Got a thought or idea for our BI Bloggers ? Email biinactionblog@ebizq.net
BI in Action Blog
|
« Informatica Jumps Into Bed with SAP | Main | Oracle 11g -- Not "The BI Release" You Were Hoping For » July 25, 2007Breaking BI Bread, or Mixing Livestock Fodder?
Vendors and analysts continue to talk up various trends in business intelligence. But what's real and what's hype? That got me thinking about bread…yes the kind you toast and butter. Bread is made from wheat, which is first separated from the chaff by a thresher-like machine. Wheat is useful. Chaff is, well, chaff. So I thought I'd apply my own mental thresher to some of the hyped up trends that vendors and analysts are talking up today. Will they become bread or livestock fodder? Data Warehouse Appliances Wheat or Chaff? Love them or hate them, data warehouse appliances are not a passing fad and are here to stay. Driven by the relentless growth in transactions and data volumes, and riding on the back of Moore's technology cost curve, scalable and performance-optimized data warehouse appliances have the potential to add value to existing EDW initiatives without expensive upgrades. Appliances hitting the market today are unrecognizable from the proprietary and expensive appliances of yesteryear. A new generation of turnkey appliances built on commodity hardware and open source software components is starting to flood the data-warehousing market. Start-ups like Netezza and Datallegro have worked hard to make appliances a viable product category. Interest in appliances from big vendors like IBM, Hewlett-Packard, and Sun, has more or less validated the market. The next challenge is to create a critical mass of customers. Some big wins are expected in 2007. Most of these will be as an adjunct to enterprise data warehouses, but perhaps some will take business from market incumbents like Teradata, IBM, and Oracle. There will also be some failures. Getting the appliance model right is tricky, and several promising appliance start-ups have already dropped out after trying to plug in appliances built on proprietary platforms. Open Source BI Wheat or Chaff? It is hard to tell. Open source BI is aimed at lower-end of the market, namely basic reporting and the Java development community (as is the case with Actuate), rather than complex analysis or large enterprise-scale BI deployments. It's true that JasperSoft and Pentaho are assembling more ambitious BI suites, but they still tend to be feature-limited, and are unlikely to overtake and replace Business Objects or Cognos installations anytime soon. The question is whether this is something that customers will be prepared to wait for? Complex and enterprise-scale BI and analytics will traditionally remain licensed software for the next five years at least. Mobile BI But isn't this deja vu? Remember portable, wireless, mobile, or whatever you might wish to call it, BI was all the rage five years ago at the height of WAP-mania and the start of the PDA device boom. Back then software vendors scrambled to recreate their desktop interfaces on mobile devices, without giving much thought to the value and usefulness of doing so. But the idea has persisted and seems to have a second wind with Cognos, Business Objects, and MicroStrategy all launching fresh mobility features that plug into their core BI platforms in the hope of making BI more pervasive. Wheat or Chaff? Mobile BI has promised to go anywhere, but so far it hasn't. What stopped it doing so in the past was less about the technology (the bandwidth and rendering issues are being fixed) and more about presenting a compelling business case for the investment. That still holds true today. Only a few specialized functions really benefit from mobile BI, notably field sales and support optimization, territory analysis, and customer communications/management. Unless someone comes up with a killer app for mobile BI then it will simply be another bell and whistle function of your BI platform. On-Demand BI They have a strong case. Established BI software providers like SAS Institute, Business Objects, and Cognos, as well as new entrants like Oco Software, Seatab, Host Analytics, and LucidEra, have all recently made SaaS BI plays. And Informatica, in partnership with SaaS poster child Salesforce.com, is also pioneering the delivery of hosted integration services to systems integrators and business process outsourcers. Wheat or Chaff? SaaS, like open source BI, is still a nascent trend in BI. It is gaining in popularity, particularly at the lower end of the market where cash-strapped and IT resource-challenged SMBs are looking for painless entry into BI. In many ways SaaS represents a significant paradigm shift in how enterprise software is deployed and managed. It is a direct response to customer demand for less expensive software licensing, simpler implementation, and more widespread adoption. BI fits the bill well since it has traditionally been expensive and complex to deploy and has yet to penetrate broader business user audiences. It is too early to gauge the impact of this software delivery model. The newer start-ups need to acquire a critical mass of customers in the SMB segment. But these vendors are also up against a formidable and entrenched competitor in the SMB space, namely Microsoft and its SQL Server 2005-based BI system. At face value, SaaS offers a much cheaper and easier alternative than on-premise software. But moving to SaaS is not as easy as simply having your BI software hosted. Companies should carefully consider several issues, not least security implications that are more pronounced in BI given the sensitive nature of the data, managing organizations' access to the data and applications through multiple client tenancy arrangements, and how a vendor's SaaS architecture fits with their IT standards and its impact on interoperability with their own internal systems, especially enterprise BI systems already in place. Vendors offering their BI solutions as SaaS must also overcome the significant hurdles of data privacy, reliability of service, working out commissions for channel partners, and grappling with a new sales model. Moreover, the majority of SaaS deployments continue to be focused on individual departmental initiatives. No provider yet offers the functionality or end-to end process management awareness of on-premise software to support the analysis of cross-departmental business flows. Finally, there is the issue of customizability, which has always been the bugbear of SaaS applications, and one reason why most SaaS BI offerings today are best suited for "commoditized" BI tasks like query and reporting. But BI delivered as SaaS is also likely to get uptake from larger enterprises that have prior experience with on-demand CRM solutions. Posted by madansheina in Business Intelligence | Digg This | Add to del.icio.us Trackback Pings TrackBack URL for this entry: I agree that SaaS isn't for everyone, but it is a good option or alternative for mid-level companies who may not be able to afford the initial investment in on-premise software. On-demand is a good way to get proof points or to develop business justification for making larger investments in BI infrastructure. Certainly the commonly cited upsides of SaaS such as quicker time-to-market and lack of hosting overhead issues do actually translate into value-add for certain organizations. It will be interesting to see how this SaaS develops in the BI space. Posted by: Laura Wang at July 27, 2007 09:24 PM I agree with Laura, we still have time to see how SaaS develops in BI Space. Posted by: Jay at July 31, 2007 01:42 PM Post a comment |
