Ireland the “goodest” country in the world? 5 common sense rules for data analysis

By Martin Duffy, SAS Ireland

So Ireland has picked up another glowing accolade for the national mantelpiece, this time for our positive contribution to humanity and the planet. That’s right, we’re the “goodest country in the world”. It’s a catchy headline but, I have to ask, what does it mean? I’m not alone in my cynicism; the author of the report said he’s had a deluge of emails from Irish people taking issue with his analysis of the data.

The fact is, good is a relative concept. We often, rightly, turn to data analysis to help us draw conclusions and make decisions. But there can be a tendency to completely overlook common sense in the interpretation. That’s fine for headline grabbing surveys; but, in business, we really need to define what ‘good’ is before we take action. A healthy dose of human experience and common sense should always be applied. So, here are my five common sense rules for making ‘good’ data-led decisions: 

1.      Analysis has a best before date

Think of data analysis as fresh produce: it can go off. In the business world, external factors may affect the validity of models (e.g. the weather), or actions taken since the analysis was done. For example, using a propensity to churn model to drive incentives will change customers’ behaviour and therefore the validity of the original analysis. So, before making decisions, always ask yourself: has the world changed since this analysis was done?

2.     The right decision can be: no decision

Just because a piece of analysis says you should do something, it doesn’t necessarily follow that you should. Analysis based on incomplete information may lead to ‘bad’ decisions. Or perhaps the time isn’t right. Don’t blindly make decisions based on the findings; always apply your own logic when interpreting results…which leads us to Rule 3.

3.     Know what you are agreeing to

Just like humans, analysis can have bias. For example, a model may tell you which customers are likely to leave, but that doesn’t necessarily mean these are the ones you should spend money and time trying to keep. For that, you’ll need information about past relationships and customer behaviours. Always question what the analysis isn’t telling you, and what assumptions have been made, before you make any decisions.

4.     Actions speak louder than words

Decisions are only worthwhile it they are tied to an action. So it’s important to think about your ability to execute. For example, if you decide to reduce churn with an incentive: how quickly can you execute, will it be in time, do you have the resources? My last blog gives a cautionary tale of a customer who failed to do this, and wasted a lot of time and money in the process. Remember: the analysis may be driving you towards and action, but you need to be aware on the constraints.

5.     Not all decisions are equal

Not every decision needs to be made using analysis, so prioritise the questions that warrant the effort. Decisions around the safety of your nuclear reactor are probably worth some serious analysis; choosing which colour branded pen to send customers less so. Just because you have a sledgehammer, you don’t need to use it, sometimes a spoon will do just as well.

A combination of analytics and these common sense rules is the secret to ‘good’ decisions and actions. For a master class in how to avoid making errors when interpreting statistics, I’d recommend How To Lie With Statistics by Darrell Huff. Or follow SAS Ireland on Twitter or LinkedIn for more thoughts on how to get the most from your data.