Employee performance to support work productivity: a Pls approach in agro-input suppliers company

The purpose of this study is to analyze the performance of employees in supporting the productivity of PT Petrokimia Kayaku Gresik. In this study, there are three variables, namely independent variables, work motivation, job training and leadership style variables, two dependent variables, namely employee performance and productivity. The method used is Sem-Pls and uses the Warp PLS 7.0 software, while the measuring instrument in this study uses a Likert measuring instrument starting from the smallest 1 to 5 of the largest. The things examined in this study are the direct influence, or without using the mediating variable (intervening) and the indirect effect or without using the mediating variable, the mediating variable in this study is employee performance. The results in this study are employee work motivation, employee job training and the response of a positive and significant leadership style to the performance of PT. Petrokimia Kayaku Gresik employees. The third variable is a variable from the company and has a significant positive effect on productivity.


Introduction
An organization in carrying out activities to achieve its goals has several interrelated and influential factors. One of these factors is very important, which is used to drive other factors, namely human resources. Therefore, organizations are required to manage and optimize human resources. Employees are the principal capital of an organization and have the most critical role in an organization, namely as stakeholders or stakeholders in all activities in an organization or company.
According to Nawawi 2011 workforce, workers, employees, human potential as the driving force of the organization in realizing its existence or potential which is an asset and functions as non-material capital in a business organization that can be transformed into a real potential physically and non-physically in realizing the existence of the organization. Employees can work well if they have high performance so that they can produce satisfactory achievements. In order for company activities to run well, the company must have employees who are knowledgeable and skilled and have high loyalty to the company. The success in achieving performance is not only determined by the mastery of in-depth knowledge but also strongly influenced by the behavioural attitudes that employees have in dealing with work. According to Rivai 2015 work performance is the result of an increase of a person in total during a specific period in carrying out a task compared to various possibilities, such as work standards, target or target or criteria that have been previously determined and agreed upon. In order to achieve maximum employee performance, employees need work motivation, job training and functional leadership responsibilities. According to Sinambela 2012 states that employee performance is defined as an employee's ability to perform specific skills. Employee are essential because this performance is not known what can be done by employees in the tasks assigned to them. For this reason, it is necessary to determine clear and measurable criteria that are jointly determined as a reference. In terms of motivation, several important indicators are salary, health insurance and old age or position to work more diligently, and there are bonuses or praise for satisfactory work results and promotion of promotion. The importance of employees in the organization, employees must pay more serious attention to the tasks they do so that organizational goals can be achieved. With high work motivation, employees will work harder in carrying out their work. In fact, with low work motivation, employees do not have enthusiasm for work, give up quickly, and have difficulty completing their work. This motivation is following Oloko (2015). Motivation is the key to an organization that is successful in maintaining survival in the organization with a reliable way and assistance for survival. Motivation is giving them proper guidance or direction, resources and imbalances to get them inspired and interested in working the way you want them to.
In addition to work motivation factors, job training such as attending seminars or training for employees is no less essential to improve employee performance to support work productivity. With job training for employees, employees will experience a process of systematically changing employee behaviour to achieve organizational goals. This increase in expertise and knowledge is by the opinion of Widodo (2015), job training is a series of individual activities to systematically increase skills and knowledge so that they can have professional performance in their fields. Job training is a learning process that allows employees to carry out current jobs by company standards.
Human resources play a role in processing and utilizing resources and materials so that later they become a product. Therefore, to improve employee performance, it is necessary to pay attention so that human resources can work efficiently and show performance that can contribute to productivity is a fundamental problem of various management and leadership concepts. Not only in terms of production but human resources also affect the company in achieving the profit target that has been set by a company. There fore human resources must always be upgraded to the quality of the way they IOP Conf. Series: Earth and Environmental Science 803 (2021) 012054 IOP Publishing doi:10.1088/1755-1315/803/1/012054 3 work so that their performance will be maximized.
The role of leadership in leading its employees and making decisions in an issue is significant in an organization or company because good leadership will have an impact on the success or failure of an organization, both business and public-oriented besides that, leadership is also very strategic and essential for achievement. Mission, vision and goals of an organization are one of the motives that encourage people always to investigate the ins and outs of leadership.
PT Petrokimia Kayaku is a company engaged in the agrochemical sector which produces main products in the form of pesticides, biological fertilizers, livestock probiotics and mosquito repellents.
PT Petrokimia Kayaku's products have been recognized according to FAO International Standards -specified for plant protection products, thus meeting safety standards for crops. With the COVID-19 pandemic, PT. Petrokimia Kayaku, as a company engaged in agriculture, the company has no problems with turnover or reducing the number of employees. According to observations and interviews with the head of the finance department of PT. Petrokimia Kayaku, This is because production is still running and has not decreased the number of demands from consumers.
The problems encountered at the PT Petrokimia Kayaku company include. First, the Human Resource department often holds job training for employees in the company, especially training that is prioritized for new employees, where frequent job training has an impact on work buildup in several departments, both things. What becomes an uneasiness for employees at the company is the lack of communication between the leadership and employees, which has an impact on the lack of familiarity between leaders and employees in the company. Therefore, the urgency in this study requires further research on the effect of employee motivation, employee job training and leadership style on employee performance to support the productivity of the company at PT Petrokimia Kayaku Gresik.
The purpose of this study is to analyze the performance of employees in supporting the productivity of PT Petrokimia Kayaku Gresik.

Materials and Methods
The research approach used is a quantitative approach, which is an approach that requires using numbers starting from data collection, data interpretation, to the appearance of the results. This approach uses primary and secondary data, where primary data is the result of observations, interviews and questionnaires, and secondary data in the form of documentation and supporting literature. The influence of the variables under study can be explained descriptively by describing a systematic description and explanation based on the data obtained from the study. PT. Petrokimia Kayaku which is a subsidiary of PT. Petrokimia Gresik, which is engaged in agriculture and is one of the largest pesticides producing companies in Indonesia. The choice of location is also based on the existence of human resource problems in the company so that research is needed in order to improve human resource management. In addition, considerations of cost, time and research personnel are also factors that determine the place of research.
Determination of the sample in this study using a non-probability sampling approach using the resampling method of boostrap. The sample used in this study includes permanent employees without exception the leaders who work at PT. Petrokimia Kayaku, which amounted to 120 people, so that all of them were used as samples in the study.

Variable Descriptive Analysis
The description of the characteristics of the respondent's answer to each variable in terms of the results of the frequency distribution and average value. Measurement of each item from each variable used a Likert scale with a score of 1-5 According to Solimun, et al. 2017. Each of these variables will be seen its frequency distribution and average value. Measurement of these variables using the Likert scale method. Descriptive description of each variable will be described in Table 1. 1 and the maximum value is 5. This shows that the assessment of all statement items has a varying value with a value range of 1 to 5.

Model Equation Structural Equation Model
The data analysis used in this research is the Structural Equation Model (SEM) method based on Partial Least Square (PLS). The SEM-PLS analysis method has two stages, namely the evaluation of the measurement model (outer model) and the structural model (inner model). The software used in the analysis was Warp PLS 7.0. Source: primary data processed in 2020

Convergent Validity Test
Convergent validity is the extent to which a measure is positively correlated with alternative measures of the same construct. Using the domain sampling model, reflective construct indicators are treated as a different (alternative) approach to measuring the same construct. The p value or probability value or asymptotic significance is the probability for a statistical model, where the null hypothesis is true, then the parameter will be equal to or greater than the true value. While the loading factor is a coefficient that explains the level of relationship between indicators and latent variables. In general, the higher the loading factor, the better, and values below 0.40 cannot be interpreted.

Discriminant Validity Test
The validity test of discrimination in the SEM-PLS analysis was evaluated by looking at    Source: primary data processed in 2020 Table 5 shows the value of the square root AVE for each variable. The value of the AVE squared is indicated by the number enclosed in brackets. The criteria for the fulfilled AVE squared value is that the AVE squared value must be higher than the correlation between latent variables in the same column above or below it. The table above also explains that the square AVE value on all variables has a higher value than the correlation between its latent variables. For example, the work motivation variable has an AVE square value of 0.729, higher than the work motivation value for employee training of 0.551, the value of work motivation on leadership style is 0.509, the value of work motivation on employee performance is 0.717, the value of work motivation on productivity is 0.607.

Evaluation of the Structural Model (Inner Model)
Inner model evaluation or structural model evaluation is assessed by looking at the effect of the significance value on the R-Square value, Full Collinearity VIF, Q-Square, and Goodness of fit (GoF).
Evaluation is carried out to ensure that the structural or inner model made is accurate or good. The value of R-Square is the value of the coefficient of determination on the dependent variable, Q-square is a value that indicates the existence of predictive relevance or relevance between exogenous latent variables to endogenous latent variables. Goodness of fit (GoF) is a value that validates the structural model as a whole. In the evaluation stage, the inner model aims to see the influence between the variables contained Source: primary data processed in 2020 AFVIF values in this study are ideal because the resulting values are ≤ 3.3, namely 1,810 and 2,022, so that this study does not experience multicollinearity problems between indicators and between latent variables, multicollinearity is defined as the existence of inter-correlation between several independent variables. The correlation between the independent variables causes the partial regression coefficient to be less reliable (Dajan, 1986).
The GoF value in Table 8 shows a large value because ≥ 0.36, namely 0.561, so the model can be accepted in this study. The SPR (Sympson's paradox ratio) value and the RSCR (R-squared contribution ratio) value in this study can be said to be ideal because the results are 1. The criteria for SPR and RSCR are accepted if ≥ 0.7 and ideal if = 1, while for RSCR it is accepted if ≥ 0.9 and ideal if = 1. Ideally this value means that there is no problem. Sympson's paradox means a phenomenon in probability and statistics, where trends appear in several different data sets but disappear or reverse when these groups are combined. in this research model and model is also free from negative R-Squared.
The value of the SSR (Statistical suppression ratio) obtained is 1 and meets the requirements ≥ 0.7. This means that 100% of the paths in the model are free of statistical suppression. The last value measured in Goodness of Fit is the value of the NLBCDR (Nonlinear Bivariate Causality Direction Ratio). The value of the NLBCDR in this study has met the criteria because the measurement result is 1, greater than ≥ 0.7, which means that 100% of the associated paths in the model support the hypothesis or there is no causality problem in the model.

Determination of indirect (mediation) and direct models
This chapter will discuss about the indirect model determinant and the direct model.

Analysis using an indirect model
The following is a model test using an indirect model, namely a model using the employee performance mediating variable (Y1). b) The yellow ellipse shows the exogenous variables and the red ellipse shows the endogenous variables.
c) The direction of the black line arrow explains the influence between work motivation, job training and leadership style variables on the productivity variable through employee performance variables, which we call this employee performance variable as a mediating or intervening variable. This is the reason why this analysis is called an analysis with an indirect model. d) P: probability / opportunity, β: The risk measure used when comparing one variable to another, while R2: The magnitude of the influence of certain independent latent variables on the dependent latent variable.
e) The β and p-value of each variable will be compared with the significance value specified in the Solimun (2017) book, which is <0.05. f) If the significance value is less than 0.05, it means that it has a positive and highly significant effect and if it is greater than 0.05 it is said to be significant.
 The influence of the work motivation variable (X1) on work productivity (Y2) through the employee performance variable (Y1) produces a β value of 0.624 and a p-value of 0.001 is less than 0.05 which means that it has a positive and highly significant effect. This is also in accordance with an incident found in the field, namely that employees will be more enthusiastic in doing their work and achieve high productivity with the work motivation of the company for its employees.
 The effect of job training (X2) on work productivity (Y2) by means of the employee performance variable (Y1) produces a β value of 0.125 and a p-value of 0.08 is greater than 0.05, which means  that it has a positive and significant effect. This is also in accordance with an incident found in the field, namely that employees will find it easier to do their jobs and achieve high productivity IOP Conf. Series: Earth and Environmental Science 803 (2021) 012054 IOP Publishing doi:10.1088/1755-1315/803/1/012054 12 with job training conducted by companies to hone the skills and talents of their employees.
 The influence of leadership style (X3) on work productivity (Y2) through the employee performance variable (Y1) produces a β value of 0.043 and a p-value of 0.32 Which is greater than 0.05, which means that it has a positive and significant effect. This is also in accordance with events found in the field, namely that employees will find it easier to do their jobs and achieve high productivity with a good leadership style from the leaders to their employees.
 The value of the employee performance variable (Y1) which is obtained from the variable value of work motivation, job training and leadership style on work productivity (Y2) results in a β value of 0.33 and a p-value of 0.001 less than 0.05, which means that it has a positive effect. and highly significant. This is also in accordance with an incident encountered in the field, namely employees will be more enthusiastic and easier to do their work so that they can achieve high productivity with good employee performance from employees to the company.

Analysis using the Direct Model
a) Based on the significance test, the direct effect on the productivity variable can be described as follows: b) The influence of the work motivation variable (X1) on work productivity (Y2) results in a β value of 0.210 and a p-value of 0.001. less than 0.05, which means that it has a positive and highly significant effect. This is also in accordance with an incident found in the field, namely that employees will be more enthusiastic in doing their work and achieve high productivity with the work motivation of the company for its employees.
c) The influence of the job training variable (X2) on work productivity (Y2) produces a β-value of 0.18 and a p-value of 0.02, less than 0.05, which means it has a positive and highly significant effect. This is also in accordance with an incident found in the field, namely that employees will find it easier to do their jobs and achieve high productivity with job training conducted by companies to hone the skills and talents of their employees.
d) The influence of the leadership style variable (X3) on work productivity (Y2) produces a β value of 0.016 and a p-value of 0.04 less than 0.05 which means that it has a positive and highly significant effect. This is also in accordance with events found in the field, namely that employees will find it easier to do their jobs and achieve high productivity with a good leadership style from the leaders to their employees.  13 Based on the table above, it can be seen that the direct effect of the work motivation (X1) variable on productivity (Y2) is 0.210 while the indirect effect is 0.950, from these results it can be seen that the indirect effect of the work motivation variable (X1) on productivity (Y2) is more bigger than he direct influence. This states that good work motivation will have an impact on employee performance which can later affect work productivity Picture 2. Analysis using a direct model The direct effect of Job Training (X2) on Productivity (Y2) is 0.180 while the indirect effect is 0.460, from these results it can be seen that the direct effect of Leadership Style on Productivity is smaller than the indirect effect (through the employee performance intervening variable). This states that with a good Leadership Style it will have an impact on employee performance which can later affect work productivity.
The direct effect of the Leadership Style variable (X3) on Productivity (Y2) is 0.160 while the indirect effect is 0.370, from these results it can be seen that the direct effect of Job Training on Productivity is smaller than the indirect effect (through the employee performance intervening variable).
This states that interesting job training will have an impact on employee performance which in turn can affect work productivity.