The CEOs of SAP, Microsoft and Adobe recently announced plans to team up and launch an Open Data Initiative, with the aim of helping customers extract data from their siloed systems of record and create a single view of the customer. The announcement has attracted attention, with it being perceived as an attempt by the vendors to create a standard customer data model.
However, how realistic is this? And how far do the vendors plan to go with this? Also, what is SAP’s motivation? Attempts by industry to create standards often materialise into nothing more than press releases. Will this be different?
I got the chance to sit down with Moritz Zimmerman, SAP’s CTO of Customer Experience, at SAP Customer Experience Live in Barcelona this week, where he explained that the decision has been driven by customer demand to apply machine learning capabilities to data across systems in order to differentiate, as well as being an internal drive for SAP to be more open.
However, at least initially, the Open Data Initiative will start small - focusing on elements of creating a customer data model - and grow from there. It shouldn’t be perceived as an attempt to try and standards on all data entities.
Zimmermann said that the initiative is based on a “few fundamental beliefs”.
It’s our customer’s data, we don’t own it. It’s not ours. And that data should not be a lock in factor, or a prohibitor of adoption, for products. Our products or any other products out there. And many of our customers now see that data becomes an important area for them to innovate on with machine learning.
The idea is really to create one common data model, starting with the customer entity. It’s a lot of work in progress. The big insight, the thing that allowed us to create this partnership, was to recognise that we do compete on the application, and sometimes on the infrastructure, and we will continue to do that, but to not to make that part of the discussion. Everyone might have their interpretation of where to store the data, or what to do with the data, where the data should best sit, but let’s make sure it’s portable.
An SAP ambition
Zimmermann added that although the Open Data Initiative is key, and the three CEOs are having a monthly call to drive it forward, SAP is already working internally on creating a ‘master’ customer data model, as this is core to what’s been promised as part of the C/4 HANA suite. He said:
We are starting with the customer entity. So rather than having a customer record in marketing cloud, sales cloud, in ERP, you name it. There will be, very soon, a common cloud-native web service that you will call and give you the central entry point for customer data. Not just for C/4 HANA, but for the whole of SAP, and everyone else will do the same thing.
This is then the front runner for other master data services for the intelligent enterprise suite. You can can think tickets, case, product, order, whatever. We are doing that right now and the effort is very much aligned with ODI for that. We need to have a common API that defines, what is a customer record? Now, the question is, do we make that API private or do we make it open? And others can call it as well.
For us, what’s important is that we need to do a lot of the work ourselves internally, because to make C/4 HANA a reality there are seven qualities for a suite. One of them is also to have a common set of master data services, customer clearly being one of them.
However, Zimmermann admitted during the discussion that aiming for a common data model across everything is a “non-starter”. It gets too complicated and wouldn’t work. But that doesn’t mean SAP - and other vendors (and the ODI is open to others, apart from Microsoft et al) - shouldn’t aim for some commonality. He said:
I agree that having a common data model across everything is a non-starter. But I think having a common data model across a few entities is important - and possible. I think even if we just achieve it for the SAP product, it will be a huge, huge advantage for our customers. For a minimal set of customer attributes, or product attributes, can be done.
Maybe the idea is to say, there’s a core canonical model on some core attributes. And then maybe there’s a harmonised look-up approach to, say, ‘Dear Adobe Software, Dear SAP Software, tell me more’. This is more of a look up approach, but even there I think you can standardise.
The threat of open source
However, following a long history of enterprise vendors locking customers in, SAP now recognises that there is a shift taking place in the marketplace, where buyers are less willing to accept this. Open source is growing in popularity in the enterprise and SAP recognises that it has to adapt. The Open Data Initiative is a direct response to that, said Zimmermann.
There’s a desire for more openness. Companies are realising that IT has changed from supporting a little business process, making a sales team be a bit more effective, to a core value proposition. And the lock-in aversion becomes much more critical now. Companies have gone through these painful exercises, where they’re reducing their footprint year over year over year, yet they pay more and more and more. What’s the other alternative? Go open source.
That’s why we open sourced Kyma (SAP’s Cloud Platform extension) and other things. We believe it’s the natural way for software to exist. If you look 15 years back, it was a few crazy guys. Now it’s like, 30%? Fast forward another 10 years and 80% of software will be open source. I’m not saying free, I’m saying open source. If a vendor breaks the promise, you can at least take the source code and go somewhere else. The ODI plays into that as well.