High Performance Analytics Teams: Retail
At a glance
We helped improve the quality of interactions between the business and Ai teams and help them ask better questions in order to enhance their understanding about customer dynamics.
The atomization of (customer) knowledge in the business is trapping value. We unlocked this value by fostering a leadership behaviour change towards ‘a duty to share’.
AI and analytics teams feel like ‘stick fetchers’ for the business; hastily gathering reports rather than acting as team members “on the bridge”. We changed this by creating “Messy Teams” - a high performing framework of interaction and capturing all perspectives.
The Challenge
We worked with the group analytics Centre of Excellence (COE) in one of the largest retailers in the UK to build a high performing AI & analytics function.
Fundamentally, this was about improving the quality of questions that came from the business and connecting them to an analytics solution, leading to a better understanding of the customer in a highly complex setting, and breaking silos of insight across multiple business stakeholders.
A Complex Customer View
The Chief Customer Officer and senior leadership at large demanded two things; enhanced understanding of the customer and drive for simplicity within the business to help them make faster, more effective decisions. The challenge for the analytics team was that leadership talked about and viewed the customer in a multitude of ways, and as such, when it came to connecting this to ways of working with the analytics function and harness the collective power of leadership, it was proving to be a near impossible task. The analytics team was running back and forth fielding requests that whilst adding short term value, failed to realize the potential value to increase broader customer understanding and fuel the required adaptability and foresight for the business, critical to winning in a complex retail environment.
One Language
Below is a ‘simplified’ illustration of how leadership in diferent parts of the organisation viewed and talked about the customer
You could walk into any business team meeting and hear one of the terms on the outer circle that the team referred for their customer strategy. What was clear form the onset, was that all of this work was driven by positive intent and by some experienced and talented leaders. Added to this picture was of course the fact that customers are dynamic and all interact with different brands in different ways at different times.
The organizational response to this complexity was to attempt to simplify the way they talked about the customer, reverting to a single persona. Whilst this satisfied communication efforts across the business, the hands of the analytics team were tied as they struggled to gain the required clarity for the question formulation process.
All the potential value creation of AI was left on the table as the organization failed to deal with the complexity it faced. The challenge was culture and ways of working with AI
Building understanding & creating competitive advantage
During the discovery phase of the project, we addressed with leadership the topic of orientation. It has been our experience at Pelatum that too often executives use data to support a decision that has already been made.
In a highly operational environment that was retail, it was a mindset shift from “What do we need to do” towards “what do we need to know?”.
This was a shift from a classic retail mind-set of efciency, hunch and instinct towards that of data driven teams. We aspired towards a culture that connected experience and insight to formulate better questions.
Building A High Performing Analytics Team
The Chief Analytics Officer knew two things; Firstly, analytics was going to help business leaders make better decisions against this very complex customer picture and secondly, with the right high performing ways of working, the analytics COE team could add huge value by connecting insights across stakeholder groups, break silos and ‘fuel the collective intelligence of the organization’. Starting with each member of the analytics team and the business teams that they supported, we ran a series of workshops to align against a set of high performing skills.
This allowed the analytics centre of excellence to run at a faster pace. Using a common language and framework, analytics team members could share best practice, run effective peer reviews, share insights and most importantly, improve the quality of decision making in the business using AI.
5 skills of an AI augmented team
ONE: Align analytics to critical business challenges
Business Leader: Frame the challenge. Clarify end state objectives. Confirm essentials of the goal, key themes and messages.
AI Expert: Scrutinize (own) situational understanding. Check assumptions and information. Offer alternative explanations.
Business Value: Develop wider understanding of the business challenge facing domain and analytics teams. Closes the cognitive gap between analytics and domain expertise.
TWO: Articulate focused questions to data
Business Leader: Clarify gaps in your understanding of the business challenge.
AI Expert: Lead the development of accurate and relevant AI questions.
Business Value: Trawler fishing to spear fishing… we need to use AI in a way that is going to improve decision making and drive value for our customers.
THREE: Combine analytics with business experience
Business Leader and AI Expert: Challenge ideas and theories. Apply contrasting views. Test hypotheses. Identify flawed assumptions. Identify faulty logic & inconsistencies
Business Value: Closes the cultural and semantic gap(s) between Domain and AI expertise.
Robust baseline for further analysis and improved end (data) product
Alternative outcomes identified.
Strengthen decision making with constructive data driven feedback and insights.
FOUR: Follow-through on the data
Business Leader: Reallocate resources and refresh focus to team.
AI Expert: Communicate & inform to analytics COE. Request further support / insights from COE. COE AI peer reviews. Peer coaching.
Business Value: Unlock collective intelligence. Break atomization of knowledge
FIVE: Share with the AI ecosystem
Business Leader: Translate decision to team and decision makers.
AI Expert: Translate insights to analytics COE to enhance AI ecosystem.
Business Value: Expands organizational understanding by allowing further analysis from the AI COE and further refining of key data regarding business value creation.
From reporting to better understanding…
Retail leaders have been brought up in an environment of operational reporting around Business Intelligence and visualization tools.
We were clear that we had to maintain this capability and ‘maintain operational stability’, but also re-orientate leaders towards a different way of working with advanced analytics.
This new orientation was about looking forwards and not backwards and seeking to understand what was going to happen next. We embedded a simple framework for the analytics team(s) to have these conversations with the business as part of their meeting rituals.
With permission from senior leaders opening the workshops that we ran - sending a consistent message, the high performing analytics team created a safe environment to draw out ideas and discovery with their respective business partners.
The value here was to connect the hard earned experience of retail leaders with the skills of the analytics team by changing the way that they worked.
Seeing Around Corners
The analytics approach to answering these questions varied, but the real value was in the team coming up with better questions to ask and building trust in the team
Summary
Through building a high performing analytics culture inside our client, we were able to achieve value in the following ways:
- Generate and improve company wide demand for the services and resources of an analytics centre of excellence by asking better questions.
- Bring valuable questions to the forefront of the analytics COE and innovate between analytics peers.
- Provide a framework for analytics and business team interactions.
- Leverage hypothesis-led thinking and provide structure to higher quality collaboration.
- Encourage members of various departments and analytics to collaborate for shared outcomes.
- Develop the right questions to reduce the actionable-insights gap.