top of page

How to reduce queue sizes at the till area?

Sensor based customer flow data analytics

Improving CX by using digital tech at Wembley IKEA store

At the Booth

Michele Usuelli

Lead Data Scientist at Microsoft Consulting Services, designing and leading advanced analytics projects for customers and internal stakeholders.
Author of two books about machine learning and host of workshops and training classes, having diverse audiences such as commercial stakeholders, business analysts and technical specialists.
Previous hands-on work experience designing and delivering advanced analytics solutions using tools like R, Python, Hadoop, SQL/NoSQL databases, cloud ecosystem on Azure.

LinkedIn profile: https://www.linkedin.com/in/michele-usuelli-1b84b460/

Bernd Heßbrügge

Bernd Heßbrügge is Architecture Manager at IKEA, having a great interest into that area.

Accompanied several analytics projects in the past from concept to implementation, ranging from BI to advanced analytics, including structured/unstructured data, advanced and continuous analytics, using Hortonworks (Azure HDInsight) commercial packaged Hadoop ecosystem.

LinkedIn profile: https://www.linkedin.com/in/bernd-hessbruegge-589691151/

Speakers
Background

Background and Objectives

Background and Objectives

The business question from IKEA Store Wembley:

How can we understand and act proactively on queue building in our store till area?

Strategic partnership between Microsoft and IKEA, here related to the impact of Cognitive Services and Artificial Intelligence with the objective to

  • Understand positive impact of digital technology on both people (co-worker and customers) and processes

  • Test and learn the Microsoft Data Science Methodology and Data Platform

  • Build a demo to predict the queues in IKEA stores to have actionable insights

Explorative engagement of IKEA with ModCam, here related to the use of sensor devices in IKEA stores to

  • Capture unpersonalized customer flows and volumes in IKEA stores

  • Joined explorative test & learn approach together with Microsoft to use customer flow data in customer queue analytics

  • Visualization of customer flows and volumes at different areas in IKEA stores to sales co-worker

  • Providing customer flow and volume data to queue prediction analytics

Example API call return result visualized above

Customer count visualization for cashline 7, 10 and self-service

Demos

Approach
Exploration

Exploration

Exploration

Path to value of the queue management use-case

Approach and Methodology to Data Science

Key phases to develop and test a product

Microsoft iterative methodology to develop projects

DEMO

Findings

Findings

Performance of the models

Way Forward

Way Forward to build a first MVP

Suggested Approach

  • Work with UK SO to further refine the success criteria

  • Create SOW with Microsoft and Modcam to understand the intended scope and responsibilities

  • Extend Modcam cameras to all 20 cash lines in Wembley store

  • Setup agile squad team to create and deploy first MVP at Wembley store

bottom of page