The Tablo Case in Universities: All About Software Scalability

What do Tablo, Ca’ Foscari University of Venice, and software scalability have to do with each other? On May 11, Mabiloft had the pleasure of bringing these three topics together in a lecture delivered by our Head Frontend Engineer, Giulio Trolese, at Ca’ Foscari University in Venice.

The lecture took place as part of the Software Performance and Scalability course taught by Professor Andrea Marin within the Master’s degree program in Computer Science and Information Technology.

Mabiloft has long maintained a relationship with Ca’ Foscari University as an internship partner, which gave us the opportunity to return to the classroom—this time not as students, but as professionals ready to share our experience.

By connecting theoretical concepts such as architecture, databases, and algorithms to a real-world case study we faced—Tablo—we addressed a topic that, sooner or later, every digital product must deal with: software scalability.

What software scalability means

What does it mean to scale a piece of software? Software scalability is its ability to handle increased workload, such as that resulting from a significant rise in the number of users.

It’s not just about avoiding crashes, but also about maintaining performance in line with user expectations. An app that doesn’t scale well becomes slow and unusable from the user’s perspective: according to a 2018 study, 38% of users are ready to abandon a page that doesn’t load within 5 seconds.

And yet, any skilled developer is capable of making the right technical decisions to allow an application to scale without issues. The real challenge, however, is reconciling the technical side with business needs. The solution that is optimal from a technological standpoint is not always compatible with what the product can actually afford.

The Tablo case

When we launched Tablo V2, we knew scalability would be one of the critical points. The new version of the app had improved the user experience by redesigning user flows, but this led to an increase in the number of users—something the existing infrastructure could not sustain.

During peak usage times, for example just before the weekend, the application began to show signs of strain. In particular, retrieving the list of events could take up to 8 seconds, well beyond the tolerance of the average user.

We needed to act quickly, before users started abandoning the application. And there was a solution. Unfortunately, that solution was not applicable to Tablo.

Users were split between those who had already installed the new version of the application and those who were still using the previous one. Taking the necessary measures—the most obvious ones from a technical standpoint—would have excluded users who had not yet updated the app. In practice, the risk was that these users would go from a subpar experience to not being able to use the application at all.

We therefore had to find a solution that could simultaneously:

• Ensure optimal performance
• Be released to production as quickly as possible
• Avoid losing a portion of our user base

We won’t bore you with the technical details, but if you’re also experiencing difficulties scaling your application, we at Mabiloft can help. To explore more case studies or book a no-obligation introductory call, visit our website.

The lecture on software scalability

But how did the lecture at Ca’ Foscari University of Venice go? Student participation proved to be a valuable addition to the presentation of the Tablo case alone. Questions, observations, and curiosities allowed us to reflect on the decision-making process that leads a team to prefer one solution over another.

In an ideal scenario, everything is optimized to its fullest potential. But this often clashes with reality, where scalability must also be evaluated in light of the strategic decisions that need to be made throughout a startup’s lifecycle.

As for the message we wanted to leave to the students, below is a summary of the main points we covered during the lecture.

Don’t scale everything right away. Designing an oversized system too early can slow down development and drive up costs. Especially in a startup context, it is often more effective to postpone optimization to a later stage. A good architecture is not one that is perfect from the very first release, but one that can evolve alongside the product.

Optimization should be driven by data. Without performance and usage measurements, there is a risk of acting on non-critical features. Not all functionalities have the same impact. Optimizing the wrong areas is a waste of time and money for a digital product.

Solutions don’t have to be perfect, but functional. The ideal solution is not always applicable, whether due to feasibility or cost constraints. Often, a solution that works is all that is needed. Only later, with a larger budget and when the product requires it, can a different approach be adopted.

Moments like the one we shared with university students are valuable to us, as they allow us to convey our approach to software development: it’s not just about writing code, but about designing digital products that can grow, adapt to circumstances, and generate value over time.

We sincerely thank Ca’ Foscari University of Venice, the students who participated, and Professor Andrea Marin for giving us this opportunity. And if you’d like to hear us speak about the topics we’re passionate about—whether in your classrooms or your offices—we’d be happy to contribute. Get in touch using our contact details or the form available on the dedicated page.