De-risking innovation for a global commercial business
Big data.
It’s a trendy buzzword that’s supposed to be the solution to all of our problems.
But how does a multi-billion dollar business leverage years, decades even, of their data to derive business value?
And how would they apply it?
Service design, business & data analysis as part of
Shell Agile Hub’s Advanced Analytics team.
Jun 2019 - Feb 2020
How might we de-risk high value and innovative business opportunities by leveraging data?
The Role
Conduct user interviews and design user journeys to provide context to business problems
Understand users’ requirement and interpret business data to propose service improvements to meet their needs
Develop ideas using sketching, prototyping, and iterative design
Collaborate with cross-functional team to conduct data and business analyses
Manage cadence of team meetings with scrum practices and ceremonies
Manage library of process documentation for business and user requirements, research materials, and artefacts
The Challenge
Over the course of half a year, General Managers from Shell’s downstream business (read: lubricants) (like, motor oil) shared use cases from within their business units that had unique opportunities to be improved through the use of data. The brief given to them was simple:
What areas within your business can be optimised through advanced analytical analysis?
My team consisting of a data scientist, a data engineer, and a business analyst (myself) would work in sprints to then figure out if there was any way we could utilise the data they shared whilst also understanding the business context, opportunities present, and significant blockers, if any. This was the discovery work before the discovery work — the philosophy was if we couldn’t find significant potential value in the data, the use case was not a priority.
The use cases and data varied from analysing the ordering patterns of US customers in frequency and volume to predict their next order, to optimising the supply chain in India to optimise ‘on time in full’ (OTIF) success rate, and measuring the price elasticity of products in Thailand. Due to the global nature of Shell’s downstream business, my team was constantly working with, interviewing, and demonstrating our work to teams from all over the world.
The Outcome
Through data analysis and service design methods, my team at Shell Agile Hub’s Advanced Analytics team managed to unlock value through rapid testing and prototyping of innovative business use cases.
Predictive modelling methods showed the business potential gains of:
60% increase in customer stickiness
48% value gain through cross-sell and up-selling
35% improvement in supply chain operational efficiency
$20m in additional yearly profit
The Process
Work redacted for NDA.
01. Hypothesis generation
The sprints would follow the classic agile and scrum structure: cross-functional and self-managing teams, rapid iteration cycles, and daily stand ups to share learnings and decide as a team whether to pivot or persevere.
We called these investigative sprints ‘Exploratory Data Analyses’ — or EDA for short. Each EDA would have predefined objectives in the form of hypotheses, value levers, and dedicated subject matter experts that we would touch base with every few days as we dug deeper into the data.
02. User research
Since EDA teams were lean (most of the time only 2 people at a time and never more than 3 people) and time was never enough, I utilised Service Design methods to map out the end-to-end service by conducting interviews with the subject matter experts.
We set a cadence with the subject matter experts to check in with them every 2 days within the 2 week sprint, sharing our progress and concerns — usually to do with the data. Within this time, the Data Scientist on the team will work their magic and run test cases with the data based on hypotheses generated before running the sprints.
03. Synthesis of insights and playback to GMs.
At the end of the 2-week sprint cycle, findings (both qualitative and quantitative) are gathered and insights are synthesised for a cohesive story. Though each use case came from different parts of the business, the end goal was the same: de-risk innovation by rapidly testing a small sample of business data.
My team would then share our findings and recommendations to the respective General Managers on areas that could unlock business value and what steps to take to get there.