There was an era when Analytics was construed as largely Dashboarding and Reporting. If you had smart people that could play with some bit of available data and categorize it as Red, Green or Yellow and create status reports, that solved for a great deal of data requirements of a business. Businesses, then, were only keen on understanding their performance vs. key metrics. After all didn’t Peter Drucker say, “What gets measured gets managed!”
This was the analytics of the ‘80’s. Come 2000, slowly, the need shifted from simple dashboarding to more complex reporting and enhanced data visualization tools. The need for Business Intelligence and Data Mining experts was huge. Analytics professionals that could crunch large data sets and generate meaningful insights or could use some of the flashy visualization tools to create attractive dashboards were very much in demand. Those were the days of descriptive analytics where the business need was limited to the use of historical data to generate actionable insights resulting in smarter decisions and better business outcomes.
Today, the world has moved far ahead in the analytical journey where prescriptive analytics is very much in vogue. However, like any other business analytics stream, talent analytics too has been part of the same evolution chain. Today, businesses are not comfortable sitting on a heap of attrition reports that gives them data analysis on who left, when and why. Businesses now want to predict the future. For example, given the historical data, they would like to know which class of employees has a greater propensity to attrite and when? Such information is of great value to any business as retention efforts can then be channelized towards a targeted pool rather than an intuitive effort on a larger employee base that usually has no recognizable outcome or benefit. Data Scientists can make that happen. Talent analytics can help the people function move beyond the past and the present and help predict the future with an acceptable level of accuracy. But, if one assumes that a data scientist alone can solve for world hunger, such expectation is unrealistic. Before we embark on this long journey, it is important to debunk some of the myths and understand the golden rules of Talent analytics.
A Data Scientist can solve every business problem: Isn’t this equivalent to saying that a surgeon can perform any surgery? A cardiac surgeon cannot perform an orthopedic surgery or vice-versa. You need domain expertise to get accurate results. Therefore, in a talent analytics project, it is not only important to have a team of highly capable statisticians, but also domain experts who know the function, the systems and the data very well. To explain this simply, every analytics project has four layers to it:
- Correctly identifying the business problem
- Converting the business problem into a statistical problem
- Finding a statistical solution to the problem
- Translating the statistical solution into a business solution/outcome
The outer two layers need Domain experts, while the inner two layers require a data scientist. It is, therefore, needless to say that without the domain experts any project would be dead long before it gained any momentum. A successful talent analytics project will have a good blend of domain experts who bring in the functional knowledge and data scientists who bring in the dimensions of mathematics and science to the art of decision making.
Get it right the first time: First Time Right (FTR) is a concept most Operations Teams are familiar with. However, that concept will not work in any analytics project given the nature, complexity and the elements of ambiguity involved. In talent analytics, when predictive models are created, the first iteration is usually far from actual reality. The iterative process is longish and painful at times. Patience and persistence is required all along the project timeline. It is important to fail fast and early, continuously learn and evolve throughout the project than invest all resources to get to a stage of perfection in the very first iteration. That would be a foolish thing to do!
The coolest tool will give you the best results: If this was to be believed, why wouldn’t the world judge a book by its cover? A great analytics organization will first understand the business problem thoroughly, outline what business outcomes are critical and focus on it. Everything else is a distraction which needs to be weeded out early on for the project to stick to its course. What is important is to stay agile in the journey. One cannot work in a silo and risk the project getting extinct due to external factors such as digital revolution. Therefore, it is more important to have a good pipeline of vendor-partners in this journey – folks that will bring in data science, technology and domain expertise. Having subject matter experts on board, who will lend their intelligence as we develop the process engine by telling us what variables we have missed and why our hypothesis might fail, is more important than a flashy tool which might just take garbage in and send garbage out.
You have to outsource every talent analytics project: Nah…not really! If your organization has deep expertise in analytics, you could even Crowdsource ….that’s the key to solving the talent pipeline issues for longish analytics projects. If you have fungible internal talent, they could easily be deployed on these projects without the project getting stuck for resourcing crunch. The other advantage with internal teams working on one common problem is it builds competition amongst teams. It could also bring out multiple solutions to one business problem.
Talent analytics project starts with data and ends with insights: A great analytics organization will know that this is not the real purpose of any analytics project. A good project would be deeply embedded with business. This means that the talent analytics project should embed itself into both upstream (understanding the data sources, system linkages and every single thing that feeds into the data which comes into the predictive analytics process engine) and downstream activities (consulting with users of data on the insights derived through the project and possible the business outcomes that can be achieved through it). The success of any project is ultimately measured by the impact created on the shop-floor. Therefore, it becomes imperative to walk that last mile and support the businesses in their decision-making process of how they could leverage these insights and benefit hugely out of the hefty investments that have been made in the talent analytics projects. In Genpact, the talent analytics journey has been about getting the HR data at the touch of a button for the decision makers. And, to get to that state, one has to go through the painful process of data cleansing, getting the system linkages corrected, understanding master databases and so on and so forth. It cannot just start with data and end at insights!
The future of Talent Analytics
While everyone will have their own version of the future state, one thing that will definitely make a huge impact is the way Artificial Intelligence could revolutionize the People function. We are already hearing about Chatbots doing the ‘Stay Interviews’ or even the initial screening of interview candidates. We are also envisaging an era of video interviews where real time analytics can be done on the candidate’s facial expression, tonal quality, body language and personal attributes which can enable the interviewer to uncover the real person behind the mask of an ideal interview candidate. Gamification has already started seeing success in the behavioral assessment of potential candidates replacing the online personality tests which are not only lengthy and cumbersome but also leave room for the candidate to tweak his responses. The results of gamification are more reliable and accurate as it becomes almost impossible for someone to lie while playing a game or in a simulation.
Therefore, the future of talent analytics is all about intelligent machines making real time decisions about people with a greater level of accuracy as compared to human intelligence — something the HR functions can hugely benefit from. AI could also reduce gauge issues making it a preferred talent analytics solution of the future.
Talent analytics holds great promise in changing the dynamics of HR organizations.
After all, who wouldn’t want to attract the right quality of hire, predict the employee performance and attrition and be able to estimate the talent costs with great accuracy?