科研管理 ›› 2020, Vol. 41 ›› Issue (10): 238-247.

• 论文 • 上一篇    下一篇

基于行为证据推理的企业研发人员绩效测量方法研究

郑毅1,2,徐芳1,2,牛华勇3   

  1. 1中国科学院科技战略咨询研究院,北京100190;
    2中国科学院大学公共政策与管理学院,北京100049;
    3北京外国语大学 国际商学院,北京100089

  • 收稿日期:2017-11-01 修回日期:2018-04-19 出版日期:2020-10-20 发布日期:2020-10-19
  • 通讯作者: 牛华勇
  • 基金资助:
     

A research on the innovative performance measuring method based on behavioral evidence reasoning

 Zheng Yi1,2, Xu Fang1,2, Niu Huayong3   

  1.  1. Institutes of Science and Development, Chinese Academy of Sciences, Beijing 100190, China; 
    2. School of Public Policy and Management, University of Chinese Academy of Sciences, Beijing 100049, China; 
    3. International Business School, Beijing Foreign Studies University, Beijing 100089, China
  • Received:2017-11-01 Revised:2018-04-19 Online:2020-10-20 Published:2020-10-19
  • Supported by:
     

摘要: 由于研发人员工作内容的复杂和产出的不确定,如何对其绩效进行科学、准确的测量和管理一直以来困扰着研发管理者。较之于社会标签、工作结果等表层绩效信息,对研发人员日常工作行为的分析研究可以帮助我们更好的理解其深层次绩效特征,进而开展有针对性的绩效管理。因此本文提出了一种基于行为证据推理的研发人员绩效测量方法。这一方法首先确定与研发人员绩效相关的行为证据,进而采集行为数据并利用证据推理规则构建起研发人员“行为-绩效”间的推理关系,进而根据推理结果对研发人员绩效表现进行预测和管理。通过实证研究,上述方法可以有效判断个体研发人员的绩效水平,并帮助研发管理者开展有针对性的绩效反馈、提升。

 

关键词: 研发绩效, 绩效管理, 证据推理规制, 行为证据

Abstract:

Today, the activities relevant to performance management (PM) can be found in every corner of business, and its importance could be described by a famous business motto that whether a company measures its workforce in hundreds or thousands, its success relies solely on performance. Despite its importance as an enabler of successful business, some issues and shortcomings still exist in the performance management research and its implementation, and the PM dilemma in the R&D management process is widely agreed as the most representative one.
Due to enormous complexities and uncertainties existing in the process and output of R&D workflows, it is always a major challenge for R&D managers to measuring and managing their subordinates′ performance accurately and timely. Traditional PM frameworks and R&D management tools are largely incapable to guide the R&D managers out of the dilemma. The emergence of behavioural PM indicators brings a new approach to measuring and managing the R&D personnel′s performance in a more convenient and convincing way since these indicators reveal more profound performance characteristics as compared with the traditional ones. Therefore, a behavioural evidence-based PM way for the R&D personnel is explored in this article. The main idea of the innovation PM approach for the R&D personnel is to first select a competency model to determine the competency factors to be measured. Then possible data sources for the measurement of the factors (these can be identified via discussion with key stakeholders) will be determined. Finally, suitable models from the artificial intelligence field will be determined to produce proxy assessment of the factors, and then competency and performance will be measured from the data sources. 
Specifically, there are four main steps to apply the evidence-based PM approach:
The first step is to establish a competency model based on the strategies & objectives of the organisation (or simply apply an existing model). The competency factors reflect the characteristics of R&D staff that affect job performance. In practice, CM selection is one of the most important issues in carrying out CM-based management. Practitioners should review a range of available CMs and compare them in terms of feasibility within specific managerial contexts. Under some circumstances, users can directly employ the selected model. However, if the company wants to develop a tailored competency model from a general one, it will need to further refine the competency factors under the competency dimensions based on their specific managerial needs. Since the PAKS model includes most essential competency factors for R&D staff, we will use this model in the case study of this paper for illustration. However, it should be bear in mind that the framework is also compatible with most of the existing competency models, and users could replace PAKS with other preferred models.
The second step is to search for supporting evidence for the competency factors. The supporting evidence is crucial in carrying out proxy assessment for the competency factors, since their qualities affect the results, and hence affect the management directly. Generally speaking, there are several common sources of evidence: 1) Past assessments: the information on past assessments reflects the performance and capability of an R&D staff member; such information is often supplied by the HRM department. 2) Biographical information: this is the background information on an R&D staff member, including gender, age, educational background, past academic performance, etc. 3) Online (Intranet and Internet) behaviour information: since most of the daily activities of R&D staff are carried out on computers, their online behaviour is one of the most important forms of evidence reflecting their characteristics and competencies. 4) Subjective assessment information: this is competency-related evidence generated from the staff member and his/her line managers and peers through subjective means such as 360-degree evaluation, self-evaluation, questionnaires and scales.
In this step, the key stakeholders should have in depth discussions about how to select and link these data sources with the competency factors. Meanwhile, according to the availability of the evidence sources, the competency factors should be altered again – those factors lacking evidentiary support should be eliminated. It should be pointed out that here we only consider the availability of the evidence in a general managerial context and adopt some of the most common evidence sources (the first four sources mentioned above). The company can further enrich the factors if they have better hardware. 
In the third step, the specific MCDM method and the corresponding measurement should be selected and applied. In the evidence-based PM approach, multiple MCDM methods can be employed to produce the proxy assessment of competency and then performance. The evidential reasoning rule (ER-rule) method was adopted in this research for the following reasons: 1) The output of the ER method is a likelihood matrix composed of multiple managerial events and corresponding probabilities. Therefore, this result can provide more information for elaboration management and managerial diagnosis compared to deterministic-output methods; 2) The ER method is also highly compatible with different data forms; both the sequencing data and continuous data can be inputted for analysis; 3) The probability-based logic of the ER method matches with the nature of managerial practices more so than numerical fitting methods since alternative options always exist in business operations. It is also possible to utilize other types of MCDM methods in this approach depending on uses′ preferences. 
The fourth step is to apply the results of the assessment to the PM of the R&D unit. The results of the assessment can be applied to the R&D PM in two layers: 1) The predictions of individual performance can help R&D managers carry out PM at the individual and departmental levels. They may intervene in both levels of R&D operations before a poor-performance situation occurs. 2) The results from the competency factor assessments can assist R&D managers in training their subordinates and assigning jobs. For instance, R&D managers can design a specific training program for a R&D staff member based on his/her competency scores; or they can create more detailed performance plans for the employee according to his/her competencies.
In the last part of this article, a case study carried out in a Chinese hi-tech enterprise is addressed for presenting detailed procedures about how to implement and to obtain managerial benefits from this innovative tool. Based on the result of the case study, it can be concluded that the behavioral evidences-based PM approach can help R&D managers to measure and infer their staff′s performance continually with acceptable time and resource cost.

Key words:

中图分类号: