Early in 2017, leading auto parts corporation LKQ started conversations with the NICD about a project that would help it discover valuable insights in its company data. A little over a year later, thanks to a successful collaboration, it is planning to invest in a dedicated team to explore advanced analytics.
This is the story of how it made the project work, and changed minds about the power of using data more intelligently.
We’ve taken the business on an important journey. It’s not just about the revenue. The excitement is in taking the whole management team on that journey by showing them the prototypes, and explaining what we can achieve if we’re open to it
LKQ Corporation is a global leader in providing specialty parts used to repair and accessorise automobiles and other vehicles. Formed in 1998, it has expanded through the acquisition of more than 220 suppliers and manufacturers across the world. The company is now a dominant figure in its market, with operations in North America, Europe and Taiwan.
LKQ is a global company, with a growing network of stockists, manufacturers and self-service retail businesses. Each of these operations retains data which could have value for the business as a whole, from increasing revenues to streamlining supply chains.
LKQ was keen to explore how strategic data analytics could suggest improvements to the way the company operated, challenge longstanding assumptions, and reveal new opportunities. It established a six-person team to explore the potential of selected data sets from within the company. The team was based out of the Newcastle Helix, a hub for international science and technology innovation in the North East.
In the space of a six-month project working alongside the team at NICD, LKQ’s small team:
- Identified “substantial” potential savings and sources of additional revenue
- Encouraged management to invest in a dedicated data analytics unit
- Gained insights on how to transform data quality across the company
- Shifted the company toward a more data-led mindset
Moving from anecdote to data
Over the last two decades, LKQ has completed a series of acquisitions which has established it as a world leader in its field. The company now benefits from the knowledge and experience of managers and business owners that know the market inside out.
However, advances in data analysis provided the company with an opportunity. If LKQ could collect high-quality data from across the company, it could mine it for insights that would maintain its position as a dominant player in the industry.
This data could also ensure that LKQ was making investment and growth decisions based on the best information.
“If you want to shift from anecdote, it starts with the data”, says Marcin Lisowski, Head of Data Analytics at LKQ Europe Limited. “But that can only be as good as the data you’re putting in.”
In early 2017, an initial meeting took place between senior management at LKQ and representatives of the NICD.
NICD agreed to collaborate with LKQ’s team on a six-month proof of concept project in Newcastle, providing guidance and experience as well as the hands-on assistance of a PhD student.
The project began in August 2017.
The ability of the whole team to deliver against the vision has been stupendous. All of the various elements came together in alignment. NICD, ourselves and everyone involved should be very proud
The team challenged with exploring the potential of LKQ’s data was not large. It featured one employee from its India team, one from Italy and three from the UK. Profile Analysis chief scientist Alan Timothy – who had worked with the company on a number of projects – was also invited.
“Previously, there had been some data analysis, but it was no one’s full-time job. No one ever got the bandwidth and the six months we got to dig, play and draw on the exceptional knowledge base we have”, says Alan.
The team devised the methodology, and identified 40 potential analytical streams of operational, sales, customer and financial data. From this list, the team identified six that had potential, based on the type and quality of data. There was no pressure to pick a particular data set; merely to choose the sets that offered the clearest results.
“The business was very open to what the benefits were”, says Marcin. “We were working within a leader in the market, but we knew this could add even more value to the business.”
Preparing useful data
The most important task in preparing for the project was making sure that the data was ready for use. Between April and July, the LKQ team had regular meetings with data experts at NICD to explain its thinking, and work on ensuring the data was cleaned and validated. This process continued into the first month of the project.
“Having cleaned it, we statistically validated what we had done”, says Marcin.
“When we went back to the business with a solution based on a data set, we didn’t want someone to query whether the data set was wrong. In that sense, the role of Newcastle University in doing the statistical validation on the data was vital.”
The original data was of varying quality. However, the team was able to convince management that it was taking a properly structured approach and build trust in its methodology.
“Only a small fraction of big data initiatives tend to succeed”, says Alan Timothy. “One factor is the data quality you have to work with, and another is the change management challenge, which can often rely on appreciation of the analytics themselves.”
A place to focus
Instead of working out of one of the LKQ offices in Europe, North America or elsewhere, the team co-located with NICD specialists in a space at Newcastle Helix. This allowed it to be within easy reach of expert help and gave it space to explore new avenues.
The team worked in Newcastle five days a week, collaborating in sprints with NICD, PhD students and University staff.
"You’re able to experiment in your own space”, says Marcin. “The environment enables you to test what something new might look like, and took us out of our normal business mindset.”
A helping hand
From the early stages of the project, the team was assisted by a PhD student from the EPSRC Centre for Doctoral Training in Cloud Computing for Big Data who helped to validate data and offer expertise in Statistics, Mathematics and Computer Science.
"They brought very useful skillsets to the team. While we have an understanding of statistics in the business, it is not at that level”, says Marcin.
“Having the student working with us was about asking questions both ways, and investigating them. We’re now recruiting, and through this we’ve developed a working relationship with people who have PhDs in Computer Science. They know us, and we know them.”