Founder and Deputy Chair of orgvue. Rupert Morrison is an entrepreneur, economist, and visionary in data-driven organizational management.
getty
At some point today, it's likely you'll come face to face with a robot. Whether it’s a retailer chatbot or some form of automation when you log into your banking app, AI lives among us. But does that mean we should trust it to run our business?
Despite heavy investments, a thriving AI industry should be approached with caution. When used to inform business decision-making, blindly relying on AI is lazy. So while everyone is shouting AI from the rooftops and millions are being invested, nobody seems to know the value that’s being driven. So, what can businesses do to demystify their approach to AI and get it right?
Don’t be blinded by the lights.
Business leaders already know that having actionable data to inform decision-making is essential. But what if you’re not measuring all that matters? This comes down to the concept that you can only manage what you measure. So by definition, if you don’t measure it, you don’t manage it.
The McNamara fallacy is a good example of this. During the Vietnam War, decisions were made based on quantitative metrics that were easily obtainable and measurable, while other metrics were ignored. Body count was viewed as the sole indicator of success, rather than one piece of a much bigger, more complex puzzle. Of course, we can see now that this is folly. Body count was an easily obtainable and measurable metric. But it’s one that ignored other more important—albeit more difficult to obtain—data points. This happens in business all the time. Organizations opt for measures that are easy, not those that are the best indicators of long-term success.
Take many current productivity measures as an example. They’re easy to measure but are short-lived and unable to answer important longer-term questions. For example, given strategy, are employees spending enough time on the activities that will materially increase your chances of hitting long-term goals? Do you know who’s working on what? If the answer is no or you’re unsure, the odds are that your business is only measuring lagging indicators, such as current productivity.
Much more important to understand is how effective your organization and your employees are at executing organizational strategy over the long term. So if the strategy is building new products or entering new markets, it’s vital to prioritize the work that maximizes for tomorrow, or in other words leading indicators, which might potentially impact your short-term productivity. If, on the other hand, the strategy is to sell what you have now, it’s crucial to prioritize the work that maximizes now, so productivity in the present has a bigger role to play.
The point is, if you don’t have an accurate view of both lagging and leading indicators, you’re not able to see the full picture. That’s why leading indicators and those that are more difficult to measure are often disregarded because it’s assumed, implicitly, if not explicitly, that they’re unimportant or don’t exist. This statistical blindness is tantamount to burying your head in the sand as it fails to optimize beyond the quarter and financial year-end, preventing businesses from planning for the right long-term strategic goals.
Crave the context.
The most successful business leaders understand that data-driven decision-making isn’t just about interpreting individual datasets in isolation. They crave answers to the right questions. How do these factors affect my workforce? What’s the impact on the skills and competencies needed to deliver? What do the best- and worst-case outlooks look like?
Building the right foundations to interpret and interrogate your data is therefore vital. Without a structured data foundation, storytelling gets lost along the way. Leaders become cognitively overloaded, leading them to misinterpret the data or draw the wrong conclusions.
This is where the difference between correlation and causation is critical. Take the old tongue-in-cheek example of data showing shark attacks increasing in line with ice cream sales. Blindly, you might conclude that increasing ice cream sales cause more shark attacks. But correlation does not mean causation. If you apply this logic to a business context, it creates an alarming reality: Organizations are too often making decisions based on correlation.
This all comes down to some key fundamentals:
1. Having the right data to understand the workforce, the tasks and the skills required to deliver on strategic business goals.
2. Agreeing on the definitions—if HR and finance aren’t speaking the same language and measuring different data points on a crucial success metric, any conclusions drawn from the data will likely be erroneous.
3. Visualizing the right story. How often is a report presented without a "so what?" It leaves too much cognitive effort for the reader and fails to contextualize.
Level up your literacy.
It’s no longer acceptable for business leaders to be happy-go-lucky with their data. Put in the wrong hands or interpreted in the wrong way, data can drive wrong or misplaced decision-making. That’s why leveling up the workforce and eradicating statistical illiteracy is crucial. In the same way that poor grammar is frowned upon in a corporate setting, poor math shouldn’t be left to pass through unchecked.
Ultimately, if a data-driven approach is at the core of an organization’s decision-making (as it should be), employers have a duty to provide the tools, training and upskilling required for the business to draw reasonable conclusions and avoid the pitfalls of narrow-minded, statistically blind data analysis.
There’s no doubt that AI and ML will continue to dominate the discussion concerning their impact on business transformation, but there’s so much more that needs to be done to drive genuine value. It’s simple. Don’t be artificial with your intelligence.
Forbes Human Resources Council is an invitation-only organization for HR executives across all industries. Do I qualify?
