5. Have a model firmly in mind

2012-10-13 15:44:47 +0000

Have a model firmly in mind. No model—no insights!! Not having a clear understanding of the real purpose of the campaign, trend, or other business problem you’re looking at is the main reason for muddy analysis and failure in data-driven campaigns. 

“But Chris, what’s a data analyst to do?" Glad you asked! Here’s a good framework to get you started:

A. Business objectives. Define the business objectives relevant to your question; Do they meet the DUMB (doable, understandable, manageable, beneficial) test?

B. Clear goals. Set some clear and measurable goals for each identified business objective. Making sure your goals are quantifiable guarantees that they’re specific enough. Specificity drives clarity!

C. KPIs. How will you measure your goals? See above.

D. Segments. If you want to take your analysis from good to great, take the time to define some segments, of people/behavior/outcomes; This will help you frame your understanding of the data.

Segmenting, and running analysis on a segment-by-segment basis, will yield such a richer picture of what’s going on. This is something an aggregate analysis has difficulty picking up.

Let’s say you sell high-end sporting equipment online and you run marketing based on a young up and coming superstar in the sport. If you segment by age you bet conversion rates will be vastly different between young folks and baby boomers.

Not segmenting might mask the true success (or failure!) of your campaigns.

That’s all for now, and should get you started on taking your analysis from woolen to great—and you from just a number cruncher to department superstar.