Editor's note: Joe Abruzzo is executive vice president, chief exploration officer, at Havas Media, New York.
Tracking research is an important means of monitoring brand health. How is the brand tracking on measures like unaided or total brand awareness? Consideration? Brand preference?
We are interested in levels and trajectories across a range of brand health metrics. How healthy is the brand? Where is the brand headed?
The problem is, brand health tracking measures tend to move slowly. Variation that accompanies smallish samples clouds the brand health picture. It’s difficult to decide when it’s time to celebrate or begin to take corrective action.
When working with survey data, it’s a good idea to understand what is being reported. What time interval does the data represent? How often is the data captured? Is the data being captured continuously? One week per month? One month per quarter? How many interviews represented by each data point?
Some data providers choose to “roll” their data, which means creating a moving average that reduces sample bounce and introduces stability into data trends. This may seem like a great idea. However, the individual data points are not independent. A three-month moving average means that each data point is only one-third current. Changes from one month to the next are dampened. Emerging trends may be obscured. A data set with discrete data points where each data point is independent of all previous data points helps to avoid such a problem.
Active investors in the stock and commodity markets deal with these kinds of decisions every day. Many use simple tracking systems that generate buy-and-sell signals, or, at the very least, indicate when to pay attention.
It’s common to track stock and commodity prices in comparison to the highest and lowest price observed over the most recent 52-week interval (referred to as a 52-week channel).