Data exhaust in SAP’s SaaS applications: An opportunity that shouldn't be missed

[richard_hirsch]

Note: I know that there is currently a certain degree of disturbance in SAP’s cloud activities.  Since broader changes in this area are in progress, I’d like to take a deep dive into one small corner of the overall strategy in the hope / expectation that not all aspects of SAP’s cloud strategy will be changed in the next few weeks .

As I was reading a recent press release on the use of HANA in Ariba, I came across an interesting statement:

The Ariba Network on SAP HANA will create an entirely new cloud offering possible only from SAP. Companies will be able to glean predictive intelligence on risks, performance, capabilities, rates and more from the more than 15 years of transaction and relationship data and community-generated content that reside in the world’s largest, most global business network. Leveraging the speed and power of SAP HANA, they can instantly process this information and enable innovative processes that dramatically improve their performance and productivity.

I found this statement interesting, because, at least for me, it implied that aggregate data of Ariba customers would be used to provide HANA-based analysis.  Ariba has customers with different lengths of association with the company – some companies had been Ariba for 10 years, others for 6 months.   The press release directly mentions “15 years” of data so my assumption is that aggregate data from Ariba’s entire historical data would be available.

As I read this statement, I was immediately aware of the commercial value of such historical data. Thinking about this feature, I was curious regarding the background (legal, technical, etc) of such data usage.

Such scenarios is called “SaaS network effect data (AKA data exhaust, aggregate application data, etc.) (and) is based on the position of the SaaS vendor as a proxy in a multi-tenant environment and the value that is created by the network effect data can be quite significant when leveraged correctly; both for the vendor and the ecosystem.” [SOURCE]

Note: I’m definitely not a lawyer, so my analysis of legal documents should be taken with a grain of salt. I’m also using publicly available documents. There may be confidential contracts between SAP companies and their customers.

One blogger suggests the following points to consider when providing such functionality:

• Make sure your User Agreement is up-to-date and is clear about your intentions (see your legal department)
• Anonymize the data before leveraging it in aggregate
• Leverage the information as soon as it becomes actionable (critical mass is relative!)
• Understand that this is part of the trade-off that you get for managing infrastructure and operational burden for the clients
• Add value to the ecosystem when you leverage SaaS network effect data

Based on this list, I decided to take a closer look at the feature in Ariba’s applications:.

• Make sure your User Agreement is up-to-date and is clear about your intentions (see your legal department)

Let’s take a look at Ariba’s Data Policy

Ariba may create high level statistical reports relating to the Solution utilizing Transaction Data, so long as such reports contain only anonymos, aggregated data that does not identify your company or any specific Transaction Data, and such reports may be reported publicly. Ariba offers benchmarking programs to facilitate deeper analysis into spend management practices for companies wishing to participate. Ariba operates the Ariba data analytics and Ariba benchmarking programs according to standards that protect the confidentiality of each customer’s information. These programs do enable Ariba to offer valuable reviews with customers demonstrating how that customer uses the Ariba Solutions as compared to typical usage of other customers by industry, company size, region or other factors. [SOURCE]

Looks good.

• Anonymize the data before leveraging it in aggregate

Based on the data policy, it looks like this condition is met as well.

• Leverage the information as soon as it becomes actionable (critical mass is relative!)

The combination of 15 years of data from all those customers in combination with HANA should meet this condition as well.

• Understand that this is part of the trade-off that you get for managing infrastructure and operational burden for the clients

The press release doesn’t provide any technical details about the use of this historical data but the use of HANA should prevent that the “operational burden” of such scenarios would be so great.

• Add value to the ecosystem when you leverage SaaS network effect data

Although there are few details, Ariba is providing this aggregate data to its customers so I’d call that adding value to the ecosystem.

Thus, it appears that Ariba meets the blogger’s requirements for SaaS network effect data usage.  If there are other legal requirements involved in such scenarios, this is a job for the corporate lawyers.

What about other SAP Cloud applications?

I’ve just focused on data exhaust scenarios in  Ariba but what about SAP’s other SaaS applications? Let’s take a look at the legal documents for such applications:

Fieldglass

Fieldglass collects, stores and uses information from the users of its application (meaning the computer software programs), and the hosted environment (meaning a hardware/software system combination under the control of Fieldglass on which the Product, or any portion thereof, is run). Aggregated versions of this information (information that does not personally identify you) may be used in many ways. For example, we may combine information about your usage patterns, with similar information obtained from other users to help enhance the application or hosted environment and our services. Additionally, Fieldglass may keep track of this information and the information that you provide to us for our internal review, business development, research, marketing and promotional purposes, press inquiries or aggregate statistical analysis, customization, demographic patterns, and historical trends. Aggregate information may occasionally be shared with our business partners and customers, and this information does not include any personally identifiable information about you or allow anyone to identify you individually. As Fieldglass adopts additional technology we may also gather information through other means. [SOURCE]

SuccessFactors Business Execution Suite

SAP may use Customer Data for the purpose of generally improving the look and feel of the Service as  well as for developing and distributing benchmarks and similar reports and databases therefrom so long as  SAP, for the latter, ℹ first removes all references to Customer and any personal information of a Named User included in the Customer Data; (ii) presents such Customer Data in the form of aggregate benchmarks; and (iii) generates and publishes benchmarks only when such benchmarks include data from at least seven (7)  other organizations. [SOURCE]

Precision Marketing

Consumer Data means data collected from or submitted by individuals accessing the Service via a device application or other digital interface, and transaction log data collected by the Service showing the purchase activity of such individuals. Consumer Data may include, without limitation, consumer preferences, buying habits and patterns, purchase locations, and other consumer behavior information. Customer agrees that SAP is free to use the Consumer Data for purposes of operating the Service, and to collect, store, analyze, process, display and distribute the Consumer Data in anonymized form as part of or independent of the Service during the term of the Order Form and thereafter without any obligation to compensate Customer or any other person or entity. [SOURCE]

Two Go

End User Data means data or content collected from or submitted by individual End Users accessing the Platform directly or by using a mobile device to remotely transmit such data to the Platform, and transaction log data collected by the Service. Customer agrees that SAP may use the End User Data in aggregated form to provide services to other customers provided neither individual End Users nor Customer can be identified from such aggregated End User Data.  [SOURCE]

I looked for such legal descriptions in the documentation for other SaaS applications from SAP but didn’t find anything of interest. This doesn’t mean that such disclaimers aren’t present but rather that my fast search didn’t find them or such passages may only be available in private contracts between SAP and customers.

Other SaaS Companies have similar ToCs

Such usage scenarios are also present for other SaaS applications. For example,

Zipinion

Zipinion may aggregate anonymous statistical data regarding use and functioning of its system by Customer and its users. Such aggregated statistical data will be the sole property of Zipinion. [SOURCE]

I also looked for such legal documentation for other major SaaS vendors (Workday, Salesforce, etc) but didn’t find anything that was publically available.

Analysis

SAP recent cloud  acquisitions are all established SaaS vendors older than 10 years.  The historical data accumulated by these firms is of real value and HANA provides the technical foundation to exploit this information. Indeed, this historical data might have been one motivation in acquiring such established firms.

Other companies such as ADP are already offering such information.  Some companies even promote this functionality as a competitive advantage:

As the first 100 percent cloud-based, multi-tenant compensation software solution in the marketplace, Xactly has the exclusive ability to aggregate, anonymize, and analyze customer data to provide the information companies need to make the most of their compensation dollars. This ever-expanding data set goes back to 2005 to deliver the meaningful analysis and actionable information the marketplace has demanded for decades.

You don’t just look at your own data, you can benchmark yourself against the market, and against your peers in your industry. THAT is usable, actionable Big Data. Plans, performance, payouts and positions compared across tens of thousands of data points over time.

Our Big Data platform is an aggregated view of Xactly data that is anonymized, not identifiable to any person or entity, and presented in a manner so that no customer’s identity may be derived. [SOURCE]

As long as the necessary legal preconditions are met, SAP should actively exploit the historical data that is stored in its existing and future SaaS applications.  As a showcase of cloud-based HANA functionality, it is too good an opportunity to miss.