Outra: 2023-2024

---

Outra is a data agency utilising analytics, machine learning and AI to sell insights and enriched property data to external clients.

Challenges

Their main challenges were poor delivery processes and execution, inconsistent data quality and huge technical debt which all lead to high levels of resource churn.

Task

Engaged as the Product Manager to implement agile methodologies to improve delivery, increase data quality and reduce technical debt across the platform services which underline their product suite.

Actions

• Successfully implemented a Servant Leadership model within an agile framework which built on the existing talent of the data science, engineering and analytical team members in order to create the environment for us to succeed. 
• Introduced quarterly SAFe PI-planning based strategic roadmapping, delivery and Scrum/Kanban ways of working in order to facilitate scale up, stabilise, optimise and formalise the overall business process
• Managed introduction of Airflow DAGs to enable automated orchestration of workflows plus monitoring dashboards and automated email alerts to overcome historic dysfunctional processes and manual resource pinch-points
• Managed refinement of raw data ingestion process from multiple sources and the address matching logic which dramatically improved the time needed to match, confidence scores and the match rates from 53% to 72%
• Managed data cleansing and data pipeline refinements which trained the data models 
• Managed shift towards DBT meant integrating seamlessly with CI/CD pipelines, enabling automated data testing and deployment consequently improving data quality and delivery of our client data exports
• Stakeholder management and reporting, crisis management

Results

• Delivery was incrementally every fortnight but intrinsically part of a cyclical quarterly three-month release schedule. This regular predictability enabled us to deliver quality over a longer period but also enabled us to pivot without the hinderance of long-term commitment. 
• Moving away from constantly firefighting meant that we had the time to develop and refine orchestration and automation which improved the end-to-end data quality and reduced technical debt. 
• Because the team were trusted and empowered to solve the main technical problems this increased employee engagement exponentially, so resource remained stable rather than churn like other teams across the business.