Phil Wainewright has a thoughtful post on the subject. He writes:
Questions remain concerning e-commerce giant Amazon's discontinuing of its wholesale deployment of Workday HCM and Payroll, which came to light this week after a report in Business Insider. Workday subsequently published a blog post confirming that the two companies had "mutually agreed to discontinue" the deployment more than a year and a half ago, which was over three years after Amazon first signed up to the deal in October 2016. The deal was announced in February 2017, shortly after Workday announced retail giant Walmart as a customer, a deployment that has since successfully gone live.
On a positive note, the project is ending without litigation. And, according to Workday's blog post, it will continue its partnership with Amazon's AWS for its cloud infrastructure, as well as its implementations with other Amazon subsidiaries, such as Audible, Twitch, and Whole Foods.
What Happened?
The Business Insider report, based on an anonymous source, says "the database behind Workday's software didn't scale as planned to fully support Amazon's rapidly growing workforce."
Workday disputes this. It writes:
This was not related to the scalability of the Workday system, as we currently support some of the world’s largest organizations, including more than 45% of the Fortune 500 and more than 70% of the top 50 Fortune 500 companies. In addition, more than 70% of our customers are live, including one of our largest customers — a retailer — across its more than 1.5 million global workers.
Workday, rather, writes that the project failure has to do with Amazon having a "unique set of needs that are different from what we're delivering for our broader customer base." It also writes, "At times...customers have a unique set of needs that are different from what we’re delivering for our broader customer base, as was the case with Amazon — one of the most unique and dynamic companies in the world."
How Do I See It?
All I can do here is read between the lines.
First, I don't think the Business Insider's claim of a Workday scalability problem is credible. Workday doesn't name its large retail customer, but no doubt it is Walmart. In 2020, Amazon had about 1.3 million employees, while Walmart had about 2.3 million. So, as my late business partner used to say, there is "an existence proof" for Workday being able to scale to support enterprises with multi-million employee counts.
Then, what about Workday's claim that Amazon had some unique requirements that are different from what Workday provides for the rest of its customers?
This has a ring of truth to it. Amazon is unique in many ways, and it would not be surprising if this extends to how it hires, retains, and manages its workforce. As a SaaS provider, Workday cannot afford to customize its core architecture and process to accommodate a single customer, even one as large as Amazon. It is commendable that Workday was willing to walk away from a large opportunity like this rather than compromise its core architecture.
On a much smaller scale, Plex (a cloud manufacturing ERP provider), in its early years, used to make customer-specific customizations to its core multi-tenant code base. Later, it paid the price to move those customers off those customizations and back to its common core. To my knowledge, it is still trying to do so. Workday is not about to make that mistake. (Interestingly, Plex itself is a Workday client and partner.)
What Happens Next?
Workday writes that it and Amazon may revisit the HCM deployment in the future. But for now the project has been discontinued. This leaves Amazon on its legacy Oracle PeopleSoft HCM system.
This is where the plot thickens. There is no love lost between Amazon and Oracle. With its Redshift offering, Amazon looks to shift Oracle customers away to Amazon's data warehouse. Oracle, in turn, looks to compete with Amazon with its own cloud infrastructure offering. Naturally, Amazon has been working to disentangle itself from any use of Oracle products in its internal operations. Having to remain on PeopleSoft has to stick in the craw of Jeff Bezos. This might explain why the project may be revisited in the future.
There are not many HCM options for enterprises the size of Amazon. PeopleSoft is a legacy platform, with Oracle's HCM Cloud as its successor. But Amazon is not likely to increase its dependence on Oracle. Workday is the obvious alternative, which is why, despite the project failure, it still might be "revisited.
So, is SAP an option?
Photo Credit: Coolcaesar, CC BY-SA 4.0, via Wikimedia Commons
7 comments:
Many B2B SaaS companies have paid the ultimate price by choosing to implement customizations in their products for enterprise customers. This is a dilemma (and a choice) that nearly all CEOs and product executives face at some point along the way.
Must focus relentlessly on target market. Great article, Frank.
Amazon having a "unique set of needs that are different from what we're delivering for our broader customer base."
By happenchance during interviews I discovered Amazon is far more granular in its data collection for business optimizations in near real-time. Thus functionality driving scalability at Amazon's size could add back pressure to Workdays OLAP based multi-tenancy. I added a comment on Phil Wainewright's Diginomica article on the different database foundations Workday vs PeopleSoft are built on. Recall Force dot com still has governors to kill rogue back pressure. As business systems move to cloud networks away from ERP era on-premises scale, how the key vendors provision scalability will be the key area to watch.
"So, is SAP an option?" Maybe...but more likely Infor. Infor HCM/Payroll runs on the AWS cloud. So much of their cloud subscription fee will go right back to Amazon.
Numerous B2B SaaS organizations have died by deciding to execute customizations in their items for big business clients. This is a difficulty (and a decision) that essentially all CEOs and item leaders face eventually en route.
This topic reminds me of the “Super Pod” instance that Salesforce spun up for HP; because HP was so big relative to their other customers. Maybe it has come back into the common instance now (if suitable governors are in place to throttle load).
Amazon is as much a tech company as it is a retailer, so it would likely have set the functionality bar quite high compared to Walmart.
Like other commercial off-the-shelf (COTS) software and services, Workday’s functionality would suffer from lowest common denominator syndrome (LCDS). That is usually a good thing; anyone who needs something off center will need to pay extra to have unique functionality built. Fortunately that is easier to do nowadays, thanks to application platforms as-a-service.
Interestingly on that topic, Workday launched their own Workday Cloud Platform a few years ago, but then not too long afterwards they collapsed it into their application suite meaning it wouldn’t be available as a standalone platform (i.e. intended for building applications that work in tandem with Workdays standard suite functionality.
Frank, I loved your article [What I Learned at Smith Tool about Enterprise IT], not only for the content, but for all the great reminiscences of Ken, Roly, Nick and others, and those early days of the APICS Orange County Dinner meetings. Such Great Memories. APICS has been very, very good to me over my career. Although in Semi-retirement, I still do Webinars and training for ERP systems. Thank you for all your history. My Best Don Lindsey
Don, great to hear from you. Thanks for the note. Best wishes in your semi-retirement! --Frank
Post a Comment