Setup for Advertising Trackers

From Q
Jump to navigation Jump to search

Advertising trackers regularly replace the campaigns being tracked. This page discusses how to setup data files to deal with this (see Data File Setup for Tracking Studies for other ways of setting up data for trackers).

There are a variety of different ways of addressing this. They can differ substantially in how ideal they are for a given project and a cost benefit analysis should be undertaken prior to making a decision. However, in general, stacking is the best strategy.

Stacking the advertising data

Rather than having a single data file for the tracking project, there are instead two files. One file which contains:

  • In ID variable.
  • The advertising data in a stacked format (e.g., if a respondent evaluates three campaigns then there are three rows of data for that respondent int he data file).
  • A categorical variable indicating the advertising campaign.

The other file contains all the other data, and an ID variable that can be matched with the stacked data. The data is then analyzed in Q using the Multiple Data Files tools with different advertising campaigns automatically analyzed by using filters.

New variables for each new campaign

From a data file creation perspective, the simplest approach is usually to have new variables for each campaign. However, this is often the least helpful way of setting up the data as it complicates the analysis (see How to Analyze Advertising Trackers with Different Variables for Each Campaign) and there is no straightforward way of automating the updating of any reporting (i.e., updating of PowerPoint, Excel, Word or Dashboards) if this approach is used.

Re-using variables

For example, if one ad is taken out of a tracker and replaced with another, then ensure that the data for the new ad is exported in the same variables as the old ad, with another variable then used to create filters. This ensures that any work done in Q is automatically re-applied.

When this is done it is generally prudent to include a variable in the data file which flags the campaign that the data relates to.