Skip to content

People Analytics: Keep it Simple and Segment

Published by orgvue


Peter Howes speaking at the Tucana People Analytics conference

Today I was at the second day of Tucana’s People Analytics conference and attended a great session from Peter Howes. Howes is one of the original founders of InfoHRM, which was bought by Success Factors and then SAP. It was fascinating to hear his 5 most important analytics recommendations:

1. Segment – don’t use aggregated data. Sales people data is very different by high performers vs low performers. Segment by personal characteristics and segment by role characteristics.

2. Integrate people data with business data, such as revenues, expenses and CSAT.

3. Integrate people data from multiple sources, such as engagement, absence and training.

4. Data quality is key – but you can’t wait for it to be perfect. You get data clean by using it.

5. Tell the story – to have an impact, you have to communicate the story in business terms.

The danger of aggregated data is worth repeating: Howes described these as ‘generally dangerous’. He explained that it is vital to drill down: “We did a study of 18 of our clients in Australia. We looked at 7 benchmarks – including voluntary turnover by organisation unit. We excluded the 1st and 10th decile. And the range was 0% – 35% while the average was 16%. But the turnover rates in one company were much higher for their high performers in critical roles. It was meaningless to compare that company’s total turnover with the average of other companies’ total turnover.”

Peter Howes 2

Interestingly, Howes suggested that Success Factors is now exploring detailed transactional benchmarks, such as high performing professionals in their 3rd year of position tenure, as a way of overcoming the aggregation issue.

Being sensitive to segments, and valuing basic steps in analytics aligns to something Max Blumberg said on Day 1 of the conference: while people describe the classic analytics journey from description, through correlation, to prediction, many organisations spend most of their time and get most of their value from the first stages.

As Howes commented, the majority of this work is applying basic maths (‘Grade 4 maths – addition, subtraction, division and multiplication’) and getting good simple descriptions of the organisation. A rough guess of activity mix and value achieved from Max & Peter’s discussion… does this look like the organisation effort in your business? 50% – Description 30% – Correlation 5% 15% – Prescription

This was Howes’ top quote from the session: “I’d like every HR practitioner to spend 5-10% of their time doing HR analytics.”

Want to find out more? Come visit the OrgVue stand at the Tucana People Analytics conference.

OrgVuePhotographs courtesy of Andreas Grieger.