Often when I talk to companies about analytics they end up just talking about building dashboards. This is not a small thing because dashboards are expensive and can easily eat up most of your analytics resources. It’s also not a simple thing because dashboards are steeped in confusion and stirred with fantasy—this creates multiple paths to failure.
The confusion begins because there are multiple uses of dashboards, here are three:
- Dashboards can be valuable in operational settings such as a factory where real-time information allows a production supervisor to tweak machines hour-by-hour to optimize productivity. There are few such situations in HR because there are few things we tweak in real-time to improve productivity. This is the first hint of fantasy, the idea that somehow a dashboard will allow real-time tweaking of humans. Pursue this fantasy and you’ll end up disappointing every stakeholder in the process.
- Dashboards can be somewhat valuable as a kind of enhanced report. Reports can be useful. It’s good to be able to look up headcount from time to time. Will better HR reports deliver fresh competitive advantage? No. Pursue this path and you’ll end up with an unexciting return from what was meant to be a transformational move into analytics.
- Dashboards can do some real analytics, albeit of a rather narrow kind. Whereas a paper report would only give us a few numbers about turnover, with a dashboard we can potentially drill down to find where turnover is highest and perhaps even find the cause. It’s narrow because the only data you are looking at is what happens to be in your HR systems, and it’s narrow because the sophistication of analysis is limited. But it’s not a bad thing. The only question is whether people really will be using this capability and how often. The risk here is that this good, but narrow capability will rarely be used, leaving HR scrambling to justify their big investment.
Dashboards are popular because they are tangible and keep everyone busy.
Those three somewhat related uses of dashboards are confusing because they all have some value, but are still likely to lead to disappointment or even out and out failure. We are even more likely to end up with failure if we vaguely pursue all three paths at once not really knowing which one we are on.
It gets worse after that because there is a whole other motive for using dashboards that has little to do with the previous three: command and control.
Dashboards can be a vehicle for reporting KPIs whose purpose is primarily to stop people slacking off. Many managers are concerned that their people are not working hard or not working on the right things. They harbour the hope that if they can just get the right KPIs they can hold people accountable. This is confusing because it’s partly correct and partly fantasy. Pursue this path and you’ll get the usual results of a KPI project. You won’t get analytics, which is a problem if the budget for dashboards was justified as an analytics project.
Why do we pursue these many paths to failure?
Dashboards are popular because they are tangible and keep everyone busy. They’re pretty and they’re fun to create. Everyone can support them because they have some vague notion that that they will help with one of the four goals. This creates just the right mix of good reason, confusion, and fantasy to get organizational support.
If you think this through you’ll keep circling back to the same point: dashboards can be kind of good for various reasons, but in practice we will probably blow a lot of time and money on this and at the end have nothing especially valuable to show for our efforts.
What should be we do instead
Given that there are some merits to dashboards, let’s not give up on them; just don’t let them lead the process. There are really two main parts to analytics: long-term advanced analytics done by data scientists and mid-level analytics done by everyday HR people who have basic quantitative skills.
You’ll succeed by putting the emphasis on the mid-level. You should be seeking out a constant stream of mid-level wins. This is the reliable means of making progress and ensuring ongoing support for the overall analytics work, including the advanced analytics work. It will also reveal where you really do need dashboards. Instead of doing dashboards simply because its seems vaguely acceptable, you’ll be doing them to meet specific needs that have been clearly revealed from the ongoing mid-level analytics wins.
I’m beginning to see that people analytics is already setting itself up for a fall because it seeks to make big investments that won’t pay off for a long time. After the honeymoon period, leaders will look at the money spent and the results delivered and suggest that the budget for people analytics be quietly shifted elsewhere.
Dashboards are dangerous because the uses are confusing and the higher hopes are pure fantasy. Go for mid-level analytics wins and from that foundation, you’ll actually transform HR.