Abstract
Goals
Better understanding of online users
The uses of Audience Score are not limited to remarketing campaigns, although it was the main use in this case. What it offers is an evaluation of how close a visitor is to completing the objective of the site.
Optimize remarketing conversion
Test the utility of Audience Score, and how well it can optimize remarketing campaigns by doing an A/B test with Meliá's existing remarketing campaigns.
Reduce the cost of remarketing
See if using Audience Score limits the number of users impacted by remarketing, reducing associated costs.
Approach
Identify variables that should have weight within the model
See which variables Metriplica's statistical model should keep in mind to generate the Audience Score. These can vary depnsing on the client, their business, and the characteristics of their website.
Implement tags to measure all of the variables
Once we identify all of the points, we have to make sure the information we collect is sufficient and of good quality to properly calculate users' Audience Score.
Analyze the improvement of the campaigns that use Audience Score
Observe the evolution of these campaigns to see how the use of Audience Score influences the conversion from these campaigns. Establish which users should and shouldn't be impacted.
Results
Improve the ROAS of remarketing campaigns by 354%
After performing an A/B test, the campaign using Audience Score obtained a 354% increase in the ROAS, 72% in savings, and a 625% increase in conversions.
Needs
Meliá is a world leader in hotel complexes, and a leader in the South American and Carribean markets, with more that 370 hotels open (or in the process of opening) in more that 41 countries.
They carry out remarketing campaigns on users that visited their website based on what products they looked at.
The segmentations that they currently use for their campaigns are the same as most other companies, and the results are far from profitable on their own.
Challenge
The main challenge of this project is to implement Audience Score and see how well this statistical model can improve remarketing campaigns.
If they see that Audience Score works, they could use it for other sectors such as email marketing, site personalization, activators and others since they can export Audience Score information to a CRM or any other tool that the company uses.
The delicate part of the process is adapting the solution specifically to Meliá since they collect a lot of information on their website and it's essential to define what the model must take into account and what it should give weight to.
Solution
We statistically and automatically identified those variables connected to the behavior of users and their interactions with the website that were related to the key actio nof making a reservation. We exported the Google Analytics data and ran a statistical model on our server. Once we got the score for each user, we uploaded it again to the web analysis tool using a personalized variable.
To perform the test, we wanted to use a regular remarketing campain that the company used before knowing abotu Audience Score.
For the test, we generated two identical campaigns in terms of content, but each with a different audience: the first campaign would only impact users with a high Audience Score, and the second, control campaign would impact any user regardless of their score. Since we wanted each user to only see one campaign, users with a high Audience Score were divided 50/50 between the two campaigns to make it as fair and equal as possible.
Results
Thanks to the use of Audience Score, each euro invested earned more that double than without it.
- The ROAS was 354%
- The increase in conversions from remarketing campaigns was by 654%
- The decrease in the cost of gaining a customer was by 79%
Otros casos de estudio
Tracking offline transactions with Google Analytics
Cuadro de mando para monitorizar el uso de la intranet
Revenue Funnel, entendiendo dónde se produce la caída de ingresos
