Microsoft, Adobe, SAP Unveil Open Data Initiative

Microsoft's annual Ignite conference got underway this morning, with day one already providing lots of news. To me, the most compelling announcement is the Open Data Initiative (ODI), jointly supported by Adobe, Microsoft, and SAP.

It appears that Microsoft wanted this announcement to have some punch, as the CEOs of the three companies -- Adobe's Shantanu Narayen, Microsoft's Satya Nadella, and SAPs' Bill McDermott -- each took turns discussing the importance of what they are jointly doing.

There were no details provided on roadmap or availability of ODI, but the concept is simple. The three companies have teamed up to enable their mutual customers to do more with their data based on three guiding principles:

1. Every organization owns and maintains complete, direct control of all of its data
2. Customers can enable artificial intelligence (AI)-driven processes to derive insights and intelligence from unified behavioral and operational data
3. A broad partner ecosystem should be able to easily leverage an open and extensible data model to extend the solution

This particular incarnation of ODI would have data from Microsoft's Dynamics 365, SAP's C/4HANA and S/4HANA, and Adobe's Experience Cloud; use the same data structure; and be made available in the Azure cloud. Businesses would then have a single data set that they could apply AI- and machine learning (ML)-based algorithms in order to find new insights that can improve customer experience.

It's been well documented by companies like Gartner, Accenture, and Five9, that in the digital era customer experience will be the top brand differentiator. This makes finding those insights in the massive amounts of data faster than the competition the number one thing businesses and IT leaders focus on.

In theory, ODI will enable businesses to do the following:

• Unlock and harmonize data that was previously held in data silos
• Bi-directionally move data to and from a common data lake
• Create data-powered digital feedback loops for greater business impact
• Improve security posture and better meet privacy and compliance initiatives
• Build smarter applications that can leverage data from Adobe, SAP, and Microsoft

During the keynote presentation, the three CEOs discussed the excitement that companies like Coca-Cola, Unilever, and Walmart have shown over ODI.

I applaud SAP, Adobe and Microsoft for the joint initiative, although I do find ODI to be idealistic and Pollyannaish at best. Using Coca-Cola, Walmart, and Unilever as "proof points" for ODI is like asking a person on the street if they want a million dollars. Any business would want to harness the power of all their data.

As I pointed out in this No Jitter post just last week, the data sets that most companies have are filled with erroneous information. The issues that I noted in the post -- workers inputting only some of the available data or even bad data -- aren't specific to Salesforce. SAP, Microsoft, and Adobe have these issues as well. This means that with data aggregated from the three platforms, ODI customers will wind up with a data lake full of polluted information. It reminds me of the scene from the Simpson's movie where Homer dumped the silo full of pig feces into a clean lake and polluted it.

I use this analogy because that's what most companies face. Dumping three sets of bad data into a common data lake leads to a bigger set of bad data. And as the axiom goes, bad data leads to bad insights. Again, the vision of ODI is great, and I tip my hat to Adobe, Microsoft, and SAP; but the quality of data that currently exists in CRM and other systems is so bad that it's hard to see how customers will get anything usable out of it.

Also, I don't believe the structure of Dynamics, HANA, and Adobe's CX products are the same, so trying to create a single data model will likely require a significant amount of data restructuring and preparation. Then the question arises, which data set is the one to use? What makes Dynamics data better than HANA data, or vice versa?

The other big barrier that I see getting in the way is multi-vendor support. It's awesome that Microsoft, SAP, and Adobe have joined hands and are willing to sing Kumbaya together, but what about companies like Salesforce and Amazon? Or what about the contact center vendors like Genesys, Avaya, and Five9? Further, would Microsoft be willing to federate data between Azure, AWS, and GCP? There's also the problem of accessing data that's store in on-premises applications.

For all my cynicism, I do believe this is how vendors should be thinking. Historically, any time vendors are willing to make their customers lives easier and open up platforms, it always creates a rising tide that lifts all boats. As an example, the use of mobile phones was limited when customers could only use them in certain regions on specific networks as well as worrying about calls that were in network and out of network. As mobile networks have become more open and we no longer have to worry about on-net and off-net, usage has skyrocketed.

With this in mind, I'll provide a few recommendations to the vendor community if they are really interested in helping customers do more with their data.

• Make data ingestion easier and more automated. The post I referenced earlier profiled a company called People.ai that did this for Salesforce. And as big as Salesforce is, it is but a small slice of the overall data pie. As long as people have to take the time to input information in the tedious manual manner they do today, something like ODI can't work because the data that feeds into the common data lake is flawed. There's enough AI, ML, natural language processing and other technology for automating data ingestion available today, that we could eliminate this issue.
• Use ML to clean up and normalize existing data sets. The first step in this is agreeing on a common data set. The three founding members of ODI talked a big game, but whose model are they going to use? If one vendor's standard is chosen over another, does this put them at an unfair competitive advantage since their partners now have to go through the work of transforming data? Also, something other than people needs to go through existing CRM and other records to clean up what's already in place. This seems like an ideal situation to use unsupervised learning-based ML. The problem is most companies won't have the skills to do this, so it falls on the shoulders of the vendors to take the lead here.
• Find a way to include on-premises systems. The cloud is great but in reality, there's far more customer data stored in private data centers than there is in the cloud. This includes a massive amount of highly fragmented secondary storage. Perhaps, a good first step is to stop thinking about data as "primary" and "secondary," and just think of data as data.
• Stop fearing being open and be willing to share data. When things are good for customers, they are good for the whole industry. It's better for vendors to have a smaller piece of a much bigger pie than trying to hold their customers hostage to their own tools and applications.

ODI is a great start but it's one small step in the long journey to being able to find those hidden insights locked in the massive amounts of data that customers have.

Follow Zeus Kerravala on Twitter!
@zkerravala

Related content:

• People.ai Offers a Better Way to Salesforce
• Microsoft Mixes Up Reality, Video Calling to Assist Remote Workers
• RPA: Becoming the Linchpin of Enterprise AI