Part 1 of this blog series depicted the differences and some subtle similarities between the well-established enterprise applications giant SAP and up-and-coming vendor Endeca Technologies. The post ended with the new fundamentals for the future of enterprise applications that were outlined at the Endeca Discover 2009 conference.
Part 2 of this blog series explored how SAP is adapting to the new fundamentals outlined in Part 1, especially with respect to the notion of BT or “business technology,” which denotes a pervasive technology in use by casual users and end users alike, increasingly managed outside the direct control of IT departments. I also explained the architecture of the recently unveiled SAP BusinessObjects Explorer product.
Part 3 continues with SAP BusinessObjects Explorer’s traits and areas for improvement, especially in terms of the user experience.
The SAP BusinessObjects Explorer Pitch-count: Strike One
In baseball lingo, by pitching SAP BusinessObjects Explorer to enable customers to load and manipulate enormous volumes of data into the in-memory database (where data can be accessed almost instantly), SAP has thrown “strike one.” Baseball aficionados will know that two more strikes are necessary to take one opposing player out, then 6 more strikes to retire the entire side in an inning, and then 24 more for the entire game. In other words, this product launch is a great beginning, but much more work remains.
During the lunch event in Boston (mentioned in Part 2), we saw two videos from marquee reference customers, since these customer representatives could not make it there in person. One executive from Molson Coors Brewing Company testified about how loading 900 million records into SAP NetWeaver BW Accelerator, and searching across all the records took less than three seconds.
The company’s business users learned how to use the software after only a “five-minute WebEx- or Netflix-like training.” For Molson Coors, the primary benefit has been the ability to reduce the time needed for a business user to make an informed decision from one day to one hour.
Another executive from Sara Lee Corporation conveyed similar sentiments and experiences. He said how the company was able to load about 300 million rows of production data into the accelerated version of SAP BusinessObjects Explorer (explained in Part 2).
The team reportedly deployed the software in three days and gave it to business users without providing any training per se. There was immediate user acceptance, which was atypical compared to their prior experience with other heavy-duty SAP products.
More Pre-packaged Content, Please
The SAP BusinessObjects Explorer front end was built using Adobe’s Flex open-source tool for building Rich Internet Applications (RIAs). Thus, “strike one” for SAP was the delivery of a product that requires no training for end-users.
Indeed, if you can shop online or search on Google, you can pretty much use SAP BusinessObjects Explorer straight away, as I too had the chance to learn during the event. There in no need for previous training on data sources and structures, since the users don’t necessarily need to know where the data is located or how to select members (rows and columns) of built queries.
But to achieve more strikes and win the entire game, SAP needs to provide more packaged content in terms of business scenarios, applications, and templates. Certainly, the ability to load and browse a gazillion records is commendable, but what useful info for different user roles can quickly be discerned from there?
Some low-hanging fruit role-based scenarios could be performed by users that regularly need “finger-tip” access to information, such as a sales executive running to a customer meeting and wanting to know what products the client has bought and if they have any open support calls. Or perhaps product managers that are on the phone or on the Internet chatting with colleagues and needing to know the sales of the product that they each manage by product version.
Other no-brainer scenarios would be end users that need to better grasp their business, such as customer service representatives (CSRs) that need to better understand customer satisfaction by region, product, version, language, etc. Or perhaps marketing folks that need to understand which campaigns were successful by demographic or regional information, campaign type, leads generated, and cost. In his article, Bruce Richardson of AMR Research also thinks it wouldn’t be too hard to imagine businesspeople deploying the tool to instantly calculate customer profitability by product or region, compare forecasts to actual results, or argue for or against various stock-keeping unit (SKU) rationalization proposals based on data (rather than instinct).
In fact, SAP pointedly asked the customers during the lunch event whether they have conducted any preliminary return on investment (ROI) analyses on the benefits of deploying SAP BusinessObject Explorer. The blank gazes and stares from the audience have told me that the 65-or-so current live customers of the product (whereas the SAP NetWeaver BW Accelerator install base is about 600) have probably been early adopters, rather.
The product is currently only used for viewing purposes and users cannot work on and change “live” data within SAP Business Suite’s (or SAP ERP’s) instances. The initial release (Wave 1) of SAP BusinessObjects Explorer does not support SAP BW data and reporting hierarchies either.
While one positive aspect of this is that there is no danger that the less clued-in users will inadvertently mess up the mission-critical system, what the product can possibly do remains limited without prepackaged tools and applications. At this stage, the product does not feature strong analytic capabilities in terms of providing non-cumulative key figures, exception aggregation, calculated measures, basic line-wise arithmetic calculations, restricted key figures, modeling based on value or value type, etc.
Wave 2, which has not been released yet officially, will reportedly mitigate most of those shortcomings of Wave 1. SAP’s process of introducing a product to market is called “ramp-up,” and it begins with a release to customers, which in this case is expected to be in mid-November.
Back to Comparisons with Endeca
The recently announced Endeca Manufacturing Suite (with procurement, spend analysis, etc.) is based on Endeca’s general information discovery concepts as explained in Part 1, but the vendor also provides the specific functionality needed for spend analysis solutions. As Jason Busch said is his blog post, the idea is to create a revolutionary procurement tool, with a functional design and user interface (UI) that helps unearth opportunities that were previously impossible to find.
This additional spend analysis functionality comes from built-in analytics and works as follows: in addition to seeing all the search results listed, the user can also see some calculated values based on the returned records. For instance, users can see all the parts bought in March 2009 in Latin America and their average and cumulative value. There is a variety of calculation functions over the visible set of data, and these functions range from simple algebraic calculations (sums, max/min values, etc.) to very complex formulas and visual graphs and charts.
Thus, users can instantly get not only a list of records but also visual graphs, and they can also continue to navigate and select sub-sets of the records. As soon as the new data sub-set is selected, the graph is instantly updated. These spend analytic capabilities are sold as an additional module on top of Endeca’s MDEX environment, at a fraction of the basic MDEX Engine price.
Becoming an Application Directory Platform?
Therefore, I concur with Bruce Richardson’s assertion that SAP BusinessObjects Explorer would make a great platform for third-party applications like Jonova’s business planning software, Steelwedge’s sales & operations planning (S&OP) applications, or Kinaxis RapidResponse. Imagine having the power and data needed to model and simulate the impact of moving to direct store delivery (DSD) or to load actual point-of-sale (POS) data versus forecasted demand data from thousands of retail outlets and plan near real-time replenishment scenarios based on balancing costs and demand.
But for users to be able to model data and scenarios via additional third-party applications, SAP needs to publish the application programming interfaces (APIs) or provide some developer tools (studios) and software developer kits (SDKs) for third-party providers to extend the SAP BusinessObjects Explorer foundation. I’m not sure whether and where that fits on the roadmap.
For now, SAP is mulling over the idea of providing APIs for users to develop their own (or use third-party) visualization tools. Additional future integration aspects for SAP BusinessObjects Explorer could be to open up SAP BW Accelerator to other non-SAP data sources through SAP BusinessObjects Data Services.
Using Endeca to illustrate another point, that vendor provides business tools for ordinary users to be able to spruce up the application. One tool is called content spotlighting, which is the ability to configure the search engine to offer some alternative suggestions besides the retrieved list results. For instance, if a user is looking for flights to Boston, the engine could suggest a few cheaper flights to the nearby Providence or Manchester airports.
Dynamic merchandizing is a variant of content spotlighting that is used by online retailers and sellers. For instance, a consumer is looking for a Blackberry phone, and the site returns the results with some iPhone suggestions on the side. Or, an online bookstore will feature on top of the page the book of the month in the genre that the consumer is interested in.
Another user tool is the page builder to create landing pages or topic pages, as mentioned in Part 1. Namely, instead of the list of possible results, the user is rather directed to a specially designed page for the query. For example, a consumer is looking for an iPhone model, and instead of the list of models, here comes a dedicated page for all Apple products. In Endeca’s lingo, a “cartridge” is a component of a topic page, e.g., a dynamic picture or banner that changes every few seconds.
Make no mistake, Endeca Information Access Platform (IAP) and SAP BusinessObjects Explorer are playing in different leagues (MLB and AA in the baseball lingo, respectively) and are used by different user audiences. SAP Explorer’s front-end UI is only focused on structured data.
Still, SAP might want to use Endeca for inspirational purposes. At least, the current keyword search should be bolstered with rich search abilities like data-driven spell correction, type-ahead search matching, faceted search, “Did you mean” prompts, and search results with guided navigation.
Endeca features all these nifty capabilities as differentiators between an overwhelmingly long list of search results (e.g. Google) and a “rich search” experience. For example, the type-ahead feature actively suggests alternatives while the user is still typing the inquiry.
SAP BusinessObjects Explorer in the Cloud?
I might be pushing my luck by suggesting SAP enable the product for cloud computing and software as a service (SaaS). At least the accelerated version would not be suitable for this mode, since multi-tenancy would not support the accelerated in-memory database performance and dedicated server blades.
Still, there is an interesting set of slides and a preview of “Explorer in the cloud” on the SAP Developer Network (SDN) here. Also on this test-drive site, SAP allows prospective customers to set up an account in a hosted environment for 21 days. Users can upload datasets, navigate with SAP BusinessObjects Explorer, and share their findings with others.
Now SAP BW Accelerator is not the only product on the market that offers a high re-usability of plain relational databases together with fast performance of canned reports and report cashes. For example, Oracle Exadata also competes in the database appliance market. However, the last time I checked, Oracle’s client-side reporting tools like Oracle Business Intelligence Discoverer still require some user savvy and training before use.
At the end of the day, dear readers, your comments with regard to the opinions and assertions expressed thus far are welcome. How important are the above-mentioned considerations in your software selections? What are your information access and discovery needs and experiences?