Technology plays an ever increasing role in our resourcing strategy and operations. Looking at technology needs and options, we see that data science is more and more defining the possible benefits that technology can bring. But though data science brings significant added value to our world, it also comes with an important health warning.
The term ‘Data Science’ is often used as a catch all term. Artificial Intelligence (AI), Machine Learning, Big Data, Robotic Process Automation (RPA), Deep Learning etc. all mean different things. They are often incorrectly used interchangeably by providers hoping to impress the market with their latest technology offerings. But even though for all of them data is at the heart of the matter, they are each based upon different forms of data analysis. Additionally, data science tools and techniques cán deliver significant benefits as part of a technology or service provider, but are not benefits in their own right.
Considerations for Data Science in resourcing
So data science based solutions should absolutely be explored because of the vast benefits that these types of techniques can deliver to our resourcing processes and operations. Such solutions will quickly become essential in remaining competitive in an increasingly competitive labour market. However it is important to understand the potential pitfalls and take a holistic approach to implementing such tools.
Quality of data
Quality of the data is crucial. The greatest of innovative, data science based solutions will not deliver the expected benefits, if the data in our own systems is not of sufficient quality to correctly inform the clever science that is looking to deliver significant added value to our operations. We’ve all heard the expressions ‘rubbish in, rubbish out’, ‘if it’s not in the system it didn’t happen’ and many other ways of stating the obvious as to the quality of data in our various technologies. These flippant but well intentioned sayings, are becoming more and more important as time goes on.
Big and small data
Our operations systems often have relatively small data sets. To develop these tools and techniques big data is used. In developing the tools and techniques that are being proposed, technology providers often use millions, sometimes hundreds of millions of data points to ‘train’ their systems. The advantage of very large data sets is that the impact of any incorrect or outlying data points is minimised. The techniques and tools themselves also learn to recognise and ignore errors and outliers. This is much harder to achieve with the size of data sets we typically have within our own organisations.
As an example, if 9 records are stating the meaning of life is 42 and one record states its 24, you can make a case for the meaning of life being 40.2. However, if there are 999,999 records stating its 42 and 1 record of 24, you’d say it is 42 (unless you go to a silly number of decimal places).
Suppose you have invested in these new amazing technologies that use big data to ensure data accuracy. Does this mitigate against any issues with your own data? The answer is no, and the reason is context.
Another example; you have invested in a system that not only tells you the price of apples, but also the variations in price based upon the freshness of the apples, the location where we are buying the apples and the variety of the apples. So we look in our ATS/VMS and see that we are looking to buy Golden Delicious apples in Rotterdam, that are less than a week old. Our new system tells us exactly the price we need to pay. The problem is that the information in our ATS/VMS is wrong and we are actually looking to buy Red Delicious apples in Rotterdam that are less than a week old, so now the price we have is wrong.
When you know your data isn’t the best, should you defer investing in such technologies? The answer again is no. And the reason is that your competitors are both investing in such technologies and addressing their issues with data. The market for talent is becoming ever more competitive and therefore doing nothing is not an option.
So what do you do? You need to adopt data quality management.
TalentIn is happy to discuss what this means in practice, however the key considerations are;
- Prioritise – It is unrealistic to ensure every data item is correct. Prioritise those data items that matter.
- Incentivise – Think of creative ways to encourage your teams to focus on data quality.
- Review – A few simple reports from your system can form the basis of regular reviews of your data.
About the author
Matt Jessop is Associate Partner at TalentIn. With his expertise as experienced technology subject matter expert and innovator he helps organisations build and execute their workforce strategy,
Matt has a great passion for data and analytics and its role in driving productive business behaviours.
TalentIn has extensive national and international experience in developing and improving strategies for the recruitment of your permanent and temporary staff. We know how these programmes can be designed and implemented successfully. We advise, but can also provide practical support. Are you interested? Please contact us for an appointment without obligation via www.talentin.eu, email@example.com or +31103075422