Experimental Design in Advertising

There are two status quos from my perspective in traditional advertising today. The first is that many companies retain an agency that provides multiple concepts for review. The other is that we have fallen into the conventional methodology of in-series campaign design. Michele Morrisette

Before staring my own firm, I spent many years in a large corporation where each year, we were given annual sales objectives and a corresponding marketing budget. Being located in the head office, I served the various regional offices in Asia-Pacific, Middle East/North Africa, Israel, Caribbean and Sub-Sahara Africa.

What drove me crazy for almost 16 years was that there was very little, to no accountability for the resources that were allocated to each region. We had anecdotal information that told us if a campaign had been successful or not and then there was the actual sale of a product or service that might indicate a campaign had been successful. However, there were always factors such as weather and economic conditions that may have influenced campaign results but their impact could not necessarily be quantified. It was always, “let’s move on to the next campaign”.

As employees in a company, I believe that we have the responsibility to manage marketing budgets as if the funds were coming from our personal household accounts. Would you spend money at the market hoping that your family will eat food that you randomly picked off the shelf without any consideration for their personal tastes or even allergies? Probably not.

In many cases, that is what businesses do every day; spend money in the hope that their product will yield additional response rates that will be converted into sales. The aim of starting my own firm was to find something new, something innovative that would really change the way we do business. Where the results could be better quantified to understand what worked and what didn’t work in traditional marketing campaigns.

My journey led me to the discovery of experimental design of advertisements. Beginning in 2003, there had been some buzz about what a California based engineer was doing by taking Taguchi Application Design of Experiments into marketing and advertising campaigns with phenomenal success rates. Most people in marketing and advertising had no idea who or what Taguchi was, including me.

Did you know that if you choose just 15 different elements in a marketing campaign to test, with two options each, you could test the equivalent of 32,768 possible combinations using only 16 tests? How cool is that?!

A new world opened up and I felt as though I had perhaps reached nirvana from a marketer’s perspective. We could test up to 32,768 possibilities in one campaign and have statistically valid responses at the end of the test.

Knowing that experimental design can help optimize an advertising campaign, why would agencies and organizations not embrace creating multiple designs and testing a sample audience to arrive at an optimal design and then deploy that message to the mass target audience?

There are several reasons. From an agency point of view it is frightening to think that their creativity might be challenged. Creative people don’t necessarily want to know the quantitative results of their campaign. While they want to please the client, they also have an objective of aiming to win the coveted creative awards that are so highly regarded in the industry.

From an Ad agency point of view, accountability might limit their creativity. From a company point of view, many decision makers are like I used to be, a bit naive and are not familiar with the concept. They just allow the agency to guide them along believing, that based on the vast sums of funds spent, accountability will be part of the process.

Another issue is that there are a wide range of available techniques; from the very basic method of assigning a priority to key words for internet and web landing page use, to very intensive on-sight brainstorming and facilitation sessions with internal and external customer input. The channel of advertising is also very critical.
While these techniques can be applied to newspaper and television campaigns, the more immediate return of results and investment will be seen in direct mail and Internet advertising campaigns.
The results can be staggering.

In a study run at Dell in 2004 the focus was on internal employees and the company e-mail ad sent to this group.An 18-test layout was created and in this case, the 520% increased sales results speak for themselves[1]:
• Click Through Rate increase: 5.2 times
• 7.1 times more sales per e-mail
• Annual sales before optimization: $8,900,000
• Annual sales after optimization:  $63,100,000

It also requires both agency and client embracing a new way of doing business. Instead of perceiving the testing as a threat, it should be perceived as a tool that can help an agency provide those quantitative instead of just qualitative results. In fact, a really progressive agency might see the potential application when aiming for that creative award.

Instead of quietly submitting the one brilliant idea, that the client considered to be too risky, to one paper for the purpose of making it eligible for judging, why not show the client how it can be tested and take away the fear factor and openly test the concept?

I believe the first agency willing to embrace the experimental design concept will achieve not only higher client satisfaction but also be able to add to their bragging rights. Now they can claim that along with being the most creative agency, they are also the one that is looking out for the client’s best interests by ensuring that the optimal campaign is deployed.

Link to the original article in Profit Magazine Experimental Design in Advertising-1

Disclaimer: All insights and opinions are my own and have been formed over many years of practical business and personal experiences; as both an employee and a business owner. They do not represent the views of  my current employer. Michele Morrisette