Warning Signs: Inside Travis Perkins’ Failed ERP Migration
November 3, 2021 Alex Woodie
When UK construction supplier and home improvement retailer Travis Perkins announced last month that it had selected Oracle Fusion Cloud’s Financials as the core of its new ERP deployment, it marked the latest in a long series of decisions about the company’s computing infrastructure. Unfortunately, it came after many poor decisions and failed ERP migrations along the way. Hopefully they will serve to educate others who might travel the same path.
Travis Perkins’ current ERP woes began in 2016 at Infor’s annual Inforum user conference in New York City. The UK company’s then CEO, John Carter, announced that it would be migrating various internal business applications systems to the Infor CloudSuite version of M3, the Java-based ERP suite that it obtained with its 2011 acquisition of Lawson (which in turn had acquired M3’s original developer, Intentia, in 2006).
According to the companies, Travis Perkins would eventually license 24,000 seats of Infor’s CloudSuite offering. Over the 15-year lifespan of the deal, it was expected to be worth $200 million for Infor, making it once of the biggest deals the ERP giant had done in the cloud. “It’s the largest cloud ERP transaction in this game,” Infor CEO Stephen Scholl reportedly said at the time.
Travis Perkins, which enjoys annual revenues around $8 billion, certainly was motivated to move to the cloud. During his remarks at Inforum, Carter reportedly said that the company’s existing ERP systems were first developed in the 1980s, and were “being held together by Sellotape and elastic bands,” according to a July 2016 story in CloudPro. “This is a massive transformation for us, and it will ensure we’ll be relevant for the next 25 years,” Carter said, according to CloudPro.
There were several 1980s-era applications that Travis Perkins sought to move away from. According to this July 2017 Diginomica story, the company relied on “three core platforms,” including “an old JDA platform,” a 30-plus year-old “green system,” and it a 35-year-old “white on black system.” (JDA software changed its name to Blue Yonder in 2020.)
Travis Perkins wanted to upgrade its back-office applications in order to keep up with “smaller digital competitors” and the move to omni-channel fulfillment, the Diginomica story quoted Travis Perkins group strategy director Norman Bell as saying. Infor’s M3 was selected to replace all of those systems, providing new general ledger, supply chain, branch management, and front-end engagement system.
“[W]hilst our business is based on relationships, digital wasn’t used to really underpin the relationship,” Bell said, according to Diginomica. “So we were developing initiatives to build the relationships, but without a strong digital backbone we weren’t able to execute on those things. We had point solutions that worked up to a level, but we didn’t have the richness that you get when you renew the backbone.”
While Travis Perkins was able to develop new functionality atop the old applications, which provided “accuracy around where the stock is, where it’s coming from,” Bell said, the underlying architectures of the existing systems simply weren’t amenable to making the changes that the company wanted to make.
More insight into the existing systems used in the Travis Perkins business comes to us from a 2014 story in Insightdiy, a UK-based publication tracking the home improvement sector. According to Insightdiy (which borrowed from a 2014 story by Computerworld that is no longer available), different parts of the Travis Perkins business were using different systems for its “multi-channel” initiative.
For example, Travis Perkins proper was relying on PICK/POS ERP system running on Linux-based HP Superdromes, while Wickes, which owns more than 230 retail locations in the UK, used a JDA ERP system, the Insightdiy (i.e. Computerworldd) story said. The Travis Perkins ERP system was being front-ended with a Sitecore customer experience management system, while Wickes’ JDA system was bring front-ended with a website based on technology from a company called Venda.
Travis Perkins soon ran into problems with the migration. The company announced a delay to its “Merchant ERP replacement programme in December 2018 as this programme has continued to face significant challenges,” the company stated in financial results from the first half of 2019.
“As a result, the Group is considering whether to implement the various elements of an ERP system as separate items, after modernising the Group’s core IT architecture,” the company wrote. The company’s board of directors also decided to write off £111 million (about $142 million at early 2019 exchange rates) in assets related to the troubled modernization initiative.
By July 2019, the entire project had fallen apart. “Following the change in approach to the replacement of the Group’s merchant ERP system announced in July 2019,” Travis Perkins wrote in its full year results report in March 2020, “the Group terminated its relationship with Infor (the software provider) in October 2019 and formally set out its damages claim.”
Travis Perkins took Infor to court, and the companies settled by September 2020, with Infor paying Travis Perkins £4.2m (about $5.8 million at then-exchange rates), according to a story in The Register. A source for the IT news site said the project was hampered by a lack of functional specifications for interfaces, database schemas, and security schemas.
The Reg wrote: “With barely any input from the IT department, business leaders in charge of the project said these aspects could be decided later during the construction phase. ‘It was just an absolute mess,’ one source said.”
Eric Kimberling, the CEO and founder of Third Stage Consulting, dug into the failed M3 implementation at Travis Perkins, and shared the results of his findings in a December 2020 YouTube video.
Third Stage Consulting CEO Eric Kimberling shared insights from the Travis Perkins ERP implementation failure.
Kimberling says that, after evaluating what has been publicly written about the implementation and investigating additional sources (such as Glassdoor reviews, where former Travis Perkins employees shared details about the ordeal), several things stood out as red flags that should be a warning to anybody else would travel the same digital transformation path as the $8-billion UK retailer.
For starters, Travis Perkins’ decision to delay the design phase of the ERP implementation until the build phase had already begun likely contributed to the problems, Kimberling said. “Now again, this is just the tip of the iceberg and this is what the company says was the problem,” he says. “But in our experience, we find that the technology design can certainly create problems, but it’s usually not the root cause of failure.”
An unrealistic deployment timeline likely accelerated the failure of this particular ERP implementation, Kimberling said. While one might think that a fast timeline should help keep things on track, it often leads to bad decisions, he said.
“It forces you to cut scope, it forces you to cut corners on things that you shouldn’t be cutting corners on,” he said. “And from what we can tell, it sounds like that’s what happened at Travis Perkins. They were trying to implement a big technological change that was quite a jump from the 1980s based technology to more of a modern-day cloud ERP system.”
Travis Perkins was also trying to consolidate multiple ERP systems into one, which increases the complexity level and also likely was a cause of the failure, Kimberling said.
“If you’re an organization that has one system that you’re trying to replace, that’s a lot different than if you have a plethora or multitude of different systems that you’re trying to consolidate into one,” he said. “That’s a pretty big leap and really navigating that technological landscape can be very difficult, very challenging.”
Another red flag was the fact that Travis Perkins was undergoing several other big changes at the same time they were trying to migrate an ERP system, Kimberling said. This includes two divestitures, including the Wickes and P&H (Plumbing & Heating) businesses, as well as reorganizing the structure of the branch managers and their compensation.
“In this case it appears as though the company had competing priorities that may have conflicted with the implementation and the overall transformation,” Kimberling said. “Either they weren’t aligned in those efforts, and/or the same resources maybe have been competing to handle those various projects.”
Lastly, Kimberling lays blame for the failed project at the feet of Travis Perkins executive leadership. “It’s important to recognize that you really need that strong executive leadership support, buy-in, vision, alignment,” he said. “All of those things are really important and that’s usually a place where you want to start your organizational change efforts is to make sure that your executive team, make sure they have realistic expectations, make sure that they are setting a clear vision for the project, make sure that they are able to help vet out and balance . . . competing priorities and competing resources.”
One factor that almost certainly was not at play was the M3 product itself. Long a staple among IBM i shops, M3 has been deployed many times, on-prem and in the cloud. “I think you could have plugged in any other technology into the same situation,” Kimberling said. “You probably would have ended up with very similar result.”
While it’s unclear whether the IBM i server played any role at Travis Perkins, the company’s travails certainly should be a lesson to IBM i shops around the world, many of whom are facing the exact same set of business and technology circumstances that this UK retailer faced. Moving away from legacy technology is really, really hard, and care must be taken to ensure that the projects do not go off the rails.
RELATED STORIES
Technical Debt: The Silent Killer
How To Not Fail With Your ERP Implementation
Excellent article, as Alex generally provides