Research Analysis - Avoid Data Paralysis in 5 Easy Steps

I've stated it before. Analyzing research data is a daunting task. It's complicated because, often times, it is difficult to know where to begin. Qualitative, or behavioral data analysis, seems more manageable, because there is a structured methodology and limited data to cull through. Quantitative data analysis is an entirely different beast. Huge sample sizes, hundreds of measures, and results which include "the passage of time" as an axis, offer challenges that (as Information Architects) "put our name to the test" (King Leonidas, 300).

As always, solving a problem begins with "breaking it down." Abstract the solution, and develop an outline that is the framework for a persuasive argument. Draw the lines first, and then, fill in the details.

Here's an approach to analyzing quantitative data, in five (5) steps:

1. Define objectives - Does this 1st step need to be restated in every blog entry? Every challenge has the same beginning. Consultants succeed only when they achieve client objectives. Research analysis needs to support these objectives, otherwise, research is a wasted investment for a business (we can debate the value of "research for the sake of research." I understand the NASA model, where innovation and "spin-off discoveries" lead to practical applications. I'll argue, that is, in itself, an objective.) In a business context, where clients are spending money in a crap economy, Information Architects must do everything in their power to prove the value of the research they are conducting. Spell it out, and be promotional about the fact that the research findings will directly support meaningful objectives. This exercise has the added benefit of enabling you to develop a framework for your research analysis.

2. Develop theories based on prior knowledge - Most likely, clients really do have a base of understanding and knowledge about their customers. Often, theories exist that project a relevant set of customer behaviors over time, or provide an explanation to previously recorded behaviors. Borrowing from the "scientific method," Information Architects can develop a measurement framework that is aimed at supporting, or disproving, theories. List the theories that are related to key customer behaviors that support business objectives. Understand what action items are a result of proving or disproving these theories.

3. Conceive "psuedo-measurement equations" to support objectives and theories - It is now time to understand how data actually proves or disproves theories. You aren't going to even look at the data, yet, to complete this step. It is much easier to dig into the intimidating pool of data, once you understand what to look for. Creating the measurement equations is what this step is all about. Take a theory like, "customers are abandoning the shopping cart because there is an offer that appears on the shopping cart page that is drawing them off the path of conversion." What is an effective measurement equation to assess the validity of this theory? Maybe, "of the X% of customers who abandon the shopping cart, X% are clicking on the offer that appears on that page." The relevant data reports that need to be generated, include:

> Measurement - percent cart abandonment
> Measurement - percent cart abandoners that click on Offer X

Identifying the key measurements that prove or disprove theories about relevant customer behavior is a matter of developing a logical argument. Develop these arguments for each theory that exists to support a business objective, and you will have defined the necessary measurements required to achieve the next step in the analysis process.

4. Look for actual data to support "pseudo measurements" - This step, which is the one that must cause the greatest amount of anxiety, is actually the easiest step. Step three (3) entailed defining the actual measurements to make. Step four (4), is a simple matter of generating the data reports to support defined objectives and theories. Once the relevant data has been retrieved, analyze whether or not the numbers prove or disprove your theories.

5. Develop recommendations - Once again, from a business perspective, data is only useful when it is applied. The last few steps to analyze data have been in pursuit of supporting objectives and proving and disproving theories. This step requires the definition of actionable steps that need to be taken to increase specific customer behaviors, or reverse negative trends in specific customer behaviors. In essence, this step is about making meaning out of the data. It is, perhaps, the most difficult step of all. Drawing from your own expertise in the medium, knowledge about the customer, competitor tactics, and proposed tactics, select a set of tactics that addresses the need to support unsupported objectives, and reverse negative trends in customer behaviors. List these tactics as recommendations based on the analysis.