In my opinionated SAP Sapphire 2022 Orlando news review, one of my top three stories was BTP isn't going away:
BTP's long/protracted history has caused some SAP watchers to write BTP off. Time to reconsider. BTP is centrally positioned in SAP's RISE pitch, and in all S/4HANA migration conversations... One SAP exec told a reporter that BTP is implicated in almost every RISE with SAP project.
When you hear SAP leadership talk about transitioning to a "clean core," and phasing out code customization, that doesn't mean the end of customization. BTP is SAP's go-to solution for that - for any need to extend SAP.
An instructive story of BTP-in-action came the next day, via my sit down with SAP customer Bristol Myers Squibb (BMS). But it's not the use of BTP that's instructive. It's how Bristol Myers Squibb's use of BTP allows them to pursue a long-term cloud apps strategy - without the deal-breaker of moving their entire SAP ERP system to a standardized cloud instance overnight.
Quick context from ASUG CEO Geoff Scott, who was in Orlando for the ASUG Accelerate program. In my underground SAP Sapphire '22 podcast review, I asked Scott about the customers on the ground again - for the first time in three years. What are their top issues? Scott:
What are they talking about? What do we do in a post-pandemic world. We know some realities. We know that virtual work is here to stay, that the way we do big projects - the way we do things in the future - is going to be different. So how do we do that? And how do we think about that? The great thing about bringing people back together again face to face is: they can have those discussions.
Making the S/4HANA business case
What I heard from Bristol Myers Squibb (and others) reinforces that. So I asked Michael Di Novi of Bristol Myers Squibb something I know nothing about: how do you work for the same company for 30 years? You earn your way into meaty new roles, that's how. Di Novi, now Executive Director of Enterprise ERP, cut his teeth in manufacturing, as a master scheduler for a Bristol Myers Squibb manufacturing plant in Greensboro, North Carolina.
Bristol Myers Squibb has a deep history with SAP. So when it came time to upgrade from SAP ECC, there wasn't another ERP vendor up for consideration. That's no surprise when you consider the type of trust a customer can develop for a system that runs and runs, at massive scale. As Di Novi told me:
We have roughly 21,000 batch jobs. It comes down to a batch job running every four seconds. And so I always remind people: 'I'm not getting phone calls - and yet something is happening in my system.' So it is a very well-engineered system. And we do a good job of implementing, I think in a very standard and robust way - and it runs every day.
So, in October 2017, Di Novi's team took on a monster project: take their global ECC 6.0 system, and move it to a single S/4HANA system that runs the whole company - and shift from an internal data center to AWS.
Why upgrade? The conversation started when SAP announced its ECC end-of-life date (originally 2025, now 2027 for standard support with additional years of extended support options). Di Novi's team made good use of SAP's ECC Enhancement Packs. He didn't want to find himself on a system where, after Enhancement Pack 8, new Enhancement Packs no longer shipped:
I looked at how much capability we took out of Enhancement Packs over a five-year period. I said, 'Hey, if we had to build all of that type of stuff custom, maintain it, and then take it out when the time came to go to S/4HANA, how much would that cost us? And it was substantial. It was probably almost 80% of the cost to convert to S/4. So we can either convert for say, X, or we could wait and convert for 1.8X. The technical debt that we would have built up would have been significant.
I won't lie: I never like to hear about ERP upgrades that are driven only by end-of-life concerns. But for Bristol Myers Squibb, the reasons to upgrade didn't stop there. Di Novi:
The second piece is: we really wanted the foundation in place with the speed of the HANA database, the new capabilities, the simplified data structure - for us to enable our business going forward.
Then, in 2019, Bristol Myers Squibb acquired cell therapy provider Celgene Corporation. After the S/4HANA system went live in February, 2020, it had a chance to provide itself:
There are some very nice features inside of S/4 that really allow us to enable that business in a way we would not have been able to before.
Any other business benefits? Di Novi:
The speed of the database allows us to do document splitting... We need to do real-time response to requests coming in to schedule patients. We have the speed and the ability to process - and do available to promise. It's those types of things that allow us to move up.
We were also able to work with SAP to put a chain of identity into all of the processes, which was much easier to do in the simplified database structure. So there's a whole host of things.
On SAP BTP, the clean core, and the path to the "true ERP cloud"
If I stopped here, we'd have a solid use case. But it was Di Novi's SAP BTP strategy that really set this one apart. Di Novi gave me a big clue when he said:
We've also used the Business Technology Platform to create a few applications that also helped us enable our doing business with Celgene. The new architecture of S/4 is really conducive to us being able to deliver much more personalized or individualistic type of medicines of the future.
Okay, I'll bite - how does BTP fit in? Di Novi says he can't get to the "clean core" and, eventually, the "true cloud version of S/4," without BTP:
I bought into the basic premise: BTP will help us move the BTS-specific content out of the core, therefore, keeping the core clean... Admittedly, we still have lots of stuff in there, that's for sure. It will take us years to get it out. But we wanted to start down that path, so that someday, we'll be able to actually go to software as a service for the true cloud version of S/4, and get out of the whole hosting and managing other software ourselves, and be able to take innovation faster, because they deliver it four times a year.
No arguments here - but even if you buy into SAP's clean-core, extend-with-BTP approach, that doesn't mean it's easy to get started. How did Di Novi's team get their BTP app push going?
We did a couple of small things to dip our toes in the water, and get used to BTP and understand it. Part of that was also engaging SAP to do a few things for us. For instance, the routing guide out of the rail car, an industry solution from years ago, we had SAP put that into BTP for us. We just needed one little piece, so we had them put it into BTP.
We began to work with SAP to understand, 'Hey, what's the advantage of utilizing BTP?' We understood the side-by-side, to extend cloud applications - we understood how we can use user exits there.
To pull this off, BMS needed confidence that BTP could work effectively, even in a heavily-regulated industry:
At this point, we've made a half dozen smaller BTP applications, again, just for us to understand and begin to manage them, to understand how we manage the change process, because in a regulated industry, it's a whole new set of practices. So we have to get everything in place. We're crawling before we walk, but we have to start somewhere.
Di Novi's team may have started small, with "relatively simple applications." But moving these to the cloud built momentum:
We have one from SAP called management of change, which controls the changes to bill of material. It produces an audit trail for us. We have another one that gives us self-service. So it allows the users to make changes that would normally be config.
So we created an application where a set of users can actually go in, create and manage those MRP controllers through BTP, and it updates the configuration in all our systems to keep it all in line. And so it gets IT out of the way; it allows our business to react faster. So it's a great win for everyone.
I don't care what ERP vendor we're talking about - I've never felt that a lift-and-shift into a hyperscaler is a long-term win - for either the customer or the vendor. I haven't been shy about telling vendors that; I'm biased towards public cloud ERP, even in the large enterprise. To hear an ERP leader in a heavily-regulated industry like pharma talk about "true cloud" being their long-term destination speaks to what I personally believe will transpire. That's why I've been pushing to get an update on SAP's public cloud ERP story - more on that in a subsequent piece.
For now, SAP believes it has a good answer to the limitations of hyperscaler lift-and-shift, via the gradual move to a clean core, with BTP extensions taking the place of custom code, an unsavory option which I perceive, perhaps unfairly, as instant technical debt. It's another thing entirely to hear proof points from a customer at this scale. In some ways, it's early days for Bristol Myers Squibb here, but it's a story worth watching.
I should mention Di Novi's team also has Accenture as their services partner, a firm that is well-versed in BTP extensions (he speaks highly of their relationship). Whether all customers can have BTP successes like this one will depend on their partners, and continued improvements in BTP's ease of use.
As much as I'm biased towards the kind of "true cloud" future Di Novi describes, this spring, I've run into a number of live, on-premise and/or hyperscaler S/4HANA projects that demonstrated a much more agile platform than what I've seen on ECC. That's a bit of a wake-up call for me. It hasn't changed my long-term bias towards the "true ERP cloud" for most companies/industries. But it's given me something to rethink and weigh out as an interim step.
Clearly, Bristol Myers Squibb has business wins they can already point to on S/4HANA, separately from BTP entirely. That's a good jolt for me, and a good approach for them. Escaping end-of-maintenance isn't good enough; stacking up business wins is now the name of the ERP game. How customers pull that off will vary; add this one to the list.
Updated, May 18, 3pm ET, with a few fixes for reading clarity, and one tweak to my concluding comments for the same reason.