As the days are getting shorter and people spend more time inside, department stores across the UK are displaying their baking accessories. When I was thinking about talent analytics while wandering through one of those shops, I wondered what baking a cake and talent analytics have in common. I was surprised to find the similarities! 

PLAN your objective & approach

What is the occasion, who is it for, what is the timeframe? Just like your objective and recipe would be different for a wedding cake compared to one for a Halloween party, the objective of talent analytics is not always the same. I want to be clear on what I am trying to achieve with the analysis, and find it helpful to write down the question I want to answer. For example:

  • ‘Which channels are most efficient to attract middle managers who lead high performing teams with low attrition rates and high engagement levels?’
  •  ‘What are the common traits among mentors who have helped their mentees secure a promotion within three years of starting the mentoring relationship?’
  • ‘Which factors predict how much revenue a sales manager generates?’

Once I am clear on the question, I then look at defining hypotheses, assessing which data could be relevant, and which statistical methods are applicable in finding an answer. Researching similar cases and exchanging thoughts and problems in the real and in the virtual world help to get another perspective.

GATHER your ‘ingredients’

Just like the ingredients determine the quality of a cake, having the right data determines the quality of the analysis. I often find that talent or HR analytics is limited to data from internal HR and talent systems. The most valuable insights are generated where that data can be linked to other internal and external data, such as financial measures, risk measures, or data from social media. One could be identifying the sales representatives that bring in the most profitable and loyal customers, and looking at whether there are any predictors in their assessment scores, training records, or social media usage. You could also identify predictors of leaving the company, check if any of your identified successful sales representatives are likely to leave, and take countermeasures to try to keep them.

Put it in the machine and OBSERVE

You put the cake in the oven, or the data into your statistical analysis, and then you wait and observe what happens. Just like you use different senses to observe when your cake is ready, you use different statistical indicators to assess the quality of the analysis in terms of reliability and validity.
When the technology has done its job, the part I find most exciting begins: Deriving insights from your analysis and translating those into recommendations. This is when I would look at the results, with my original question in mind, and pull out what story the data tells in answer to that question. Stepping even further away from the detail, I would look at the answer, ask myself ‘So what?’, and come up with recommendations from these insights.

PRESENT your results in an attractive way

The first bite is with the eye! The presentation determines if the product is used, whether that is a cake or the insights of your analysis. I find a pyramid communication approach usually works best, where you start with the key message and then diverge into more and more detail. I would start with the answer to the research question, complemented by a visual representation of the main findings. Then I would present the recommendations, visualise the next layers of detail, and include background information in an appendix.

SHARE the joys

Would you want to eat a whole cake by yourself? Why not share the results of your analysis? Depending on the insights and confidentiality, it might be appropriate to share at least some of your findings with other business units, your organisation’s leadership, all of your organisation, or to present them externally at conferences, in forums and think tanks, or through other social media channels. Sharing your approach and findings might spark some new ideas somewhere else.
I hope you can enjoy talent analytics as much as cake!