IW: Does mixing the cloud-deployment decision with the Hana move complicate matters?
Plattner: No, it simplifies and accelerates the move. Even if it's only for the deployment phase, it's clearly the superior platform. If companies are hesitating and don't want to make the decision, cloud or on-premises, we suggest to them, "let's do the project on our computers in the cloud. Once the system is ready, we can discuss where to do the production deployment, whether that's cloud or on-premises."
IW: That's where the Hana Enterprise Cloud is different than other clouds in that it's a private, hosted cloud, not a multi-tenant service. Does that make larger customers more comfortable with cloud?
[Want more on real-world deployments? Read 6 SAP Hana Customers Share Early Lessons.]
Plattner: I don't understand why people are hung up on this multi-tenant thing. We do multi-tenancy for Ariba. Do you think multi-tenancy is less safe? Shouldn't Ariba be safe?
IW: Is that an argument to do everything in a multi-tenant environment?
Plattner: Multi-tenancy is good when you have many relatively small communities doing similar things. Ariba is one example. Talent management is one. Salesforce automation is one. Salesforce automation for a company with 10,000 salespeople is a [case for] a separate system. Any of the larger ERP system are dedicated -- that's what I call it instead of "hosted" -- because they have to be standard systems. That have to be much more standardized, otherwise we do not get economies of scale.
Multi-tenancy is a less-important part of the cloud business. The important part of the cloud business is to offer rental. If people think dedicated systems are safer, fine, it is a safer environment. We don't share hardware between an Exxon and a British Petroleum. The big change is that the software can be rented. In the early days of SAP, long before we became public, we had rental software. In the end you make more money.
The most important thing in the cloud is that the manufacturer of the applications can provide the service of the application at any time of the day from any location, permanently. That means we can permanently improve and update the service. This is huge. Our simplified systems enable us to add to or change the system constantly, not by replacing but by putting something in place in parallel.
IW: Simple Finance is the first "simplified system," but you've announced that many more will follow. How should we think about these simplified offerings, because they're not really "new" applications?
Plattner: It's a component -- a piece of the application that handles certain functionality. These components are linked together by the [Hana] platform. The next might be, say, 50 new Fiori apps for certain functionality. The customer can traverse between the current system and the new system, and when the users are ready, they can switch over. They don't have to switch over on a certain day determined by IT.
IW: Do customers have to worry about losing customizations from their current systems?
Plattner: We carry everything the customer has deployed on [SAP Business Suite] forward. Just looking at accounts receivables within SAP, we have 44 different layouts in our shared services center. Our users have defined them. We cannot ask customers to redo such customizations only because we want to move them forward to Simplified Finance. Their old transactions work on the simplified data model. That's the first transition we did. The second transition is that we've put in a new application and a new accounts receivable line-item display into the system that carry forward the customizations of existing applications. So when customers go to the new applications, they will still see their layouts and will not lose their customized views.
IW: And yet Simple Finance is described as introducing new user interfaces and stripping out old code. Doesn't that introduce change?
Plattner: This is something SAP handles and it's non-disruptive to the end user. If users cannot adapt to new user interfaces within two-and-a-half hours, then we're not doing it. Any new interfaces introduced have to be intuitive... Since the system is so much faster, the learning curve is faster. John Deere planned to have 14 work days for engineers to use their [Hana-based predictive maintenance] system. They cancelled the education after two days because the engineers said there's nothing more to learn; we know the system. Our brain does not work well when the partner we have a dialogue with is slower than we are. If something is extremely fast, you can learn to operate the system on the fly.
You can use distributed databases without putting your company's crown jewels at risk. Here's how. Also in the Data Scatter issue of InformationWeek: A wild-card team member with a different skill set can help provide an outside perspective that might turn big data into business innovation. (Free registration required.)