In his OOW17
pre-match pre-conference story, Phil Wainewright had this question about SaaS:
- SaaS — I’m wondering what’s happening with Oracle’s portfolio of SaaS applications, many of them acquired over the years. How well integrated are these solutions into Oracle’s core products such as Fusion and PaaS?
Phil will have his take in due course . This is mine.
My story covers some of the announcements and a pre-conference conversation I had with Steve Miranda, executive vice president of Oracle Applications product development. I'll preface this by saying that Miranda and I have a long history. I've always had respect for both his straight talking and willingness to answer questions that are not necessarily on Oracle's immediate AR/PR playbook. We truly have conversations that represent a candid exchange of views.
I've been following Oracle's SaaS story for some years. Unlike others, I wasn't particularly interested in the Fusion days. Neither was I especially interested in the time it took Oracle to get the apps portfolio out the door and into the hands of customers. I've always been much more concerned with what customers make of what Oracle has to offer and much less on what it promises.
The story so far
The last few years have been interesting. In 2015, the company said it had a full in-house developed cloud ERP suite including supply chain management (SCM.) That got my attention because at that time, no one else was claiming the same. A trip to Moscone West where Oracle had a tiny SCM related booth quickly revealed that yes, it had some capabilities, but it was missing many of the essentials that those who use SCM expect. Last year, the company had made good progress but it was still a work-in-progress. This year, Oracle is starting to showcase customers.
At OOW17, Oracle can claim to have caught up with itself to a degree, but the more important point is that while it has taken time, Oracle is clearly aiming its offering at complex situations where customers can configure to their hearts content without worrying about the hardware and software infrastructure that goes with these business critical systems. As they say, 'it is the journey not the destination' but whichever way you look at it, Oracle has a shot at converting its SCM customer base to cloud.
Add in the fact its HR and finance solutions are much more mature (the HR solution has been doing good business in the M of SMB space as defined by Oracle) and its CX apps are coming along nicely and you have the making of a nice looking package for back office, discrete manufacture, and an emergent front office capability. What does this mean in terms of numbers? As you might expect, the actual figures are relatively small but they are walking-talking customers:
- Procurement - 1,281
- Inventory management - 543
- PLM and product hub - 310
- Transportation and global trade management - 300
- Order management - 306
- Manufacturing - 166
- Supply chain planning - 166
This follows the kind of pattern I'd expect and dovetails nicely with what Oracle told us at Cloud World at the beginning of the year. These things matter because to gain credibility, Oracle has to demonstrate continuing momentum.
I asked Miranda about customers in advance of his session which is live just about now but which is not being streamed. He has Pirelli, which uses Oracle for CX and modern marketing and Carbon 3D which is a greenfield Oracle customer using back office, manufacturing and IoT.
The emphasis is on speed and payback. You'll see in financials and others talking about the business payback. Carbon is pure cloud, frankly hybrid is very common but we don't get the feedback to focus on the integration part (as it relates to business app cloud to cloud.) We are trying to get many different types of customer with different profiles to demonstrate the breadth of apps even if they don't have it all at this time. The highlight on Carbon 3D is they are manufacturing so they are using our manufacturing and supply chain plus ERP and financials so they are truly end to end for us. Pirelli, big, international well known brand is big on CX but likely has other third party apps in their landscape.
This is important because Oracle is both focusing upon live customers who are able to demonstrate how the gaps that existed one and two years ago have been filled. Note: Pirelli was always a big SAP customer and as far as I know that hasn't changed. But what about those large companies:
You'll see a variety of customers, Western digital springs to mind where, as you know, they're cloud adoption pattern is different.
IoT becoming mainstream for some customers
So good so far. But what about innovation. At this point, Miranda started talking about IoT initiatives. If that seems so-2016 in terms of technology fashion then you'd be right but as Miranda pointed out:
Five years ago we wouldn't have even thought about it but speed of innovation at Oracle is really good right now and customers don't have to wait five years to get newer technologies. In IoT it is for preventative maintenance, alerting and where they have sensors in place. There's a six month refresh cycle so customers don't have to wait. Carbon 3D's big pitch now becomes that with IoT and preventative maintenance in place, they can print replacement part ad hoc, as needed. The concept of replacement part warehouses is going away. That's sort of why I think that customers have moved to us more quickly because customers are moving much more quickly than I expected.
Impressive? to a degree and I can imagine automotive manufacturers being able to drastically reduce their tie to delivery of key components, perhaps by situating 3D printers in servicing points.
But as I pointed out to Miranda, with Oracle having so much of the technology stack at its fingertips, surely there are deals to be done where Oracle acts as the backbone IoT provider up to the point where companies can build their own business specific apps. If that happens then in my mind, Oracle becomes a strong player is what should be the embedded operating system for IoT situations. Miranda chose not to go too far down that route and I suspect there are many challenges to making my sense (sic) of sensor led innovation a reality when there are companies like GE which is not only pioneering that aspect but is also discovering just how tough it is to make these technologies truly scale.
The machine learning angle
On machine learning as a component of AI, Miranda is very much in the 'let's get real' camp:
You need the following, you need data on which to make this work. Second you need the algorithms, hardware and infrastructure so between IaaS and PaaS, we think we have those capabilities, processing very quickly to get results (see Larry Ellison's on stage demos from last evening.) Third you need applications to surface. You need then to adapt. I think we differentiate with data. Today it is pretty common to see marketing applications that take machine learning to discover which customer profiles respond to which marketing events to better target customers. But now we are applying that to HR. In recruiting candidate space it is the same thing. You put out your candidate profile requirements so you should be able to see specific jobs available to me. So it becomes what did we show them and did we recruit them? But also are those people successful i.e. highly rated people? That's where we want to tailor and tune the recruiting part.
Interesting. But Oracle is taking this much further. Where last year, these classes of application were really added on top of existing, Oracle is talking about applying machine learning to everything a person touches. In short, Oracle wants to make the work machine learning removes as pervasive.
So when you login into the application you're using, Oracle should know exactly what you need and provide the means for you to do your job as efficiently as possible. Navigation is a good start to that. Maybe when you login in at a certain point in time you might want to do expenses...people have repetitive thing s they do and if we can ease that then they can do the work that's more useful. Eventually we want to get to a push model rather than being reactive. So in an HR example, we might send an email saying now is the time to enter your time card, here's what you entered last week, please make any adjustment for this week.
That just leaves the question of consulting. Is there enough low hanging fruit that Oracle does not need a huge consulting practice?
If you're concentrating upon analytics then I think you're right and we fall ore on the tooling side of that equation but I think the bigger problem is that we are going to find that the assumptions under which we have built applications are just plain wrong. Think about search. When you're spending time to just search for things, then machine learning represents a genuine opportnity to both learn what you're doing so you don't have to search but also change the way you work in these big ERP systems. They're going to become big suggestion machines.
Is Miranda right? Many of us in the cheap seats have wondered what the next iteration of back office solutions will look like. We've tended to think about the technologies available but not necessarily thought about what it means to have so much data and capabilities like machine learning available to us. Oracle clearly has done that thinking. As I listened to Miranda, it became clear to me that despite my personal bugaboo - financials - the technology and data available today change what happens in those systems and the human work that sits around them. Debits and credits plus all that goes with that concept won't go away, but now they can be enhanced, compressed and commoditized in ways that were unimaginable just five years ago.
To that extent, Miranda's vision is pragmatic and attractive because it addresses real life problems of waste, low value work and the improvement of processes but at speed. It's the kind of message that hard pressed execs at large companies can warm to because while it uses buzzwords with which we are familiar, it roots the use cases in a tangible reality that is recognizable.
As always, we have to wait and see how this hashes put but so far so good.
Now over to the ground team to tell us more from the customer perspective.