I solemnly swear … an HR data and analytics manifesto

In the spirit of the election season, I would like to propose a type of nonpartisan party platform for candidates for analytical leader or officer positions within organizations. And even if you aren’t in charge of analytics, these are good principles to live by as someone who has to wrestle with what the analyses in your organization really mean and how best to use them – or ignore them.

1. Diagnosis before action. There should be the equivalent of a Hippocratic oath for people-related data. If doctors have to swear to first do no harm, can’t we ask the same of our data analysts and business leaders? The equivalent oath would go something like this. “I promise to take no actions based on any piece of people-related data without first doing a careful diagnosis of the drivers of individual and organizational behavior, including team dynamics.” Anyone caught violating this part of the manifesto will have to attend a daylong seminar on applying multivariate statistical analyses to social science research questions, which is a much more painful choice … trust me.
2. Benchmarks belong on the bench, not leading the team. Never make a decision based on only benchmark data. Benchmark data at best can inform the beginning of your inquiry, not the end. If you do a complete systems diagnostic and identify specific metrics that have strategic value, then you can scorecard them and use them to guide decision making. But never populate scorecards with any piece of data just because it seems like a good idea. It is better to have a scorecard with no HR data so you don’t send people off on a useless exercise of “solving” something that isn’t actually a problem.
3. Strategic analytics is a team sport. It’s not the Montagues vs. the Capulets, the Sharks vs. the Jets, or even Team Edward vs. Team Jacob. We’re all in this together – the business and HR – and have to coordinate our analytics for maximum actionable insights. Compartmentalization leads to subpar analysis at best, and to dead ends at worst. We need to all learn from Alice and avoid the analytics equivalent of Wonderland’s rabbit holes and paths that lead to nowhere useful.
4. Big data leads to little insights. Models and causal analysis are what you need to get the right insights, not correlational analysis on unstructured data. We already have almost all the models we need to map human behavior and motivation in organizations. Big data can help a lot on the consumer side of the equation, but not the employee side. Awhile back HR was accused of being “big hat, no cattle.” Using big data tools that are best applied to unstructured data, not the data we have on employees, is the analytics equivalent: “big tools, no insights.”
5. Big tech is no better. The brand new computer and information systems the consultants are selling today promise to be so versatile they will clean the house, get your kids tucked into bed at night, and pick up your dry cleaning. Collecting the bulk of your data in one place, with fancy visualization and analysis routines certainly can make your analytics tasks easier. Yet they are no substitute for knowing how to build and test causal models. Leaving those decisions to the computer and an untrained operator is like putting heavy artillery in the hands of children. Not a good idea unless your idea of fun is massive collateral damage.
6. It’s the package that matters, not the individual pieces. Whether it’s compensation, work-life balance, supervisor support or any of a number of other job aspects, it’s the total package that drives employee engagement and retention, not just any one element. You usually have to see people react to the job bundle as a whole to know what’s working and what needs improvement. In almost every other part of the manifesto, to do the right thing you usually want to avoid going directly to fancy statistical models. In this case, in contrast, there often is no substitute for multivariate models that put the factors in question through a statistical horse race to determine what matters and what can be ignored.
7. Continuous improvement is not strategy execution. Most of the day-to-day work of the business and HR focuses on what’s needed to keep the lights by incrementally improving what we already tend to do well. Improved strategy execution, in contrast, usually requires a paradigm shift: doing things in a fundamentally different way. The mindsets and analyses needed for continuous improvement are often quite different from what you need to close the gaps in strategy execution. That’s why HR perfect processes often are the enemy of what’s really needed in the business.
8. You can have too much of a good thing. It doesn’t matter whether it’s safety, retention, employee engagement, quality, go to market system efficiency, customer service, or innovation: there is always an economic limit to what you should strive for. There are so many competing business priorities and ways we can better engage our people in the work we do. Not a single one is so important that everything else should be sacrificed to improve it: you can’t get safety incidents down to zero forever without making people so cautious they stop working effectively; turnover that’s too low cuts off new blood from coming to work for you; quality standards that are too high sacrifice profits – same thing with customer service; innovation without regard for execution also won’t make any money; and so on. In each case we both have to tradeoff what’s best for the bottom line, and recognize that there is a balance to be achieved across all strategic objectives.
9. You need competitive advantage, not ROI. In the long run building and sustaining competitive advantage is the best way to make money and reap high ROI. But in the short run competitive advantage and ROI are almost always in direct conflict: the best way to improve short term cash flow, and thus ROI, is to stop investing in the long-run capabilities you need to preserve and build your market share. This is not a theoretical idea but a cold hard fact. So the next time you find yourself on the defensive, arguing with finance over an expenditure, look for the longer-term strategic benefit the investment provides in terms of competitive advantage. If you can show it, you will win the day and get finance to back down. If not, well maybe you should reconsider the rationale behind all that frivolous spending.
10. Employee engagement does not cause performance. Truly disengaged employees can sink your business results. But that doesn’t mean that measured employee engagement statistically and accurately predicts variation in business performance. It simply does not, despite the claims of Gallup and others. I’ve provided the details on this argument in my other posts and in my books. One more reminder never hurts. The bottom line is that you can use measured engagement to see where there are local management problems, and those issues often need to be addressed. But you can’t track employee engagement scores across the enterprise and push for improvements everywhere as a reliable way to improve business performance. That simply won’t work in the vast majority of cases.

This post is drawn from my books Strategic Analytics: Advancing Strategy Execution and Organizational Effectiveness and Employee Surveys That Work: Improving Design, Use and Organizational Impact.