Conversions & ROI: Creating Baseline Metrics

As a marketing professional, people ask me questions like, “How many unique visitors do you think you get me in a month,” or “How many sales will you be able to generate?” Being the truthful person that I am, these types of requests put me at a loss for words because reliable figures take time to generate.

My last post was about the challenging issue of predicting ROI and conversions. Again, these types of metrics can be viewed as a way to assess the performance of what has already been done. This is obviously different from wanting to know what will happen during the course of your campaign.

Reporting and analysis during the beginning stages should be focused on collecting solid baseline data. Baselines are used in science research. They can be thought of as snapshots which tell you about the average level of performance for a specific part of your campaign.  For example, you would have a baseline for Twitter and a different one for Facebook.

Wordnetweb.princeton.edu describes the concept of baseline as, “an imaginary line or standard by which things are measured or compared.”

Getting reliable “snapshot” data requires that you collect information on numerous instances.  More is better. For example, if you are trying to determine the average number of Facebook fans acquired per month, it’s better if you had data for 18 months as opposed to 2.

So if your campaign did exceptionally well during the first two weeks, it’s silly to use these numbers to gauge your success for the next two months. I’m sure this sounds really obvious. But a lot of people do this. Looking at large spans of time will only serve to help you in the long run.

Once you establish solid baseline figures, you can measure the performance of any new advancements or additional components to your campaign.

Because my original background is in science, it’s natural for me to see how basic practices in research design would have a fitting role in marketing. I was thrilled when I found this article which outlines a very practical method for measuring your ROI.

http://www.stayonsearch.com/how-to-measure-social-media-roi

This is definitely a post to not only bookmark, but to use as well.

ROI’s and The Scientific Process

Most of us are aware that the issue of predicting conversions figures like ROI, for example,  is a tricky one in social media marketing. Clients obviously want to make wise investment choices. And in this economy, no one wants to throw money out the door. Marketing professionals obviously want to win projects and are pressured by the need to provide assurance on the value of their time and efforts spent.

Making quantitative assessments requires the obligation of providing valid and reliable predictions. Understanding why it is tricky to promise conversions and ROI during the initial planning stages of a campaign can be explained by understanding the scientific process.

We depend on scientists to tell us valid information. A knowledge of  basic research concepts will help us make better decisions in social media marketing.

Multiple, Interacting Variables

Before I elaborate further, let’s think about how conclusions are derived in the scientific world.

1. You have an independent variable. This is whatever is conjectured to cause a specific effect.

2. You also have a dependent variable. In simple terms, think of this as the label for the results that you are testing.

In other words, the independent variable is supposed to cause the dependent variable to occur.

In a controlled experiment, you can only test one variable at a time. If you have several variables occurring, you would not be able to conclude that the results were caused by a specific factor.

Campaigns occur in the real world, not in controlled settings. There are too many variables interacting with each other.

The success of Twitter campaigns, for example, depend on factors like

• The quality of a brand’s offerings
• The popularity of a brand’s offline presence
• Whether or not a brand is natural social part of our culture. (i.e. do people talk about toothpaste and paper towels?)
• How well the Tweets are written

In the planning stages of a marketing campaign, not all the variables are identified up front. As they say, there is usually a huge gap between theory and practice.

 

 

During the course of a campaign, other variables and dynamics often occur. Although a marketing specialist can have extensive experience with a particular vertical, they also have to take into account their client’s individual characteristics as a brand. There is no point for a brand to be exactly like everyone else. They have to differentiate themselves in order to survive. These points of differentiation introduce other variables into the picture.

Repeatable Testing

Large numbers of people are necessary to form valid reliable conclusions. Are you going to make global strategic conclusions based on a group of 20 people? Or would it make sense to utilize knowledge based on hundreds or thousands of members in your target audience

Scientific conclusions are founded on numerous tests on large numbers of subjects to ensure reliability. Sometimes a result occurs due to chance. Repeating experiments is necessary to rule out chance occurrences.

 

Campaigns can be tested and measured to some degree (e.g. display ads). However the market, for the most part is dynamic and in constant motion, as opposed to static and predictable.

 

Establishing quantifiable parameters around a given marketing campaign is not an impossible feat. A lot of thoughtful consideration needs to be done to form any type of conclusion that is reliable.