Key Takeaways
- A working system can still become a bottleneck as a business grows
- Better backend architecture can improve performance without changing workflows
- The right system design should support both current needs and future growth
A system can work perfectly well when a business is small. However, as the business expands, it can be gradually outgrown.
It's what happened to one of our ecommerce clients. He had multiple Shopify stores in seven different countries and each store had it’s own spreadsheet dashboard that shows their profit and loss on a daily basis. At first, this is what this setup was used for. As the business expanded, however, that existing reporting system began to fall behind.
When a Working System becomes a Bottleneck
A lot of Google Apps Script was used in the original setup. Approximately 2 to 3 times per hour, each of the seven systems was running scripts that were making API requests to Shopify and Meta Ads.
With our calculation, all seven stores combined it was about 700 API calls per day.
Some point, it began to exceed Google's usage limits, typically at about 6 PM. The client was all but accepting this constraint, and was aware that he could rely on the reporting system all day long, but would probably have problems by night.
There was another issue too. Meta Ads needed authentication tokens that had to be updated every two months. Each spreadsheet was independent and having to maintain 7 separate tokens was another recurring task. These restrictions were starting to be a hassle for the business.
So the architecture needed a re-thinking.
Rebuilding the System Without Changing the Experience
We didn't just want to patch the API restrictions. We wanted to create a space that would accommodate the client's expanding practice, while maintaining the feel and functionality of his current system.
The idea of using a custom web application for the dashboard was discussed. The client, however, was okay with Google Sheets and his marketing team knew the interface that they used. So we wanted to leave that experience as it is. Instead, this was dedicated to rebuilding behind-the-scenes of what was happening.
We re-designed the 7 AppScripts code bases into one backend written in Node.js and deployed it on Google App Engine. All data was shifted to BigQuery, providing a single source of truth for data from all seven Shopify Stores and Meta Ads accounts.
Google Sheets continued to be the dashboard, and where the client could keep track of product information and other inputs. The difference this time was that it didn't need to process a lot of data again and again.
This provided us with a much more separated view of the reporting interface and infrastructure that supports it.
Making Data Synchronization Faster
One of the biggest engineering challenges was coordinating data from multiple sources without making the dashboard feel slow.
The system needed to retrieve updated information from those seven Shopify stores, fetch advertising data, process everything through the backend, and finally make the results available in Google Sheets.
All of this needed to happen within a few seconds.
We optimized the synchronization process so that a complete update from the different sources could finish in approximately 18 seconds. When the dashboard only needed to retrieve existing data from BigQuery, the refresh took under 10 seconds.
This meant the client could switch between stores, check individual products, and review daily profitability without waiting around for the reporting system.
Build for Where the Business Is Going
As we saw in this project, not all systems require replacement when they begin to "show their limitation”. In some instances, the interface is totally unproblematic. What is needed to change is the infrastructure below the surface.
Google Sheets worked well for the client in the initial phases. However, having seven stores, data volumes growing and the need for more and more demanding reporting was no longer the best fit for the original architecture.
We isolated the data processing from the reporting and made it a system the business could expand without having to change its daily operations.
Not all goals are about creating something brand new. At times, it's simply about a good thing working on a significantly larger scale.



