Know-How Exchange : Student Questions

What is the context of this question - a homework assignment?

For one thing, what is baseline data? Because one of the problems with data is how it is collated. Evaluating data is relatively easy when the correct questions have been asked. Which takes us to the project side of things - are the questions used to establish the project sensible?

Which doesn't help much, does it?

Okay, the importance of baseline data is that it is relatively easy to grasp. Which is one reason I dodge it and go for the jugular, but that seems a rare skill. People like the bottom line because it is easy to understand and easy to interpolate. And here, I am putting baseline data into the same broad category. It is, after all, the same level of thinking. Baseline data in this respect gives a quick answer for busy businessmen who want actionable facts.

Which is why I wrote my first paragraph. Because if the wrong questions were asked - would the businessman then be making a bad judgement? Would he even know if he was making a bad judgement - because the data told him it was a good judgement.

My point is that baseline data (= bottom line thinking) needs very little insight. A judgement if it is to be of value needs insight. Baseline data won't supply that, and questions asked without insight into the situation at hand won't supply the necessary clarity either.

Put it another way - an analysis requires a comparison. If I am looking at a direct email campaign, where you can usually track how many people one the email (and presumably read it). So I see that 6% of the emails were opened. Is that good? If you know how many people open the average email (a baseline), then you can compare this. If 6% is more than average, then you could say it is good. If less, then you would say not so good.

In cases where you don't have historical data to sue as a baseline, you need to create it. If you are looking to make a change to a web site, you use the old web site as baseline *what is called A/B testing in this case). You show the old version of the site to half the people who come to your web site, and the new version to the other half. Compare results of what they do (in things that are good for your business - like how long they spend on the site, what pages they go to, whether they purchase or not, etc.). and see which is a better web site.

In a word: control. You can't control (or predict) what you can't measure.

Without a baseline, how do you know if what you're doing is better or worse? Better or worse than what?