each stage of the pipeline must fit into 7 operations (“Magic number seven plus or minus two”).
4. If the stage of the pipeline is important, duplicate it with another performer, and compare the results in the program. If there are discrepancies in the results, after a time (the performer does not need to show that the stage is performed again), return to the over-fulfillment.
Input of archival materials, with the subsequent classification, aggregation and analysis.
1.1. Primary data is stored on paper, in various forms and contents. To enter into the system, it is necessary to select the types of data by size, type, etc. (A4, A5, statements, certificates, extracts). Sort them before entering. When entering a new type - to expand the software.
1.2. The data is entered by scanning into an electronic archive, which will be processed in the following steps.
1.3. One type of primary data introduces one performer in one cycle.
2.1. When scanning, each primary sheet receives a sequence number, which is reflected on the sheet, if it is not possible (important primary data) in the registry, which is attached to the folder with the sheets, and is returned to the archive. This will provide an opportunity to quickly find and restore scanned in case of loss. Numbering is needed for further end-to-end communication.
2.2. Scanning must fit in 7 operations:
new folder
took the folder with the questionnaires -> attached the registry or indicated the initial scan number -> sheets to the left of the scanner -> folder to the right of the scanner
Processing sheets from a folder
took the sheet -> to the scanner -> scan (the data went to the electronic archive) got the number) -> recorded the number on the sheet or in the registry -> put the sheet in the folder -> repeated
3.1. I will not describe the structure of the database, in my opinion everything is clear.
3.2. Each performer within the cycle recognizes one type of document.
3.3. If the data is listed, each line (2-3 lines) is allocated to the questionnaire and the performer works with the questionnaire. for example
3.3.1. The questionnaire represents the table: Indicator, Value. (Indicators: Name, Date of Birth, Gender, Country, City)
3.3.2. The statement is a table: No., Full Name, Date of Birth, Gender, Position, Country, City ... So, every line is considered a questionnaire.
3.4. The user for recognition to select the line (for the list), indicator and value.
3.4.1. For the questionnaire, for example, when focusing on the name field, select the indicator and value.
3.4.2. For statements, select, row, indicator and value.
3.5. In these examples, we have blocks
Indicators: Name, Date of Birth, Gender, Position.
The statement still grouped in rows.
3.6. The number of fields for input per cycle, it is desirable to limit. For example, the performer enters from the questionnaire only the full name and gender. It will quickly bring the work to automatism.
3.7. The same block is introduced by two or more performers (duplicate). Based on the reconciliation of the discrepancy sent to re-enter. This will save us from checking manually.
1. The streamlined process of converting paper data into digital presentation is the primary digital data “PCD”.
2. Scanned archive (part or all)
3. Recognized data (part or all)
Now it is possible on the basis of the PCD to organize BD2 in which to classify data by types of primary data and blocks, while not violating the PCD. For example, all statements in BD2 will be in 1 table because This is essentially a questionnaire and vice versa all the questionnaires summarized in the statements. In this case, transformations in case of errors or change of rules can be reimported. At this stage, the main thing is to develop the rules for converting DCD to BD2.
A simple example: to create (or import data into the software of a questionnaire of individuals) on the basis of a PDD. In this case, both databases will remain independent, but connected.
We organize the aggregation on the new tables, which are connected with the PCD.
Imagine the problem: In the archives of many duplicates of the full name, country, city in the Table of the Questionnaire. Delete the data from the PCD, it is not necessary, because when errors are found in the future, you can always find out at what stage it happened and where to look for the paper original.
Fig.1. Aggregation of duplicate profiles
We form an intermediate table (T2) where the Questionnaire-Personality links will be stored.
We form an intermediate table (T3) where the Personalities will be stored and fill it with non-duplicated records. When finding duplicates in (T3) in (T2), we set the correct value in the ID_Personality. If necessary, you can create a descriptive table where you specify the reasons for the association in (T2).
Fig.2. Aggregation of duplicates of countries of cities
We act similarly with the full name. Tables (T2) (T4) for Countries, (T3) (T5) for cities. (T7) for communication city + country.
- we do not violate the PD;
- it remains possible to adjust the processed data;
- the task of the contractor is reduced to “verify data that looks like duplicated”.
But at the majority of enterprises (we will immediately determine, there are small and medium businesses, state organizations, large enterprises in a state of collapse, total: where they can’t / don’t want to hire big teams with a modern automation approach) the “smart” implementation just falls apart when it comes to comes to ordinary performers. And the reason is not so much in the low qualification of middle-level employees and performers, but in the motivation of these employees. Therefore, developed certain "rules" that allow you to:
- speed up the automation process;
- find a common language with the leadership;
- do not spoil your nerves;
1. The level of implementation
Serious implementations start only with the support of management (persons who have the authority to make decisions) of the organization. Management, you must paint what tasks, in what order are going to do. Employees convey the idea that they are fulfilling the requirements of the leadership, and we can fulfill them together, or it will be done without us.
')
2. Terms of implementation
No matter how “profitable” the introduction over the years - do not spoil your karma. You need to implement quickly, the faster you solve the problem of one customer, the faster the second third will appear. With the rapid implementation of the customer, rather expand the task (and increase the budget). Otherwise, you can get a minus in reputation and a minus in payment.
3. Do not be offended by "saboteurs"
Performers whose work you are to automate in most cases are not interested in this, for several reasons:
Group 1. do not want to change the usual mode of operation because:
a) laziness - “it used to be easier”;
b) unwillingness to relearn - “I just remembered (a) what, when, where to press, why relearn?”;
c) motivation - “they won't raise my salary for retraining. They want too much from me for this RFP ";
d) “I am a guru” - “you don’t understand anything, I organized everything here and it’s better not to do it”
Group 2. are afraid to change the usual mode of operation because:
a) fear for the workplace - “they will now introduce and reduce”;
b) fear will not learn - “I will not learn this”;
Consider more.
Group 2
It is easier to work with group 2, because There is no internal confrontation, but there is a fear of "not getting into the clip." It is convenient and necessary to be guided by such employees. Internally, they are ready to help with the implementation, they show problem areas, but you need to be very careful and not “substitute” them in the person of group 1. Consider group 1 and group 2 - this is one team “implementers come and go, fraternity ring ... lecture - forever. "
Arguments for group 2
“A) fear for the workplace -“ they will now introduce and reduce ”
- the management (initiator of the implementation) is now focused on employees who will quickly and efficiently (save on implementation) reorganize processes. And we are interested, quickly and without problems to complete the task. Therefore, a quick and high-quality solution of the problem is beneficial for everyone.
"B) fear will not learn -" I will not learn this "
- it is precisely all the difficulties of new processes that we must overcome with you - so that the manual will not hear about these difficulties.
Group 1
This is a group of litmus - your work, dragging over to your side - your victory. Group 1 lost credibility in front of group 2 - your victory.
Arguments for group 2 must be technical. And organizational discussed under the auspices of leadership.
"A) laziness -" it used to be easier "
Your proposed solutions should simplify the tasks and separate them into processes.
b) unwillingness to relearn - “I just remembered (a) what, when, where to press, why relearn”;
The task must be so divided that each process can be described by an instruction and / or a scheme that fits on 1-2 sheets. The best option when the interface itself is an instruction
c) motivation - “they won't raise my salary for retraining. For such a RFP they want me to be able to (a) ";
with this problem, the only thing that helps is to ask the executor “we should solve this problem in the implementation mode - the management, or it will be solved (and not to rise more) in the execution mode - the management mode”. In fact, to raise this problem to the level of leadership, it is not worth it, it increases the confrontation.
d) “I am a guru” - “you don’t understand anything, I organized everything here and it’s better not to do it”
ask the “guru” to describe how the processes are currently organized, what are the strengths and weaknesses of the current scheme. At 99%, the process of creating circuits will begin at best.
4. A little bit about the performers or roughly about the truth.
The most important thing to understand is that the management first of all wants to automate the processes carried out by personnel that are not valued by the EXECUTORS. Management wants to:
- did not require training of performers;
- performers were interchangeable;
- the work of performers does not violate more complex tasks;
- the work scheme was conveyor;
In the actions of the performers need to understand:
- the contractor has no motivation to take responsibility, so all functions should be regulated;
- the performer does not have to make decisions, therefore all variants of events must be regulated;
- the performer, while performing the work, thinks not about her, but about personal worries, therefore the processes that he performs should be focused on simple actions: clicked, chose, entered, the “Ok” button.
And most importantly, “The performers do the work , not work .” Therefore, claims can only be the correctness of the instructions.
5. Software Requirements
1. The solution of the problem should be organized in a pipeline. The simpler each stage will be, the more efficiently the performer will work.
2. The contractor must be put on one working (day, shift, position) cycle, on one function. After a few hours of execution, he will perform it on the machine.
3. Each stage of the pipeline must fit into 7 operations (“Magic number seven plus or minus two”).
4. If the stage of the pipeline is important, duplicate it with another performer, and compare the results in the program. If there are discrepancies in the results, after a time (the performer does not need to show that the stage is performed again), return to the over-fulfillment.
6. Example of task
Input of archival materials, with the subsequent classification, aggregation and analysis.
1. Preparation.
1.1. Primary data is stored on paper, in various forms and contents. To enter into the system, it is necessary to select the types of data by size, type, etc. (A4, A5, statements, certificates, extracts). Sort them before entering. When entering a new type - to expand the software.
1.2. The data is entered by scanning into an electronic archive, which will be processed in the following steps.
1.3. One type of primary data introduces one performer in one cycle.
2. Scan
2.1. When scanning, each primary sheet receives a sequence number, which is reflected on the sheet, if it is not possible (important primary data) in the registry, which is attached to the folder with the sheets, and is returned to the archive. This will provide an opportunity to quickly find and restore scanned in case of loss. Numbering is needed for further end-to-end communication.
2.2. Scanning must fit in 7 operations:
new folder
took the folder with the questionnaires -> attached the registry or indicated the initial scan number -> sheets to the left of the scanner -> folder to the right of the scanner
Processing sheets from a folder
took the sheet -> to the scanner -> scan (the data went to the electronic archive) got the number) -> recorded the number on the sheet or in the registry -> put the sheet in the folder -> repeated
3. Recognition of scanned data
3.1. I will not describe the structure of the database, in my opinion everything is clear.
3.2. Each performer within the cycle recognizes one type of document.
3.3. If the data is listed, each line (2-3 lines) is allocated to the questionnaire and the performer works with the questionnaire. for example
3.3.1. The questionnaire represents the table: Indicator, Value. (Indicators: Name, Date of Birth, Gender, Country, City)
3.3.2. The statement is a table: No., Full Name, Date of Birth, Gender, Position, Country, City ... So, every line is considered a questionnaire.
3.4. The user for recognition to select the line (for the list), indicator and value.
3.4.1. For the questionnaire, for example, when focusing on the name field, select the indicator and value.
3.4.2. For statements, select, row, indicator and value.
3.5. In these examples, we have blocks
Indicators: Name, Date of Birth, Gender, Position.
The statement still grouped in rows.
3.6. The number of fields for input per cycle, it is desirable to limit. For example, the performer enters from the questionnaire only the full name and gender. It will quickly bring the work to automatism.
3.7. The same block is introduced by two or more performers (duplicate). Based on the reconciliation of the discrepancy sent to re-enter. This will save us from checking manually.
At this moment we have:
1. The streamlined process of converting paper data into digital presentation is the primary digital data “PCD”.
2. Scanned archive (part or all)
3. Recognized data (part or all)
Classification
Now it is possible on the basis of the PCD to organize BD2 in which to classify data by types of primary data and blocks, while not violating the PCD. For example, all statements in BD2 will be in 1 table because This is essentially a questionnaire and vice versa all the questionnaires summarized in the statements. In this case, transformations in case of errors or change of rules can be reimported. At this stage, the main thing is to develop the rules for converting DCD to BD2.
A simple example: to create (or import data into the software of a questionnaire of individuals) on the basis of a PDD. In this case, both databases will remain independent, but connected.
Aggregation
We organize the aggregation on the new tables, which are connected with the PCD.
Imagine the problem: In the archives of many duplicates of the full name, country, city in the Table of the Questionnaire. Delete the data from the PCD, it is not necessary, because when errors are found in the future, you can always find out at what stage it happened and where to look for the paper original.
Fig.1. Aggregation of duplicate profiles
We form an intermediate table (T2) where the Questionnaire-Personality links will be stored.
We form an intermediate table (T3) where the Personalities will be stored and fill it with non-duplicated records. When finding duplicates in (T3) in (T2), we set the correct value in the ID_Personality. If necessary, you can create a descriptive table where you specify the reasons for the association in (T2).
Fig.2. Aggregation of duplicates of countries of cities
We act similarly with the full name. Tables (T2) (T4) for Countries, (T3) (T5) for cities. (T7) for communication city + country.
Result:
- we do not violate the PD;
- it remains possible to adjust the processed data;
- the task of the contractor is reduced to “verify data that looks like duplicated”.
Source: https://habr.com/ru/post/145409/
All Articles