How a Telecom Company Leverages SAP HANA in the Amazon Cloud to Simplify their Landscape and Accelerate Business
Our client distributes communication products from about 380 manufacturers to over 100 countries. Their products include network system products, base station infrastructure equipment, installation test and maintenance equipment, and mobile devices and accessories. Managing over $750M a year in sales across a wide variety of products, regions, and sales reps is a labor intensive process; however, it is fundamental in driving profitability.
The Client’s Need
With lots of sales data collected, the client became quite good at performing analytics for operating decisions. Unfortunately, the data was stored across several systems, often taking days to get a clear picture of the business. Some of the data is created or updated every few seconds, so the data becomes stale almost immediately after it was downloaded. The client needed a high speed solution to pull all their data together, perform data cleansing and transformations on the fly to harmonize the data, and produce data models that would enable end users to query the system to find answers to their business critical questions. Hana was the solution. They didn’t just want a turnkey reporting tool, they also needed to know how to build the solution and support it. An SAP ETL (Extraction, Transformation, and Loading) developer, SAP Hana developer, and SAP XS (eXtended Services) developer were engaged to kick off this project. The deliverable for Phase One was as follows:
- A Commissions Dashboard
Commission plans grow in complexity the more you tailor them to align the sales representatives’ objectives with that of the organization. As a result, it’s easy to under or overestimates a reps payout until all the numbers have been tallied after the period close. This dashboard was designed with the capability of assessing commissions in real-time for each sales rep. With this information, it would enable the client to help motivate sales reps who were approaching bonus targets. It would also provide a meaningful comparison between sales reps that could be used as a status check.
- A Sales Transaction Performance Dashboard
The client knew each customer’s purchasing history could be profiled and trended. Furthermore, they knew buying certain products created a need for complimentary products. For example, if you buy a cell phone, you probably also need a case for the phone, a charger, and ear buds. Customers were now always buying the complimentary products from our client. This dashboard offered a perspective across all purchases to enable analysts to spot trends that could be used to create a customer purchase basket. It further enabled the categorization of customer accounts into profiles for comparison of similar customers to identify missed sales opportunities.
- A Historical Sales Ownership Dashboard (what-if scenarios)
With customer accounts changing hands, multiparty deals, and the client’s legacy system’s inability to track account ownership across history; this dashboard enables management to slice the account history in various ways to better allocate account ownership to provide a more equitable compensation distribution. When a sales rep is rewarded for all contributions towards servicing an account, it is less likely that the sales rep will act solely in their own interest by performing only the activities that contribute to their commission.
From On-Premise into the Cloud
Prior to the project, the landscape consisted of about 15 tables across five databases and source systems with a mixture of them being on premise and third-party systems.
In the new landscape, Data Services and Hana were installed on the Amazon cloud (AWS). Source data was extracted, transformed and loaded into Hana. Hana Models were created and exposed to the XS engine for consumption and presentation to the user community.
Starting mid-year with these three dashboards was possible but not practical. Therefore, a deadline to get Phase One productionalized was set to coincide with the client’s fiscal year end. Everyone knew the timeline for phase one was extremely aggressive, but the benefit would be worth it in the end.
Going into the Phase One development the functional design of the dashboards were about 85% complete, and Technical designs were about 20% complete. In some cases we had to identify source systems and tables, in other cases we had to standardize the data and cleanse it before it could be incorporated into the Hana data models. Sometimes, logic had to be created for newly defined and evolving design requirements. Ultimately, the success of our project was riding on our ability to communicate effectively as a team, and drive new requirements across each functional area quickly.
Finally, the what-if functionality built into the Historical Sales Ownership Dashboard was extremely complex because the data had to be aggregated differently before being tallied, depending on the how client choose to run the scenario.
Customer Value: In-situ Training, Streamlined Development, and the Model – Model
Accomplishing the workload by the six week deadline was going to be a real challenge, and we knew we had to work efficiently. Each developer was paired with one or two members of the client’s team who were expected to take ownership of that project function after SAP consultants were gone. As we began new tasks, we provided a short tutorial of the tool, gave them a task, and then we developed around them until they got the hang of it. That approach enabled them to quickly ramp up their skills.
Every day we worked in a war room in our functional groups across the table from the other groups. Our work stream was a highly collaborative data-pull model. To get the project started quickly, the source tables needed to be loaded to Hana as quickly as possible so the Hana Modelers could begin modeling. Development of the load jobs to the cloud were expedited using the Data Services workbench. As the Hana modelers came across data cleansing issues and data transformation needs, those requests were handled with top priority by the ETL developers. As the XS developers came across modeling issues, those requests were handled with priority by the Hana modelers. Optimization, tuning, and documentation work were performed with a lower priority as time allowed.
You can only build in so much functionality into a data model to increase the system’s flexibility before begining to erode the system’s efficiency. We reached this limitation trying to model the what-if scenario in the Historical Sales Ownership Dashboard. Finally, the accepted solution was to create multiple similar models to separate the complex aggregation of each sales team role per account.
Beyond Phase One
Better organization of teams paved the way to better execution not only in our workstream approach for this phase of the project, but also for the remaining phases.
Several experienced colleagues questioned the tight project deadlines. We even considered asking the client to reduce the scope of the project to allow sufficient time for development and training. Providing knowledge transfer on an ongoing basis is just another example of how SAP consultants find creative ways to get things done. However, providing ongoing knowledge transfer may be a double-edged sword. Since our client was quite tech savvy, they were able to use the demos to build upon their understanding; however, a client with less of a technical understanding might have become frustrated or disinterested without first being able to see the big picture.
I advised our client to shift their mindset on how they use Hana for their upcoming project phases to unleash some of Hana’s untapped potential. Sure it can aggregate the data from several systems faster than ever before. Sure it can be used to build and query larger and more complex data models than ever before. However, if you use Hana to solve the same questions you were asking your old systems – only faster; you will miss out on one of the greatest advantages Hana has offer. Redefining what information you want your information system to provide on broader terms is how to take it to the next level. For example, instead of asking how can Hana provide the data faster so I can perform a manual analysis to help me make a business decision; you need to ask how Hana can be used to collect the data, automate the decision process, weigh the alternatives and select the optional plan of action. Incorporating more of the decision logic that traditionally has been performed manually will accelerate business processed and produce a game changing advantage for our client.
To learn more about how SAP HANA Services can help you can eliminate barriers to fast, informed decisions, please visit us online: http://www.sap.com/services-support/svc/in-memory-computing.html