12+ years making digital products and experiences more exciting, more usable, and more human.

Cousteau Data Mining Tool

Industry 4.0 Initiative, KD Innovation Hub

 
 

Situation

Koerber’s Technical Engineers travel all over the world fixing highly specialised industrial machinery. Data on machine fault symptoms, problems and root causes are under-utilised, and as the most highly trained technicians retire, the repair times are growing longer.

Task

Lead the research, development and design of new digital tool to support the technician’s machine interventions. Build an MVP, gather feedback, & establish a product roadmap.

 
 

Design Process

 
 
discussion2 copy.jpg

01. Problem Discovery

To better understand what factors slowed or expedited machine interventions, I travelled to Italy and held user-interviews with 11 technicians and department service personnel. This investigation allowed me to deeply understand the machine service process.

 
Artboard 1 copy.jpg

02. Synthesising Findings

To get clear insights by distributing painpoints uncovered in the user interviews on a time-line. Produce personas, an user-experience map and a service blueprint to capture insights. Communicate findings with stakeholders and project team.

 
Artboard+6.jpg

03. Solution Ideation

I held lightning demos to showcase other state-of-the-art tools that help surface meaningful insights from data sources.

Using a “How Might We” list created with the technicians, I then facilitated rounds of crazy 8s design sketching sessions. Using dot-voting to refine ideas, we established a core concept.

 
Artboard 2.png

04. Refining Concept

To illustrate the new user experience of a data-mining research tool, I created a storyboard and wireframes to demonstrate the functionality associated with key user-stories.

I used these tools to demonstrate the concept and get initial feedback, and refined the concept based on these outcomes.

 
interview.jpg

05. Prototype Testing & Iteration

Next I designed a prototype that surfaced results using an ontology of terms related to symptoms and root causes.

Upon testing, the technicians expressed their need for more contextual machine data. By cross-referencing report meta-data with external machine logs, I retested a revised prototype with more favourable results.

 
Screenshot 2019-11-11 at 12.52.22.png

06.

Designing an MVP

I built on the prototype to develop an MVP. I utilised Google’s Material Design as a UI foundation, to design 3 functionalities; Smart Search & Filter, Line Service History & Machine Base Review

The MVP was built in short release / test / iterate cycles over an 8 week timeframe.

 
cousteau mockup small.jpg
 

Project Outcomes

The MVP launched in February 2019, and usage of tool was monitored over a 3 month period. Technicians were then interviewed to get qualitative feedback:

100% said the tool was an improvement to the current system for reviewing reports.

70% of technicians reported improved repair time

Learnings

Changing how the technicians wrote initial data proved too much too soon and was dropped due to resources.

Technician’s view of themselves as experts meant resistance to the idea of getting help from other sources like the tool. 

Corporate centralisation of innovation needs to be handled sensitively to avoid “not invented here” syndrome at BAs.