Dell shares Self-Service Analytics with SAP HANA – ASUG Webcast
ASUG Volunteer Brian Durning, of Dell, shared this webcast with ASUG today.
Figure 1: Source: Brian Durning, Dell
Brian explained that Dell has been a solution provider for years, not just laptops, servers. Solutions provided are shown in Figure 1.
Figure 2: Source: Brian Durning, Dell
Figure 2 shows more of Dell’s service offerings. Customers include University of Kentucky, ATOS, Koehler, and Charmer Sunbelt. Brian also talked about Dell’s try and buy program for SAP HANA; contact your Dell Account rep for more details.
Figure 3: Source: Brian Durning, Dell
Why HANA? Brian said Dell looked at their IT strategy. They needed to make timely decisions in the areas of supply chain and customer service, to address issues as customer problems with a laptop. He said that Business Intelligence has been identified as an IT foundational capability gap by the Dell’s Business Architecture. “Dell’s Business Architecture Team looked across the enterprise for capabilities that were critical to enable Dell’s transformation from a PC company to a Solutions & Services provider” said Brian. He said that Business Intelligence is not a destination, but a transformational journey.
One of the key areas of that strategy is Self-Service BI, according to Brian, as Self-Service enables the business to discover the next KPI, look for data insights, answer one off business questions quickly and efficiently.
He added that the key to that thought is having the right governance to ensures that information is treated as a corporate asset. And the right architecture to insure that BI capabilities are delivered flexibly and in the most cost effective manner.
He also included this quote from Boris Evelson , Forrester Research, 2012
“More BI will move into the hands of end users. IT will learn not fight it or risk becoming irrelevant…Those BI platforms that support “managed” end user self-service will become more popular…Finding the right win-win combination that combines the flexibility and agility that self-service brings with behind-the-scenes monitoring and adjusting will become the name of the game.”
Need for Self-Service
Brian said that self-service allows the business and partners innovate while “driving down time to value and adoption of new insights and knowledge”. He said the needed faster ways to analyze data. They said IT can’t be the one stop shop and they needed to deliver a rich customer experience
They need to look at critical insights, drive market innovation – look at trends, calls to support center – are the memory chips burning out? They don’t want to be roadblock through having the business enter tickets and provide business justification
What kind of IT organization want to be? The want to be a value-add organization.
Figure 4: Source: Brian Durning, Dell
As Figure 4 shows, Brian said Dell was spending 80% of time on data movement and not on insights. Change was too slow. The cost to Dell was estimated to be ~$30 MM annually spent in manipulating and moving data. They wanted to look at ROI
Their average total was 47 billion rows extracted per week.
Also a lack of governance created inconsistent metric definition – CO-PA was an example.
Figure 5: Source: Brian Durning, Dell
Dell IT partnered with the business to build out the next generation BI & Analytics architecture, as they have a diverse data platforms to meet the variant needs of the business as shown in Figure 5. The purpose built marts are very focused on a business need.
On the bottom of Figure 5, HANA fits between the stream of thought and self-service rapid mart. Self-service rapid marts are built for business to do on their own.
In the center the Enterprise Data Warehouse meets the core enterprise needs.
This structure allows the business to reports on that data with views of the data that was relevant to them.
Figure 6: Source: Brian Durning, Dell
HANA is a spoke in data warehouse hub; it does not replace existing data warehouse – it augments it, said Brian.
With HANA, they use Lumira, Predictive Analysis, BI4 w/ Explorer. He explained that had 3 TB of raw data; with HANA compressed it was down to 70 GB
Figure 7: Source: Brian Durning, Dell
Figure 7 shows HANA Information Composer, which Dell used to upload data from local data store to HANA. Brian said views can be shared or not share but the intent is to share.
Figure 8: Source: Brian Durning, Dell
Figure 8 shows the use case; as an example, a manager has an Excel spreadsheet locally and can merge it with SAP HANA.
Figure 9: Source: Brian Durning, Dell
Figure 10: Source: Brian Durning, Dell
Figure 10 shows some restrictions due to it being a basic tool.
HANA Information Composer is only simple models; for complex go to HANA studio
Information is not accessed directly to tables but exposed to views, such as analytics views, attribute views and Calc views.
Figure 11: Source: Brian Durning, Dell
Figure 11 shows the benefits.
Brian said the business users achieved greater sense of self-service.
With rapid development happening, within minutes, user uploads and starts using and sharing.
He said the end users keep praising how easy it is, as they do not need a change request process.
With HANA Information Composer, they have a guided use of tool. They do not need a large IT infrastructure to support it.
Business users can easily add and combine local data
There is a change in mindset – “let’s go out to Teradata, create a report, extract, and then combine” – seeing trend the other way – tool is fast, efficient, speed of thought, slicing, and use Lumira, Predictive and Explorer. They use tools with little training.
Question & Answer
Q: How does the replication happen from Terra Data to HANA? Does SAP SLT Replication work with Terra data?
A: Chose to use Data Services as the tool to move data into HANA – HANA is spoke and hub.
Q: What is the Audit-Trail like for end-users up-loading excel-data into HANA…and then running self-service via BOBJ tools…?
A: On the backend posting the metadata information – posting it on web share as realize people have the same business needs. Want to reduce redundancy by reviewing the logs.
Q: When you say the tool is easy to use by the user, what kind of user are you talking about? Analysts, Managers, Sales…?
A: Did a sample across all users, from support secretary to VP’s and see if they could do this – large sample set
Q: How are you controlling data redundancy? Maintaining one Version of ‘truth’ etc. End users typically move on to next big thing.
A: Provide audits on data sets on HANA and look for redundancies. Will reach out to person and notify them of a similar dataset and discuss
Q: When you upload data into HANA, is there a need for you to reconcile any data?
A: Assume reconcile means validate – IT is not here to ensure data is correct; leave to business user to validate.
Q: Are there version control or data overwrite capabilities now, or planned for IC ?
A: Not sure what is on the IC roadmap – can overwrite data that is uploaded in view.
Q: What kind of training was given to the end users? Which tools and what was the delivery method?
A: Training for data modelers – 2 sessions – HA100 & HA300 from SAP. For end users – Lumira/Predictive – got their own training
Q: From a security authentication/authorization purpose can we leverage the SAP BO framework or need to rebuilt?
A: Needs to be rebuilt
Q: Do your HANA data is real time as your transactional data?
A: Not treating HANA as a transactional system. Did try replication – could do that without interruption to source system
Q: Does HANA reduce your costs?
A: It is a soft number; before spending $30M to load data; trying to reduce cost
You can meet Brian in person next week at the ASUG BusinessObjects User Conference: