Social Marketing and Message Testing: The Ipsos Approach

Traditional message testing approaches typically focus on the message alone: Is it good? Is it bad? Is it clear? Is it credible? However, our experience shows that effective messaging is a function of three variables: the Message, the Messenger and the Target Audience.

The most elegant message will not be believable if the messenger is not credible or the target audience is wrong. We believe that the research design should include all three components in order to produce the most actionable results. Let us explain in further detail each of the components.

The Message

Here, the focus is on the composition of the message, and looks at whether it meets the criteria of the 4 C's: Clear, Concise, Credible and Crisp. We also assess the extent to which it changes people's attitudes and behaviors. This is easier said than done.

The research options are endless. Our experience shows that message testing should be iterative -- that is, the message should be tested and re-tested, and, if possible, tested again.

We call this the message funneling process, and have distilled it into three phases:

• Phase 1 includes testing initial message hypotheses. Messages can be vetted--effective versus not so effective. Typically this stage is qualitative in nature--including one-on-one interviews and/or focus groups.
• Phase 2 consists of testing message statements. We look for the message's degree of believability and to what extent they can modify opinion and behavior. This stage can be qualitative or quantitative in nature, but quantitative is usually required so that statistically reliable data can be collected on what is and what is not working.
• Phase 3 re-tests the message statements in Stage 2, though with adjustments. The interval between stages 2 and 3 allows for considerable client input, using the results from stage 2 in the adjustment process.
• The final product is a message that meets the 4 C's.

The Messenger

Messages do not exist in a vacuum! Indeed, messages need to be delivered by a messenger--whether an individual or an institution.

Our experience--and the empirical evidence--suggests that if the messenger is not credible (does not have a good reputation) people will be less likely to believe the message. The more that people trust a given institution, the more likely they are to believe its advertising or be willing to participate in the given cause. This is illustrated in the graph below.

However, it is not simply about the reputation of the messenger. Instead, we are most interested in the message-messenger fit: is the message effective given the messenger. Typically, the messenger is fixed in the short-term, so it is much more a search for the right message than the right messenger (see table below). The message funneling process is where typically we determine "optimal fits".

Two caveats are in order. First, the reputation of the messenger may only play a role when the institution rates high in awareness like the American Red Cross or the NRA.

Second, in practice, reputation may be linked to the believability of the communications channel: are people more likely to believe a message in a TV ad, on the radio, from their pastor, or from their doctor?

The Target Audience

Even with the best message, delivered by the best messenger, if it is communicated to the wrong people it will have little effect. Here traditional targeting considers demographic variables in the definition for the target: for instance, Americans between the ages of 18 and 30.

Experience shows that such broad demographic definitions typically are not a cost-effective way of getting at the "right" people--those most likely to agree with the issue or those most likely to change their opinion. Instead, more nuanced target definitions of sub-groups are typical.

We call the process of targeting as target audience funneling.

General Target means targets based upon broad demographic definitions, such as Americans between the ages of 18 and 30.

Specific targets mean sub-groups within the general target. Such sub-groups have unique behavioral and attitudinal properties that make them more likely to believe a given message. This sort of specific targeting can be done with more detailed demographic data; Americans between 18 and 30 years of age that are married. However, our experience shows that sub-groups defined by behavioral and attitudinal variables are most effective: Americans between 18 and 30 years of age and who are married who believe that divorce is a sin.

The targeting segmentation schemes can vary greatly. The following traditional three-variable model serves as an example.

Here it is worth stressing that, contrary to traditional approaches, this segmentation scheme is done with actual data from the survey (behavioral and attitudinal). Finally, we believe that, to make targeting actionable, it should be further validated with research --typically with focus groups.

Effective messaging

As detailed here, effective messaging is a function of three variables: the Message, the Messenger, and the Target Audience. However, what is effective messaging?

Ipsos defines effective messaging as the ability to move the needle on measurable metrics; to change people's attitudes and behaviors on something quantifiable. To put this into conceptual terms, effective messaging is messaging which "moves people up" what we call the Advocacy Pyramid.

Simply put, a person needs to know something about you to like you; a person needs to like you before he/she will advocate for you. Effective messaging makes people advocates.

Here we can measure movement of the metrics in two ways:

• Such metrics can be tracked over-time with multiple waves.
• Metrics can be tested within a given survey instrument using a traditional questionnaire test / re-test design. In this case, we measure the above metrics; then we present the message or messages; and then finally ask the same metrics a second time.

Finally, we believe that effective messaging is only possible when research is actionable. As such, clearly tying the message to the target and then knowing how to reach the target is essential.