Making data everyone’s business

Making data everyone’s business

Key takeaway.

Here at GSoft, we’ve made it our mission to turn our data culture into a veritable engine for our product iterations, processes, and ideas. How? By making it everyone’s business.

Author :


If data once seemed like exclusive fodder for IT and tech companies, today it plays a central role across nearly all fields and industries.

In fact,  Officevibe’s Director – Data, Analytics & User Research, Élise, would argue that, in many ways, data plays a fundamental role in the very way we all assess and navigate the world.

“We consume data on a daily basis. So why would it be any different at work? But in the context of a tech company like ours, data is an unprecedented tool that allows us to better understand each other, understand our clients, and make better decisions day after day.”

Élise St-Aubin Fournier
Director – Data, Analytics & User Research at Officevibe

Companies are giving that much more importance to data in their approach to strategic decisions. But in our experience, even in tech companies, having a data team doesn’t always mean having a data culture. In fact, data expertise is often seen as a separate entity to the everyday business of building products or solutions; an external force periodically consulted for insights or quarterly reports. Yet the data sciences – a grouping of several different tools and techniques used to extract relevant information from raw data – can be an essential decision-making tool. Here at GSoft, we’ve made it our mission to turn our data culture into a veritable engine for our product iterations, processes, and ideas. How? By making it everyone’s business.

Turning data sciences into unforeseen opportunities

To us, a data culture doesn’t mean that everyone in the organization has a deep-seated understanding of AI or the advanced models of data analysis; it means that everyone understands the opportunities those frameworks can create, and has the skills required to collect and analyze the data themselves.

The big challenge of data analysis is to produce value-added products. That’s why our team supports other teams in analyzing their data so that they can deepen their understanding and learn to identify interesting opportunities on their own.

Élise St-Aubin Fournier
Director – Data, Analytics & User Research at Officevibe

That commitment to education and ownership means the data team has their work cut out for them. Not only do our data scientists and engineers take on their own projects, they also play a supporting role in the collection and analysis of data from other parts of the organization. Throughout the entire company, the data team works to empower employees to understand how data can guide and inform their product development, their client relationships, their metrics for success, and their own internal processes.

But they’re not the only ones to inspire new data projects. Within the team responsible for Officevibe, the data sciences team makes sure that new initiatives are guided by the needs of other departments. In general, these new requirements might include the following:

Could data help to sort our clients into categories according to their behaviour within the app? (sales team looking to better understand its clients)

Could it give us insights into what kinds of Officevibe survey questions correlate most with management performance?

Could our data tell us at which points in our onboarding funnel users are most likely to drop out?

The answer: **yes, yes, and yes.

Turning data concepts into tangible applications

This, for us, is where the rubber meets the road; where pivotal data science principles allow us to connect new dots, build new functionalities, and ask different questions. Here are just a few concrete examples of how data is used by our Officevibe team:

Predicting client behaviour

A survival analysis model allows us to predict the likelihood that an event X will occur in the future. This type of data analysis is often used in medical research to predict the effect of a drug on a patient’s chances of survival. Applying this model, we were able to give some of our teams (customer experience and sales teams, among others) the ability to better understand various types of client profiles.

Anonymizing and categorizing client feedback

In the context of the Officevibe application, which, among other functionalities, allows managers to gather employee feedback, the data team needs to make sure that the feedback is anonymous so that it can be used for internal purposes. The challenge is developing an algorithm that can remove the names of people and companies or other information that could identify someone (“director of accounting,” for example) without impacting the substance of the employee feedback that was gathered.

To tackle this problem, the data team has developed a named-entity recognition model to ensure anonymity, and added an NLP (Natural Language Processing) module in order to categorize the different types of feedback based on the themes they contain.

Result: the Officevibe teams were able to use the feedback they collected (anonymized and categorized) to extract relevant information in order to create marketing content that managers find relevant and improve certain product functionalities.

Transforming data into … even more opportunities?

Within the Officevibe team, as is the case with many organizations, data informs the user experience as much as it does our internal processes. Data encourages debate, calls us to question ourselves, and makes us more aware. That said, without a well-entrenched data culture to which the teams adhere rigorously, it is difficult to access the full potential of data to improve the user experience. A data-centered vision, such as the one in place at Officevibe, serves to clarify the place that data has within the organization, as well as everyone’s role within it.

Data analysis does not always give us the answers we are looking for, but it does help us to ask the right questions, ones that assist us in moving forward. The opportunities here are far more accessible than one might think, as long as data remains front and centre among other priorities.