Travis Perkins: Democratizing data analytics from the grassroots up

About Travis Perkins

Travis Perkins plc is the United Kingdom's leading supplier of materials to the building and home improvement markets. The Group operates 20+ businesses from over 2,000 sites across the United Kingdom, employing over 28,000 people.

Industries: Retail & Consumer Goods
Location: United Kingdom

About Matillion

Matillion transforms a business' data, across its various locations and forms, into cloud data warehouses to enable informed decision-making.

Travis Perkins, the United Kingdom's leading supplier of materials to the building and home improvements markets, changed the way its employees use data, with an analytics platform built on Google Cloud Platform.

Google Cloud Results

  • Halves the time required to bring new code into production, according to Travis Perkins developers
  • Enables and empowers business self-service data visualization and analysis by all employees, whether specifically trained or not
  • Improves reliability of data, with a single version of the truth
  • Accelerates project delivery to weeks rather than months

Replaces legacy sales analytics to save £250K a year

Travis Perkins is the United Kingdom's largest builders' merchant, supplying everything people need for construction projects, from flooring, to workwear, to plumbing supplies. More than 28,000 people work for the company, which comprises over 20 businesses nationwide, including builders merchants, Travis Perkins, and Wickes; a leading home improvement brand.

"Lots of our staff are not based in offices," says Steve Taberner, the company's Head of Business Intelligence. "But almost everyone receives Business Intelligence (BI) content in some shape or form, whether it's a pricing report or an email about customer trends. BI is a group function, so we support all of the major businesses within our group, helping them to use their data to support decision-making at all levels of our organization."

"The beauty of BigQuery is that we could start small and grow, without large upfront costs. That made it easy to do proofs of concepts, prove to people that it would work, get support for the solution, and grow data utilization."

Steve Taberner, Head of Business Intelligence, Travis Perkins

In the past, BI was delivered to the company by around 500 analysts using an on-premises, data warehouse solution. This process could be slow, with some analysts spending 80 percent of their time extracting, moving, and combining data, instead of conducting analysis. That's why Steve and his team decided to create a data analytics platform in the cloud, with accessible tools to help democratize data analysis, so it could be used by people across the whole organization, instead of a select few. To do that, they created a proof of concept (POC) with BigQuery.

"The beauty of BigQuery is that we could start small and grow, without large upfront costs," says Steve. "That made it easy to do proofs of concepts, prove to people that it would work, get support for the solution, and grow data utilization."

Democratizing data with faster, more accessible tools

As large companies look to leverage data to improve their services and enhance efficiency, the availability of that data for analysis is a key concern. At Travis Perkins, it was a key objective to improve accessibility and help ensure a greater proportion of the team would be comfortable using its data analysis tools.

"The tool we were using would become slow when processing complex queries, and it was very difficult to add new data or merge data sets in a quick, efficient way," says Steve. "Users would choose to export data to spreadsheets instead, where they were more comfortable arranging it into a usable format and manipulating it for analysis. That took time that could have been better spent analyzing the data directly. More importantly, it raised questions about how trustworthy the data was, because of all the transformations and manipulations it had been through."

Cost was another consideration as Travis Perkins looked for a faster, more responsive, more cost-effective solution that could embed data analysis into decision-making at all levels of the organization.

To do that, Steve and his team created a cloud data analytics platform using BigQuery, Cloud Dataprep, and Google Data Studio. "Google Cloud offered one of the most powerful and progressive Big Data solutions available," says Steve. "It also offered integrated tooling that made it more accessible. We already use G Suite, and we are seeing more and more how Sheets can connect with BigQuery for more seamless integration across our corporate tool set."

Because BigQuery is a managed service based in the cloud, Steve and his team could set up the solution without either the expense of setting up servers or the maintenance burden of keeping them online. "It's very important to us that the solution is serverless, because it means we can focus our resources on analysis, data preparation, and benefiting the business, instead of administering environments," says Steve. Fast uploads and high availability on BigQuery also mean the data warehouse is there when it is needed, with recovery times reduced from one day to under an hour, when the team needs to reload data.

"We have heard from senior business stakeholders that the whole team has benefited from a single version of the truth. This means we can now trust the data and focus on what the analytics tell us. That's a massive improvement for our operations."

Steve Taberner, Head of Business Intelligence, Travis Perkins

The company's existing ETL solution was complex, so the team looked for a more agile alternative, ready for use with the data lake. "We explored our options and Matillion stood out," says Steve. "It ticked all of our boxes in terms of focus and agility. It uses SQL, and we have a strong basis of SQL capability, which made it easy for us to adopt it. Because it is based on a Compute Engine instance, all of the processing is done within Google Cloud Platform (GCP). One developer said to me: 'This is approximately twice as fast for me to code in and get to a state of production compared with the old tools.' As a result, we saw a massive increase in our ability to deliver new data sets."

On the frontend, Data Studio offered a simple tool people could use to access the new cloud data resources. "Data Studio is very collaborative and easy to use, which really helps to get people in the business excited about using data in an interactive and agile way," says Steve. "Because it's a free browser-based tool, it helps us leverage the data in BigQuery really quickly. It doesn't require training, almost everyone in the organization can pick up Data Studio and create a nice little dashboard without any sort of coding ability."

The first POC built on top of the cloud data analytics platform created reports on the effectiveness of deliveries to customers. It proved a success and rapidly went into production, remaining in use with Travis Perkins companies today. The next project was to deliver a supplier portal for Wickes, providing information such as sales data, stock levels, and the status of purchase orders. "It gives suppliers information to help them help us sell more of their products," says Steve. "It's a win-win and funds the BI solution."

Behind the success of the cloud data analytics platform and all of the projects built on top of it, is confidence in the data it provides. "We have heard from senior business stakeholders that the whole team has benefited from a single version of the truth," says Steve. "This means we can now trust the data and focus on what the analytics tell us. That's a massive improvement for our operations."

Growing grassroots support through hackathons

Following the success of the proof of concept and the Wickes project, the Travis Perkins BI team focused on promoting grassroots engagement with the cloud data resources. Central to this strategy is an ongoing series of hackathons — 15 of them, so far — each with between 5 and 12 participants.

"We pull together people from across the Travis Perkins Group and facilitate a day to investigate, and play with, the Google cloud tools," says Steve. "We focus on Cloud Dataprep, showing them how to add data from a CSV file to BigQuery, and on how to use Data Studio to analyze it. Usually, by lunchtime, we start to see dashboards appearing with visualizations of data. We can see the reactions in the room as people realize how accessible the data is, and how easy it is to create quite sophisticated analytics, with minimum training."

"Hackathons have completely revolutionized the way we think about data as a business. In the past, everyone relied on the IT department for its end-to-end analytics delivery. As word of mouth spreads that colleagues can do this themselves, our employees are driving more value from their day-to-day work."

Steve Taberner, Head of Business Intelligence, Travis Perkins

That engagement and enthusiasm is filtering back into the company as stakeholders share knowledge and show their colleagues how to self-serve data analytics. As new initiatives from that growing grassroots support are presented to senior management, managers are encouraged to invest in the cloud data analytics platform and make use of it in their decisions.

"Hackathons have completely revolutionized the way we think about data as a business," says Steve. "In the past, everyone relied on the IT department for its end-to-end analytics delivery. As word of mouth spreads that colleagues can do this themselves, our employees are driving more value from their day-to-day work."

Generating major efficiencies with applied data

Thanks to the new solution, Travis Perkins has both accelerated data projects and generated significant savings. "The time and resources we spend on data projects has changed totally, thanks to BigQuery, Cloud Dataprep, Data Studio, and working with Matillion," says Steve. "Instead of spending nine months on a project, we recently did a project in seven weeks that delivered a massive data set for sales analysis. In addition to the savings this generated, this meant we could also switch off a legacy solution that cost hundreds of thousands of pounds a year to run."

"Among our early deliveries was a comprehensive daily sales capability, delivered first thing every morning to the company executive. That's part of step change in the way BI is regarded within the business. When we talk about projects now within BI, we don't talk in millions, we talk in tens of thousands of pounds," says Steve. "That's mainly because we can use BigQuery to present data in a flattened format to the business without additional staging and manipulation."

Now the BI team is looking to expand their cloud data analytics platform, aiming to offer more services to customers through new projects, including a pricing engine built with machine learning tools.

"We love how GCP products constantly improve," says Steve. "We actively talk to Google about the data products, and it's great to feel that we are listened to, and even have some influence over that evolution. Features that we've asked for in the past suddenly appear as if by magic. We're not the biggest organization in the world, but we feel that our voice is being heard."

About Travis Perkins

Travis Perkins plc is the United Kingdom's leading supplier of materials to the building and home improvement markets. The Group operates 20+ businesses from over 2,000 sites across the United Kingdom, employing over 28,000 people.

Industries: Retail & Consumer Goods
Location: United Kingdom

About Matillion

Matillion transforms a business' data, across its various locations and forms, into cloud data warehouses to enable informed decision-making.

Google Cloud Platform logo

12 Months FREE TRIAL

Try Kubernetes Engine, BigQuery, and other Cloud Platform products with $300 in free credit and 12 months.

TRY IT FREE
Google Cloud Platform logo

12 Months FREE TRIAL

Try Kubernetes Engine, BigQuery, and other Cloud Platform products with $300 in free credit and 12 months.

TRY IT FREE