From on-prem to cloud native - the top transformation lessons from Oracle cloud customers

(Oracle's Safra Catz and Grupo Bimbo)

All enterprise software vendors have this much in common: the goal of achieving customer trust and credibility. Unfortunately, events don't necessarily advance that cause. If you put too many shiny new toys on stage, you lose the chance to put customers front and center.

This year, Oracle set a decidedly different tone - opening with a different type of keynote by CEO Safra Catz. Catz brought customer after customer on stage. That continued throughout the show - including a slew of diginomica interviews already published.

I've documented enough use cases from enough vendors to believe this: multi-tenant cloud applications bring a different type of value  - as does modern ERP (see: Inside Oracle's B2B commerce and apps platform news - customer and partner impact). However, that doesn't change these hard truths:

• Enterprise-wide transformations are enormously difficult. Yes, a broad transformation is a far better pursuit than scattered digital projects, but that doesn't make it easy.
• Even the most successful projects have moments of truth, if not crisis.
• Establishing the right metrics - and achieving them - isn't a piece of cake either.
• If you don't stack up ROI wins along the way, your transformation will stall.
• Users have the final say. If you don't achieve user adoption at scale, you have cloud shelfware.
• Cloud isn't some magical panacea. Cloud expenses can pile up, and it could get worse. Narrow cloud "business cases" built only on IT needs or efficiency don't hold up. A broader business case is needed.

No easy solutions to all that - but we can start by learning from peers. So for this piece, I've picked out top project lessons from Oracle CloudWorld customers.

Cloud infrastructure and tech improvements can be a good first phase - if you build on it

As long as you have a bigger business case in mind, starting with IT and infrastructure can make sense. During an Oracle cloud customer session for media, Gordon McKenzie, CTO of Deutsche Bank, spoke to the technical underpinnings of a sweeping transformation. He also spoke to the importance of a trusted, hands-on partnership:

One of the problems that we had: over the years, we built up a whole range of shared service Oracle platforms. It's more difficult than you think to do that, and to run lots of those in parallel. What we were faced with was: what's the right answer to take us forward?

We went back and forward with Oracle... We entered that new deal about 18 months ago; it's been phenomenally successful. We've already moved thousands of databases to the cloud. We'd particularly pleased with the two sets of engineers working together. We're coming to Oracle with product features; in terms of how the relationship is working, it's really good.

We're starting to see the benefits we expected in terms of stability, cost, and our ability to implement the security posture that we wanted. All of those things are going to come to fruition... There's a long way to go. It's a five year journey for us.

I think we have one of the largest cloud customer deals, so it's a significant journey, because it is a lot of our mission-critical systems. But 18 months in, I think we're happy with the choice.

ROI can come from surprising places - but you must find it

During our media roundtable with CMO Oliver Hughes, CMO of Red Bull Racing, he shared their extensive work with Oracle - from engine building to data analysis to cutting edge CX. We asked Hughes how he assesses results. He responded with one of the most unique ROI examples I've heard:

One of the key elements in a race weekend is actually tires. Tires are a really interesting differentiator of tire wear, and when to pit versus your competition. Because when tires wear, your lap times decrease, so everybody has to pit at least once during the race.

One of the really good use case examples with Oracle: we run our in-race simulation on something we call the massive Monte Carlo simulation. It generally came out of the casino houses: as instances happen, outcomes can change.

So we run this Monte Carlo method on Oracle Cloud. What you want to do during a race is have a really good burst capability. What you're doing is building a massive data center that you only use at full capacity for a few seconds or a month. What Oracle allows us to do is run that full simulation on their cloud. We can run 25% more simulations than we could in our old on-prem system. That's given us a huge competitive advantage.

We have live photography on all of our competition, and on us. We are live monitoring the  tire [conditions]. And then the simulation tools are actually telling us what's happening with everybody else, and what the optimal window is for us to pit in. So if you've followed the races at all this year, there's been a number of instances where we've made really bold pit stop moves, and then won the race. That's really all capable because of the Oracle AI tools that we have, that are predicting what's happening, and telling us what to do.

Autonomous AI is overrated, but not the convergence of machines and humans 

I worry about the human consequences of giving AI more leash than it can handle (example: screening mortgage or job applicants). But providing humans with machine feeds - and actionable data from diverse sources - is another matter entirely. Continuing with the Red Bull Racing example:

It's not just tires; there's a lot of inputs into it. We have radio feed from other drivers, where they're talking about how to entice field engineers; we have imagery that we can use. We have our own eyes too. This model takes millions of inputs and then spits out outcomes, and then ultimately humans will then go make the decision. It's also not a case of 'Just change tires,' it's 'Change to what tires.'

Not every customer is ready for a full cloud apps move - that's where OCI comes in

It's a reality cloud ERP vendors with on-prem install bases must face: some customers just aren't ready to move to full-on "public cloud ERP," for whatever reason, whether it's industry requirements, regulatory hurdles, broadband access, prioritizing other projects, or simply lack of appetite for a major ERP overhaul at this time.

But in the last year, I've seen vendors get much more creative about helping customers take manageable,  gradual steps towards cloud-based operations. Oracle (and its partners) have a good roadmap for these moves, via OCi (Oracle Cloud Infrastructure). I'll have more on this later from my 1:1 with Oracle partner Syntax, but for now,  my colleague Derek du Preez documented a revealing example, via his talk with Scott Petty, Vodafone CIO:

Petty said that whilst Vodafone is building new digital platforms on AWS, GCP and Azure, it has a huge number of Oracle databases that it needed to modernize in the least disruptive way. He added:

' Moving from on-prem to cloud native is a big step for a company to go through. You're touching every business user, you have to upgrade the application, you have to change the business process - you can only do so many of those at any given time or you crash the business.

We got very excited about OCI because it enabled us to build cloud infrastructure in our own data centers, modernize the databases, but only touch the application if the business wanted to touch the application.'

Modern ERP is about adaptability -  not static reports and customization requests

My advocacy of multi-tenant applications didn't come from vendors - it came from their customers. For those who don't think that modern ERP can be different, check out this excerpt from my du Preez's conversation with Heathrow Airport's Caroline Knight, Heathrow's Head of Technology:

Heathrow completed the project and went live approximately a year ago. It now has a DevOps team that manages the quarterly releases from Oracle, assessing how the business can constantly adapt and change. Knight said:

'We went from a world where we had E-Business Suite R12, which we implemented and it sat still and we didn't touch it. Then maybe five years later we'd do an upgrade, it would cost a fortune, and no one was really happy about doing it.

Whereas now, every fortnight we are continually evolving the system with the DevOps team and it's working very well.

We've got a big recruitment task force at the moment and we've put quite a few changes into the system to make our recruitment process more efficient, to try and speed it up.

And that's the beauty of the solution we've got now. We're able to change it and change it ourselves, because it's all just configuration change. That's just something that we couldn't have done in our previous world; we'd still be sitting here now trying to write the contracts to change the system. Whereas, we've just got on and made the changes ourselves.'

All good projects - and partnerships - will be tested

We can make happy talk about modern ERP all day long. Deadlines come to a head, and partnerships get tested. Too many vendors shy away from this on the keynote stage. At Oracle CloudWorld, we got a memorable example:

My take

Transformation projects need milestones, but they are never truly done, and that's a good thing. We all need that slow-burn urgency. In the case of cloud ERP, the greatest benefits typically come well after go-live - if the company pushes for them (I have documented this in a number of posts, including Extracting value from cloud ERP in a customer-first world - what should we be pushing for?) 

No one expressed that conviction to extract value better than Raul Obregon, CIO and Chief Transformation Officer of Grupo Bimbo. As he told Catz onstage:

Six/seven years ago, when we charted our digital journey, we took two decisions. One was to be a cloud-first company, and the other was to leverage all the data that we have.

With ERP fusion with you guys, it's been a journey. We've, I would say, masked it as an upgrade. Actually, it isn't. It's been a very hard job of many. We're in 33 countries, so by year's end will be about 70% done. But I think the real challenge is the next step. It's about leveraging all the Oracle technology that we've deployed, and actually seeing the end-to-end value chain that we have.

We operate 206 bakeries worldwide; we have a 55,000 distribution network that serves, give or take, 3 million points of sales every week. So for us, all that data and all that visibility will truly push us to be even more efficient, and to have fresher products.

This is not a typical upgrade - and that's a good thing also.  During our media session, Obregon added color:

I think I joked in the keynote that we always talked about an upgrade, but for our associates, actually that wasn't right - it was a new ballgame. And that helped us a bit to grow faster. So that's where we are. We hope that by the end of next year, we should be done  - and now, we move on to harnessing all the data and the technology stack we laid down.

Yes, attendees had fun in Vegas; I heard Steve Miller put on a great concert. But to me, hearing so many customers speak so openly about their projects is the biggest CloudWorld win.