Does your data need a human-in-the-loop?

The human-in-the-loop model is a way in which people and machines work together to create, refine and improve things. Think of the way machines and people work in harmony - in a physical environment, that might be a aeroplane simulator. In computing it can refer to a way to create machine learning algorithms for things such as sentiment analysis. In medicine, software and doctors work together to create a more accurate diagnosis than either can do alone. So what does this have to do with business data?

Most teams have a wealth of data, and often tools and systems to order it. However, you need a human-in-the-loop to derive the maximum value from it. Without one, you just have a load of information that doesn’t have any context, force action or create wisdom.

Take for example the engagement survey. HR teams are usually responsible for selecting the platform, setting up the system, the internal commutation, then analysing the results and owning the subsequent action plan. Or think about those companies that have or are planning to implement a new HR system. It is a complex and costly process, both in terms of time and money.

The common thread that HR teams are investing in new and shiny software platforms. Yes, it does make the graphs look flasher, and potentially makes you look more up-top-date. It can also cut down on administrative time and generate more business intelligence. However, to what extent are these projects really transforming the effectiveness of your specific function and more importantly the performance of their organisation?

How many projects come with a business case that is signed off beforehand or evaluated afterwards? The answer depends on the human-in-the-loop. We believe it is a precision skill and there are certain characteristics that will make a successful human-in-the-loop:

  • Data literate - ability to process and understand numbers.

  • Insightful - someone who can to take data and turn it into something meaningful. Then accurately and with a degree of wisdom draw conclusions.

  • Ability to deal with ambiguity - the world is a complex place, rarely is there a simple answer, a full set of data, or a stable situation to base your thinking on.

  • Storytellers - how can you take the insights and craft them into a compelling case?

  • Impact and influence at the highest level - insights without action is just overhead. You need to be able to make things happen.

Organisations usually run with the smallest data teams possible. Most mid-sized and smaller businesses simply don't have much capacity this type of resource. Where organisations do have analysts, they are often regarded as reporting functions. This means they churn the data which goes elsewhere for other people to use. It fulfils the first set of characteristics, but not the others.

If you want to become proactive rather than reactive, your human-in-the-loop and their capabilities become essential. Some disciples have mastered this art. Take marketing: typically comfortable taking data, generating insights, building business cases, following through to action and analysis. 10 years ago, it was playing catch-up in this area, now it is one of the most data-led disciplines.

Furthermore, it is happy to invest in the external resource where required. Try and find a marketing director without a little black book of agencies there to help them solve specific problems. HR on the other hand has been notoriously slow at getting external help. Yes, HR folk are largely generous and willing to share best practice, but it is still unusual for an HR director to look outside their business for expertise. I have lost count of the number of engagement programmes I have led with clients who don’t even have the most basic data analysis capabilities internally. That’s only a problem if they are unwilling to get help to sort it out. With the abundance of HR data, and the structural importance of an effective HR department, doing this right now is of paramount importance. So whether you plan on getting internal or external resource, make sure you get it. Now is the time for HR to get their human-in-the-loop sorted and transform how they operate for the better.