Upskilling the workforce to the digital age is one of the most critical priorities for implementing the highly desired People Analytics function. In this article, I will list eleven main points to take away for immediate action.
Literature often mentions HRBP’s as the target group that need learning new skills. However, surveys indicate that only 34% of HR professionals state that their company offered training in analytics. Engaging with HR is vital due to their proximity to senior management. HR professionals possess a unique understanding of what the organisation requires to prosper. Furthermore, they understand the business context. The entire organisation can benefit from new skills. HR is having a transformational impact on the business.
The question is now: Upskilling to learn what skills in particular? HRBP’s do a whole spectrum of different tasks. Still, only very few are suitable for leveraging in the digital age. It is easy to agree that organisations should point to harness the power of Artificial Intelligence and Machine Learning. But, in practical terms, how does it work?
1. Recognise problems suitable for digital improvements
First of all, HRBP’s should recognise highly repetitive and information-based tasks to prioritise. They shall look them under the magnifying glass. The more the tasks are specific to the business context, the more likely they generate positive returns on investments. There are low hanging fruits everywhere. Please subscribe to my mailing list below to receive the bonus content. You will receive a list of ideas on where and how to implement AI at the workplace. The most valuable insights are in predictive modelling.
2. Define well the problem
The first step in any AI implementation is to define well the problem. Therefore, HRBP’s should define the problem clearly, provide some figures on the frequency and time spent. HR problems are notoriously difficult to define. Does a process description exist? If not, create one as a priority. Have you tried to write a flowchart of the different process steps? How about utilising swimlane diagrams?
3. Collect data
The next step is to collect data on those tasks. Are the data already standardised? Have the date formats all written in the same form? Did you check for any missing values? How complete are the data? Are the data in an unstructured text format? How can you make the data easy machine-readable?
4. Clean the data
HR practitioners spend a considerable amount of efforts in cleaning data to get it ready for analysis. Still, there has been little research on how to make data cleaning as comfortable and practical as possible. Tidy datasets are easy to manipulate, model and visualise. Hadley Wickham defined tidy data in his path-setting article. HR practitioners need to arrange data sets so that each variable is a column and each observation (or case) is a row.
Indeed, tidying up messy datasets is a skill to learn. It is nice to hear that only a small set of tools are needed to deal with a wide range of untidy datasets.
5. If data are not available, create data
In other occasions, data are sometimes not obtainable. Therefore, one needs to collect data from scratch. However, it might be annoying for employees to fill out another questionnaire. Employees might feel they have been asked for feedback too many times. As expected, they become disengaged. Likewise, employees should be clear about what the advantages are for them. If possible, leaders must grant more control to individuals, so they can manage and even own their data.
There are also alternatives to surveys to generate data. For instance, it is possible to amend the process so that it creates electronic records at certain moments.
6. Create a fundamental model of relationships
Identify some first models based on your intuition. If you fiddle with parameters and the feedback cycle is short, you get the hang of it. Eventually, you might also succeed in thinking of a possible algorithm. Similar to, when this happens, then it means this. In other words, write down how the different variables hang together. Afterwards, you can call in the professionals in algorithms to verify your intuitions against the data. Remain sceptical of results. Design experiments that make it hard to fool yourself.
7. Encourage knowledge and data sharing within the organisation
Leaders need to encourage staff to make use of all the free resources available online: sites, blogs, books, conferences and forums. Free resources are available from Google re:Work to HR Open Source. For instance, these sites are full of templates, tools, frameworks, and resources on modern HR practices. Life-long learning is costly to private organisations. Community learning is the new mantra. Websites similar to mine are awash with helpful ideas.
8. Implement quick win AI projects and produce a series of skunkworks
For this purpose, I am listing seven quick wins in AI projects. You will find them in the bonus content below. It will be available to those who subscribe to my mailing list. A skunkworks project is a project developed by a relatively small and loosely structured group of people. They research and develop a plan primarily for the sake of radical innovation.
9. Define and communicate an Ethics Code of Conduct for data and analytics
Technology is a double-edged sword. At times, it has unintended consequences that impact individuals and society. It can itself help to fix these downsides. For instance, by addressing bias, by assisting workers in controlling and by sharing their data securely. 92% of employees are open to the collection of workforce data about them. In truth, only if it provides them with personal benefit. Ethics is the foundation of good people analytics.
10. Provide sponsorship for more significant people analytics projects
Granted that, Human Resources leaders need to be actively involved in people analytics projects and in upskilling the workforce to the digital age. Most importantly, they should support key staff members who “get the job done”. Despite hard work, results might not come at the first attempt. Eventually, it might be necessary to start all over again. For instance, improve data collection, re-do the data cleaning. Finally, repeat the entire process several times and results will improve. However, one should always remember Goodhart’s law summarised nicely by Marilyn Strathern: “When a metric becomes a target, it ceases to become a good measure.” This motto is particularly true for humans adapting to new company policies rather quickly. Thus making the whole people analytics intervention redundant at one point.
11. Presenting results
The results of complex thoughts are meaningless if they have no impact on the business. Showing results is good practice. It provides learnings one can build on next time. Fortunately, here comes Jason Brownlee with an excellent way of presenting results:
- Firstly, define the environment in which the problem exists. Additionally, set up the motivation for the research question.
- Secondly, concisely describe the problem as a question that you went out and answered.
- Thirdly, concisely describe the solution as an answer to the question you posed in the previous section. Be specific.
- Likewise, make a bulleted list of the discoveries you made along the way that interests the audience. They may be discoveries in the data, methods that did or did not work. The model performance benefits achieved along your journey.
- Moreover, consider where the model does not work or questions that the model does not answer. Do not shy away from these questions. Defining where the model excels is more helpful if you can determine where it does not shine.
- In conclusion, revisit the “why”, research question. Revisit what you discovered in a concise format easy to remember, repeat for yourself and others.
Final thoughts
In conclusion, upskilling the workforce to the digital age requires a cultural change very much. In fact, “to play well at tennis, one needs to have the right partners, who are playing with you”. This is my old boss’ adage for similar circumstances.
Needless to say that upskilling will fail for the same reasons, many other projects are failing:
- At first, I shall mention the lack of Top Management Support and Leadership.
- Secondly, sometimes it is due to lack of competences or sensibility to the problems. The primary reason for failure is always caused by the human variable.
- Above all, errors in selecting the right initial project. They must be chosen based on good sense and the ROI indicator.
- Lack of training. Machine learning and AI are robust. However, employees need training on these modern tools. There needs a more scientific attitude towards problem-solving in HR.
- In particular, staff development should go hand-in-hand with a cultural change.
- People analytics projects are likely to fail, in particular, if not directly linked to the organisational strategy.
- Moreover, HR should work only on things that are important to the business.
- Finally, there may be many reasons for a project to fail. The important thing here is to identify, dispose of the problems and to make sure these do not happen again.
What is your plan for upskilling your workforce to the digital age of People Analytics? Please leave a comment and share.
Recommended reading: 11 tips for #upskilling your #workforce in the digital age of #PeopleAnalytics Click To Tweet