Mediator transformation

Recently, I carried out a questionnaire about the effect of a digital communication channel on willingness to buy a personalized advertised product. In this research, two mediators are tested, namely privacy concerns and attention. The first mediator is examined by asking a question where the respondent answers on a Likert scale, so this data is easily processable in SPSS. However, the second mediator (attention) is examined with this question in the survey: Arrange the following channels according to the level of attention you would give to a personalized advertisement if it is exposed to you through these channels. 1 is the advertisement of the channel that you would give the most attention to, 3 is the advertisement of the channel that you would give the least attention to.

-Email advertisement
-Facebook advertisement
-SMS/MMS advertisement

Now is my question, how should the answers of this question be transformed in SPSS to make the data workable, so that I can measure the presence of this mediator?
I've had two suggestions so far, but I am not sure if these are appropriate: One solution is to transform ranking to Likert scale: so the most preferred brand will take the value of 5(if the max. in the Likert-scale), 3 for second and 1 for last preferred option.
Another alternative is to have a dummy variable, which takes value one when a channel is preferred most. Therefore, for instance, when the social media is ranked 1st, then the preferred channel will be 1 and other two channels will be zero.
Note: The questionnaire is carried out already, so changing the question to one with a Likert scale answer is not a solution.

Thanks!