Measure What Matters: 4 Steps to Making Sense of Your Data

Data can be intimidating. The co-author of the recently released ‘Sexy Little Numbers’ helps you determine what kind of information your business should be tracking — and how to make sense of it.

With the perspective gained from helping client after client make the most of their data, I’ve found that one of the biggest problems they face is figuring out how to make a coherent narrative out of the wealth of data they accumulate. All the information in the world doesn’t add up to anything if the companies collecting it can’t figure out how to use it. How can you make sense of all this data constantly being gathered on the Web?

The answer is measurement planning. At Ogilvy & Mather we use a process called OgilvyEvaluate, which has become the standard measurement approach for our global advertising and marketing network. While it started as a tool for Web data management, it soon became clear that it could be used for any type of measurement.

Let me show you the four steps to that process, as discussed in my book “Sexy Little Numbers,” written with Paul B. Brown:

1. Prioritizing objectives. This is the most fundamental part of the entire process and the one companies most often skip when implementing a measurement system. And that’s a real shame — because when you get this wrong, everything else you do from a measurement point of view won’t be effective.

To get it right, participants first define what success will look like by writing the objectives of the program they want to develop. If you don’t make the objectives crystal clear, then people will very quickly develop their own thoughts about what success means. The consequences can be incredibly counterproductive.

2. Rewriting objectives. Rewriting the objective is a simple but incredibly useful exercise. Objectives are only real objectives if they have a metric, a benchmark and a time frame.

If objectives don’t have each of these three components, they’re more loose aims or aspirations. So, during this step in the process, we go through the list of objectives from the previous exercise and reformulate them to make sure they contain all three components.

At Ogilvy, for example, our mission is “to be most valued by those who value brands.” That, in and of itself, isn’t an objective. To make it so, we might say, “To achieve our goal, we will __ ” (and here, I am making up the numbers):

- Increase revenue per employee by 22 percent annually

- Increase the number of employees who can use our knowledge management system from 65 percent to 80 percent within two years

- Reduce forecasting errors from 20 percent to 5 percent by Q3

Each of our additions has a metric, a target and a time frame. This approach is similar to “SMART objectives,” which you may have heard of in the context of discussing the best way to achieve a goal.

A SMART objective is:

- Specific. Objectives should clearly specify what they want to achieve.

- Measurable. You should be able to track whether you’re meeting the objectives or not.

- Achievable. Stretch goals are encouraged.

- Realistic. Can you really achieve the objectives with the resources you have?

- Time-based. When do you want to achieve the set objectives?

By reformulating your objectives to make sure they’re SMART, you automatically create a list of your most important Key Performance Indicators. This list is the starting point of the next exercise. But before we discuss the KPIs, we need to spend a moment on how we’re going to judge how well they’re doing. And to do that, we need to talk, briefly, about benchmarking.

Benchmarking is about context. You can choose the right metric. But when you get the numbers back, how do you know if they’re good or bad? That’s the question benchmarking is designed to answer. By comparing your company to others that are doing the same thing, you have a way of rating your performance.

There are two ways to obtain benchmark information: from external sources and by developing your own internal data. And there are all kinds of applications and products to help you there. (We talk about these specific tools in our book.)

But you may not have to do a lot of research to judge your performance. Often the best benchmark is a comparison of your own performance over time. Yes, it’s a very common measure, but you’d be amazed how often marketers forget to use it.

3. Action learning indicators. Putting in place and measuring the right KPIs is often not enough. You need to know why a certain metric goes up or down. Action learning indicators come into play by providing insights into what drives certain KPIs. It’s important here to have a broad range of metrics that will not only enable you to assess the performance of your campaigns, but also let you understand why you’re performing as you are.

4. Data source mapping + measurement plan. Once the list of KPIs and action learning indicators is complete, you can move on to mapping metrics to data sources and actually begin the tracking to see how well we are doing. A typical multichannel measurement system will use data from a broad range of sources. Some companies have these sources integrated into an all-encompassing corporate data warehouse. But most don’t. They’re spread out all over the place. The implementation stage usually involves a lot of data integration.

Once you know how to find the data you need, then you can start measuring performance.

This article is excerpted from the book “Sexy Little Numbers: How to Grow Your Business Using the Data You Already Have” by Dimitri Maex with Paul B. Brown.