Cloud Hosting

Today's Oracle Live 2020 promises a close look at the customer impact of Oracle's Cloud Applications (see Den Howlett's event preview, Oracle kicks off The Time Is Now with a slew of apps updates but the focus is on supply chain and data).

This event should be a welcome break from the insane surge flurry of breathless TikTok headlines, eh? Over the last couple years, Oracle has made big strides here - surfacing a number of notable Oracle Cloud Applications use cases.

Along with the Oracle Live event, Oracle issued a fresh batch of Oracle Fusion Cloud news, pertaining to both ERP and EPM (Enterprise Performance Management). But if you're like me, you want to know: why should we care? Why does this news matter? That means heading into the debate on public cloud ERP - and whether all ERP clouds are created equal.

For context, I spoke with Juergen Lindner, SVP, Global Marketing Head SaaS at Oracle, prior to the Oracle Live event. I asked him the exact question: why does this news matter? Lindner says it starts with the pandemic, and how our obligations to serve each other shifted to a different level. As he told me:

The context of the news - and why we're having this event next week in general - is really that through the pandemic, I do think our context has changed significantly - as it has just about any corporation these days. We do fundamentally believe how we interact with customers during this time of crisis will be remembered coming out of the crisis.

For the Oracle Fusion Cloud team, if you want to serve customers properly now, you better have an industry-focused partnership. Generic, "horizontal" ERP doesn't get the job done right now. Lindner:

We very much tried to turn from a being a classical technology partnership into a business partnership, by extending a lot of advisory into our accounts and prospects by teaming up, for example, with companies such as McKinsey to say, "Okay, how do we expect the recovery to take shape," because as you would likely agree, the recovery will look very different for different industry sectors.

That uncertainly lead Lindner's team to wonder what the demand for cloud applications would be. But almost across the board, cloud apps uptick is on the rise:

To our very positive surprise, I feel that customers now have a much, much higher appreciation for the flexibility that cloud can bring - simply in context of "I'm still having to close the books for the first time - remotely." That's a pretty heroic endeavor for some who have never done that before. Or, they need to run through an increased pace of scenario and business modeling.

There's no end in sight for that, right? We're seeing case numbers spiking again, so there could be a very different coming out of the crisis as people were thinking about. So in that context, we were very glad to see that our customers continue to invest into cloud technology. And we saw that in the first quarter results.

If customers need different software, they need a different software partner also:

We are very positive that customers have experienced this differently. For example, we have been extending software for free into the market that we've never done before to help customers through the crisis - such as, for example, giving them free financial scenario planning tools, or employee care packages. There are certain business processes that were in the spotlight - supply chain, scenario planning, workforce management, all those type of pieces - and just try to be a very different partner than potentially customers have perceived as in the past.

But here's the problem I have: just about every ERP vendor, regardless of their technology stack, is making noise about "cloud ERP adoption" right now. How are customers to understand the differences? I put it to Lindner: why should customers care if their cloud ERP is a multi-tenant, public cloud type of solution, or privately hosted in a service partner's data center? Lindner responded:

An ERP system in the hosted cloud will not have the sustained business advantages that we think a public cloud will have - for multiple reasons. We do think we can be a better business partner to our customers by being super-responsive to where the business turns. Some of the announcements are in direct response to the crisis.

Previously, it would have taken us some time, from the time we write the code, five to seven years until it arrives to the customer site for business impact. Here, we have the opportunities every 90 days to infuse new functionality that customers asked for, which we think are essential for them to have to drive new business models forward and stay ahead of the change. The customer can decide which ones they activate, similar to in iPhone; you decide when you update them.

For years, Oracle Fusion was a mysterious product Oracle was supposedly working on that never seemed it would see the light of day. But Lindner insists that "not taking the easy way out" and rewriting the entire ERP applications for the public cloud - on the same data model - was the right move. Beyond shipping new functionality quickly, Lindner points to the ease of extending that functionality. Take subscription management - Lindner argues that you could get into trouble extending subscription management across multiple vendors, doing configure/price/quote with one customer-facing vendor, revenue modeling with another, and the supply chain aspects with another. He says this leads to a counterproductive "cloud hairball":

Often times, customers fall into this very tactical cloud adoption trap, which I call lovingly, the cloud hairball. Eventually, you'll choke on the complexity.

Mergers and acquisitions muddy the waters. Lindner referred to a Fortune 20 CIO who told him the effort to manage various line-of-business cloud apps across divisions and acquired companies was not working. Lindner called this "accidental architecture" - something no CIO wants.

I showed him a slide: "Look, this is the cloud hairball." He said, "Holy crap. That's us."

My concern about single-tenant/hosted "cloud" ERP goes beyond ease of functionality - into the problem of siloed data. How can you realistically expect to apply machine learning and workflow automation across processes, when a core ERP system is moved into its own separately-managed private cloud? I'm not knocking a money-saving lift-and-shift in pandemic times. But - it's not the end state. Lindner's reaction?

I do think lifting and shifting past investments into the cloud is going to have some savings. But every deployment option that you strike, potentially with a hyper-scaler, has certain fine print that you might want to be aware of. And to your point, oftentimes, what we're seeing is that customers need to resort to very costly data warehousing constructs, because [their applications] are not on the same data model.

That leads into a debate on why machine learning should be embedded in workflows, not applied as a data warehousing layer. But for now, I needed to ask Lindner about the EPM issue - given that Oracle Live features its share of cloud EPM info and news. In pandemic times, the need for much more agile/flexible planning processes is clear. The question is: how do you get there? Pure-play EPM players believe they have the answer. No surprise - Oracle sees this differently. Lindner:

We think strongly that EPM should not be an addition sitting on top of existing systems... Once again, you would be extracting data from your core system, into a third party type of tool, come up with a recommendation, potentially, and then need to execute it - in your core systems. Once again, for us, it needs to be seamless.

Plenty of provocative talking points for customers to consider here. I do agree with Oracle's contention that EPM should span across all areas - it's no longer a finance-only function. From workforce management to supply chain, the planning capability must be fluid. Whether EPM should come from the same system as your ERP - that's a question for rigorous evaluation per-customer.

I would be remiss if I did not mention that around the time of Oracle Live, Oracle announced some ERP hosting options that effectively blur the lines between multi-tenant and private cloud options (for details, see: Oracle Cloud@Customer). This is a complex issue that deserves its own piece - which diginomica contributor Brian Sommer has already done, in: How Oracle's latest announcements change multi-tenant vs. on-premises decisions.

Does this muddy the waters around Oracle's public cloud ERP messaging? Not necessarily. Even a public cloud advocate like yours truly will concede that in certain industries, data compliance and, perhaps, IT data philosophy means public cloud is not an option. My concerns would be whether some customers would choose a private hosting option more based on unsubstantiated cloud FUD (fear, uncertainty, doubt), perhaps enabled by service partners who see a revenue opportunity. It's the vendor's job to make sure that the private cloud option isn't over-used. I stand by the economies of scale of the public cloud architecture; it will take a lot more proof points for me to move off of that into private ERP cloud acceptance - though I'll concede what Oracle has done here appears to reduce the private cloud downsides and maintenance costs.

As for the "cloud hairball" issue, that's undeniable. Josh Greenbaum referred to a variation on this unmanageable condition this week, as extreme heterogeneity. But there is a huge difference between cloud hairballs and strategic integrations. At Oracle Live, I'll be looking for examples of Oracle effectively integrating Oracle Cloud Applications with a customer's third-party app of choice. Pushing a platform is a good thing. But when vendors revert back to "we have all the apps you need," that doesn't resonate with today's large enterprise customers. Hairball or not, business users are more than willing to revolt/boycott/ignore if your software falls short in their area. Let's see what Oracle puts forth this week.

Some may raise eyebrows at Oracle's vows to be a different kind of partner to customers. I don't have to take a position on that, except to say that customers will ultimately make that call, and will vote with their wallets. Raising the bar on such things means you have to meet it - so we'll be watching.

End note: by mutual agreement, we did not get into the Oracle TikTok topic in this interview. For more on that issue, refer to my colleague Stuart Lauchlan's TikTok, Oracle, Walmart and Trump - the shape of things to come? (Updated).

See the rest here:
Oracle Live 2020 - not all ERP clouds are created equal. Oracle makes the public cloud ERP case - Diginomica