One of our clients, a large financial institution focused on Car Loans, asked us to find hidden opportunities for better targeting customers. This is how we did it!
By analyzing Google’s category interests information connected with actual data from clients loan applications we were able create statistical models to find hidden customer segments with higher conversion rate potential. By micro-targeting those clients we managed to increase conversion rate by 15%, attracting new customers while decreasing Cost-of-acquisition.
Our client, as many brands do, was facing a challenge - their advertising campaigns were plateauing.
Without additional investments, they could not attract the number of clients they needed to reach sales targets. Also, their digital marketing campaigns were based on very generic segmentations, without taking into account what profiles of customers they were actually attracting and converting.
With stalling results, pressure for sales and a need to improve their data driven marketing strategy the client came to us with one question:
How to use data to improve advertising campaigns?
Our solution was to look into current clients to understand micro segments that we could use as references for new adverting segmentation and specific messages to attract new customers
To answer that question we used the category interests provided by google analytics as well as the actual information provided by clients at the purchase funnel [automobile they wanted to finance, age, Gender, City/State, among others)
From there, we created a data set that connected the category interests from our web analytics tools to actual car loan applications by clients. We created a classification tool to segment car models into specific types, to help understand why people were buying those cars.
Also, by analyzing the customer journey from top to bottom of the funnel, we could actually measure the conversion rate from specific profiles in order to understand each ones had higher potential for sales.
From our original generic ad set we managed to create more than 10 different segmentations. Each specific audience had a potential conversion rate higher than the client's average performance of 2.5%.
Urban women under 40 without interest in cars.
The one segmentation that caught our attention were urban women under 40. They were a small percentage of the original target audience, but they had much higher conversion rate than the average. Interestingly, they were not flagged with an interest in cars, but with decor and home improvements. By analyzing the actual cars being financed by people with the same profile, we realized that they were family cars. That gave us our main insight about a group of clients purchasing cars much more as a tool to solve a problem than a object of desire.
This segmentation and message was than applied to several medias in order to attract new clients with that specific profile.
Taking the data and turning into a digital marketing plan
From that original data set we helped our client design an specific audience for that public, advertising their focus on family vehicles. That specific segment had a conversion rate 25% higher than the average for the channel, decreasing the overall cost of acquisition for the channel. By targeting this specific group, we were also able to increase their overall participation in the number of loans sold.
Show me the code
In order to solve that problem, and provide our client with better insights to improve their digital marketing ads, we first started by completely redesigning their web analytics structure.
By correctly applying Google Analytics 360 Tags, we made sure we were able to properly collect every interaction users had with the platform. Our tags made sure we were able to collect information from every step of the funnel, including all of Google´s rich Demographic and interest categories. That information would be latter used to help us design our target profiles for all platforms.
By seen the changes in performance throughout the funnel we could identify the conversion rate for each profile that we were attracting
We also carried all of Google Analytics information to a separated database in order to enrich it with other data to refine our analysis.
All of the information provided by customers during the online loan application was already collected by the application to be used for processing requests. We also collected the non - identifiable data (to respect privacy laws) from the application to a separate that information the database already populated with GA data
Sessions x Users
Unfortunately, Google does not give us the Interests information on a user level, only sessions. To bypass that, we used some advanced statistics to create cohorts that combined both levels into a understandable audience, with specific demographics interests and car applications. Those were the basis for the audience design as well as creative briefings.
want to boost your business?
send us a message and we will get back to you as soon as possible!