Aeropost stands as the only end-to-end eCommerce and logistics company in Latin America and the Caribbean. They have maintained a partnership with scandiweb for nearly a year for QA and analytics projects, where we extend project management to their development team. This case study zeroes in on our work in analytics, delving into how we tackled their specific business needs and problems and the smart solutions we came up with to ensure they gather all crucial data for insights and data-driven decision-making.
Business needs and challenges
The project revolved around fulfilling crucial business requirements and tackling some unique challenges. Firstly, with the sunsetting of Universal Analytics, we had to migrate the client to GA4 while simultaneously enhancing and expanding their entire data tracking system. Then, to allow them to leverage the wealth of information offered by the upgraded GA4 data, it was vital that we set up Business Intelligence (BI) reporting. Regular BI reporting is invaluable in data-driven decision-making for eCommerce.
However, the project encountered distinctive challenges due to the client’s use of a custom eCommerce platform to host their website. Unlike widely used platforms like Magento and Shopify, this custom solution did not provide pre-built dataLayer and Google Tag Manager (GTM) modules. Consequently, the project involved the intricate task of constructing the data layer from the ground up, underscoring the unique nature of this project. Furthermore, the situation became more challenging because the client operated in many different markets. This meant we had to build a strong and flexible tracking solution that could fit their various requirements.
Approach and solutions
Creating a tailored solution for our client’s unique needs was a complex journey that demanded careful planning and technical proficiency. Our initial challenge was developing a data layer from scratch, as no existing modules were available for their custom platform. This process was a multi-step endeavor.
Developing the data layer
First, we defined the specific data requirements, planning how the data would be structured and organized efficiently. Detailed data layer requirements were then prepared, providing developers with a clear roadmap.
It’s worth noting that this project had a unique feature—the client had their own development team that scandiweb managed. When the developers had built the data layer, the QA team from scandiweb conducted extensive quality assurance testing on staging and production environments to ensure data accuracy.
Setting up GTM
We set up Google Tag Manager (GTM) to manage and deploy tags seamlessly, complementing the data layer.
We had 2 GTM containers: one used for the old tracking and another for the new setup. We migrated all of the existing client marketing scripts to the new container and optimized them where necessary. GTM tags underwent rigorous testing to confirm their functionality.
Additionally, when the need for tracking user interaction with specific homepage elements arose and the developers were tied to higher-priority tasks, our web analytics specialists came up with an immediate solution. They set up tracking with custom data layer pushes from cHTML tags in GTM, which gave the client the means to obtain the information they wanted until the developers were available to create a long-term solution.
After creating the GA4 property, we configured various settings such as enhanced measurement, data retention, data collection, referral exclusion, and custom definitions. This ensured that the property collected the right data accurately. Additionally, we customized default reporting by setting up exploration reports and adjusting default dashboards to fit our client’s specific needs. This made it easier for our client to analyze their data effectively.
Setting up the BI dashboards
We used Google BigQuery for advanced reporting. We gathered the client’s reporting requirements, established a data export pipeline from GA4 to BigQuery, and created tables for report generation. Finally, within Looker Studio, we crafted and extended dashboards, delivering intuitive Business Intelligence (BI) reporting to empower data-driven decisions.
Initially, we built dashboards directly connected to GA4, which worked well. However, a couple of months in, a problem emerged. New GA4 API request quota limits caused these reports to break.
Luckily, unlike the free Universal Analytics, Google Analytics 4 has native integration with BigQuery, a data warehouse from the Google Cloud Platform. GA4 allows the export of raw unsampled data to BigQuery, and the connection between the BigQuery data source and Looker Studio is not subject to the aforementioned API request limits. Even better, reports based on Big Query data are faster and allow for more advanced data manipulation.
So BigQuery became our go-to solution. We made the switch and developed the custom code necessary for the dashboards to operate smoothly.
Since the default export of GA4 data is raw, many fields cannot be accessed directly. This is the reason why we had to develop custom BigQuery SQL code to create custom tables—to serve as data sources for the dashboards. Once the data source was established, we scheduled the queries so that they ran automatically at the frequency requested by the client.
It’s important to mention that we had the client’s GA4 connected to BigQuery almost from the very beginning, which is what we always recommend, so they didn’t lose any historical data. As the direct GA4 data export to BigQuery can’t be done retroactively, establishing the connection between GA4 and BigQuery as soon as possible is paramount.
As for the dashboards, it’s worth emphasizing that extending them is a continuous process. We work closely with the client to constantly improve the BI reports and respond to new business needs as they arise.
The data layer is fully operational, effectively tracking essential data that fuels our Business Intelligence (BI) reporting. Notably, the BI reports we produce are actively utilized by the client, indicating its value and practicality.
Right from the start, expectations were set high, and we’re proud to say that we’ve met and exceeded them. As the client expanded into different markets, our commitment to data accuracy remained steadfast. There have been a few bumps in the road, but we regularly conduct extensive QA checks for each market to ensure that data tracking is precise and issues are addressed immediately.
At the end of the project, we extended our services by introducing Facebook Ads tracking for the client, and work is currently underway for the implementation of Google Ads. These additional tracking capabilities add another layer of insights for the client’s business, enhancing their ability to analyze and optimize their advertising campaigns effectively.
Are you in need of analytics and BI solutions? Want to leverage data for better decision-making? Connect with us at scandiweb. We specialize in crafting tailored solutions to unlock your data’s potential. Reach out today to start your data-driven journey and transform your business. Your insights are waiting—let’s make them work for you.