First, a cause-related marketing campaign, spread virally through social networks is capable of causing a positive impact on audience attitudes towards both the brand and the cause. However, the brand seems to benefit more than the cause under this type of strategy. The most appropriate partner must be selected for the specific achievement of certain objectives, and the stimuli should be designed accordingly since both partners can be favoured to a different extent.

Second, affinity (i.e. previous experience with the brand or the cause) can be considered a moderating variable of the effectiveness of a cause-related marketing campaign.

Both conclusions converge in a relevant managerial implication: brands and causes should be aware that the most convenient strategy is to focus exclusively on a specific audience segment, considering its previous level of affinity. If brands address several segments, different messages should be prepared and adapted for each one, incorporating the most appropriate arguments into each of them according to the target audience.

MATERIALS AND METHOD

Design and general description: WHEN and HOW?

A quasi-experimental study was proposed, following similar previous examples (Basil & Herr, 2006; Grau & Folse, 2007; Samu & Wymer, 2009). Our experiment has been developed in two phases:

• PHASE 1. An experiment was done based on a pre-test and post-test design for the same group of individuals. Their attitudes towards a brand and a cause were measured before and after visualizing a cause-related marketing campaign. This campaign was done by the brand to promote the cause. Then, previous and subsequent average attitudes towards the brand and the cause were calculated and compared using two t-tests for related samples.
• PHASE 2. Our sample was separated into three sub-samples. That is, in this second phase, a 1x3 experimental design was prepared, considering the degree of affinity (high, intermediate and low) as a moderating factor. A multiple one-way analysis of the variance (ANOVA) was used to compare these three segments attitudes before and after watching a viral, cause-related video ad. A total of four dependent variables were considered in this phase: (1) pre-attitude towards the brand, (2) pre-attitude towards the cause, (3) post-attitude towards the brand and (4) post-attitude towards the cause. The means of these variables were compared using multiple one-way analysis of the variance (ANOVA). Therefore, this experiment consisted of: 2 (pre attitudes and post attitudes) x 2 (allies: brand and cause) x 3 (affinity levels: high, intermediate and low).

Stimuli selection: WHAT?

To improve the external validity of the experiments (Alexander, 2009), a unique, cause-related marketing ad was presented in video format, in which Pedigree (i.e. a brand of pet food) encourages the adoption of dogs. Previously, that cause-related marketing ad reached significant levels of virality: more than 16,900 views on YouTube and more than 4,900 Facebook shares.

Dependent variables (scales used)

Four types of attitudes were measured on the same scale: (1) pre-attitude towards the brand, (2) pre-attitude towards the cause, (3) post-attitude towards the brand and (4) post-attitude towards the cause. The most common measuring instrument is a semantic differential seven-point scale, with the following four pairs of adjectives: (1) bad–good, (2) negative–positive, (3) unfavourable–favourable, and (4) I do not like it–I like it (Coulter and Punj, 2004; Nan and Heo, 2007).

The corresponding Cronbach's α-coefficients confirm the reliability of the scale in all cases: pre-attitude toward the brand (α = 0.97), pre-attitude towards the cause (α = 0.91), post-attitude towards the brand (α = 0.98) and post-attitude towards the cause (α = 0.94).

Independent variables: affinity level

Before the exposure to the advertisement stimulus, both cause and brand were presented to the respondents to gauge the level of affinity. Based on the seminal work of Macchiette and Roy (1992), three levels of affinity were identified:

1. High affinity: the respondent has purchased the brand’s product (dog food) AND experienced the promoted cause (dog adoption) in the past.
2. Intermediate affinity: the respondent has either purchased the brand’s product (dog food) OR experienced the promoted cause (dog adoption) in the past.
3. Low affinity: the respondent has neither purchased the brand’s product (dog food) NOR experienced the promoted cause (dog adoption).

Sample selection and data collection procedure: WHERE?

The sample selection of this study was carried out on the Internet. Participants were invited to answer an online survey before and after viewing a viral, cause-related marketing ad stimulus in video format. A non-probabilistic, snowball sampling procedure was used. Several Spanish animal charities spread the survey’s link on their Facebook profiles (see Table 1). We chose Facebook because it is one of the world’s most popular social networking sites (Garg, Gupta, Dzever, Sivarajah and Kumar, 2020). Only participants from one country (Spain) were approached to avoid cultural bias. Also, the multinational brand that we used in the present study operates in other countries but promotes other causes and uses other media. Other studies limited to a single country are, for example, those of Jung, Shim, Jin and Khang (2016), who interviewed undergraduate students from Seul; or Cheng, Fu and Vreede (2017), who collected the opinions of 115 undergraduate students from a Chinese university.

Following this procedure, a total sample of 360 Internet users was obtained.