Data Science and Digital Performance Driving BIXI's Growth

Context

BIXI's bike-sharing system has never stopped growing since its launch in 2014! Twelve years in, the non-profit organization now operates a network of over 12,600 bikes and more than 1,000 stations across 13 Quebec cities, including Montreal, Longueuil, Laval, and Sherbrooke.

This remarkable growth is in large part the result of a major strategic shift in acquisition and digital performance, carried out between 2021 and 2024. At the time, the organization was looking to better recognize and track its customers across various digital touchpoints, in order to invest its marketing budgets in the right places and maximize acquisition results.

Mandate

To help optimize its acquisition campaigns, BIXI entrusted the Click & Mortar team — now part of Tink — with a two-part mandate.

1. Data science: centralizing data from multiple sources and linking it to a unique BIXI account ID to better understand the customer journey and identify which marketing efforts are actually driving results.
2. Digital performance: leveraging that data to deploy more targeted acquisition and retargeting campaigns across the entire conversion funnel.

The synergy between these two areas of expertise made it possible to turn raw data into concrete, measurable acquisition levers.

Objectives

• Significantly improve BIXI's user recognition rate across various digital channels
• Implement a custom attribution model to accurately represent the real causes behind conversions in acquisition campaigns
• Develop a predictive customer churn model to automate and refine retargeting strategies
• Surpass previous seasons' results in terms of app downloads and subscription purchases
• Define and deploy an integrated client acquisition system covering all of BIXI's marketing initiatives, at every level of the conversion funnel

Achievements

Data Centralization and Custom Attribution Model

For this first phase of the project, we built a data centralization structure that linked marketing attribution to a unique BIXI account ID. This effort provided complete visibility into the customer journey, from the very first interaction all the way to conversion.

Building on this infrastructure, we designed and implemented a custom attribution model specifically tailored to BIXI's reality. Notably, this model gives more weight to the top of the funnel compared to classic attribution models, offering a more accurate picture of the role each marketing channel plays in the acquisition process.

Predictive Customer Retention Model

Our data science team also developed a predictive model based on BIXI users' buying behaviours. This model identifies individuals at risk of not renewing their subscription and automates retargeting actions directed at them. The customer retention team can then intervene proactively, at the right moment, with the right message.

Acquisition and Digital Performance Ecosystem

On the digital performance side, we defined and implemented an acquisition system fully integrated with all of BIXI's marketing initiatives. This system covers every level of the conversion funnel through a structured behavioural strategy, including Meta campaigns and newsletters optimized using attribution data. This approach made the retargeting ecosystem significantly more precise, directly contributing to the overall performance of the acquisition campaigns.