Sales and Marketing: Analytics in the Cloud

Originally published March 24, 2009

I’d like to thank Patrick Callahan, Senior Director at AmberLeaf for his contributions to this article.

Based on the interest from my last article on hosted analytics, I decided to continue the trend and talk about businessintelligence in a software as a service (SaaS) mode. Customer relationship management (CRM) applications flocked to SaaS models based on the success of salesforce.com. Traditional on-premise vendors like Siebel, SalesLogix, Microsoft, Oracle, Onyx and the rest followed suit by providing SaaS solutions. The lure of no administration, no hassle upgrades, low switching barriers and fastimplementation times attracted many customers into the new model.

However, business intelligence, database marketing and data warehousing have typically lagged these types of trends. It reminds me of the shift to the offshore implementation and support model. Datawarehousing lagged this trend due to the closer relationship between users and their support staff, the numerous requests for small enhancements that need to be performed quickly and the massiveamount of systems integration between legacy systems.

Trepidation with SaaS models for business intelligence (BI) appeared for the same reasoning. As business intelligence became the center of large data integration and data quality efforts, concernsover the scalability and sophistication of SaaS vendors in this regard raised cautionary flags. In fairness, CRM systems are not without their data integration challenges, and the SaaS vendors haverisen to the occasion with APIs and other connectivity tools.

Nevertheless, progress cannot be stopped. It is BI’s turn to flock to the cloud and the software industry is responding.

Data Warehouse Infrastructure

The first major technologies from BI SaaS models are coming from the infrastructure side. New startups are aiming at the Teradata/Netezza marketplace as highly scalable, incredibly fast, query(and data loading) processors. Using proprietary databases (sometimes columnar, inverted index structures with unique compression algorithms, sometimes with proprietary hardware, sometimes not), SaaSvendors are raising the bar in price per performance benchmarks. Though Netezza is probably still the leader in the performance category, Netezzainstallations charge a large premium for the effort.

Vertica may be the leader in this category. With some formalized benchmarks against Oracle, some price/performance studies against Netezza, andover 60 clients, they have grabbed some early market share. They have opted for an open architecture by creating an ODBC interface to their database so you can continue using your favorite reportingapplication. Business Objects and Cognos on the business intelligence side and Informatica on the ETL side have all certified on the Vetrica engine. Some new skills will be learned as the physicalmodel for a columnar architecture is slightly different than a typical relational model. Vertica provides a data administration application that lets modelers design the logical model and Verticaconverts and generates the physical model.

Birst differs from Vertica in the sense that it provides an automated solution with its own front end to its database and a data loader. Basically,you load up your data, and Birst figures out the optimal data model given its technology. The metadata is captured and automatically generates the user interface for the reporting piece, with sometailoring capabilities available. With its dashboard, Birst’s SaaS offering provides a consistent and central location for reporting output, making it easily accessible to many consumers.

With similar aspects to Vertica’s and Birst’s solutions, PivotLink’s SaaS delivery model includes a columnar architecture for their data store for performance, integrating data from various sources, and an optimized front-end to generate and deliver reports and dashboards for on-demand needs. With the front-end capabilities, PivotLink also provides report subscription and alert capabilities. Originally started as SeaTab in 1998, PivotLink has a solid customer list and several recent awards; PivotLink offers a strong SaaS BI option.

Other niche players worth keeping tabs on in this space are Information Builders, Oco and Good Data.

Depending on the organization, you might find the open architecture of Vertica or the turnkey aspects of Birst and PivotLink, or the vertical analytics of LucidEra more advantageous. Each promote jumpstart implementations and offer attractive pricing options and each should be closely evaluated for your specific BI needs.The SaaS model and each of these solutions are not right for everyone – if they do fit, they provide a nice advantage. Consider the following when evaluating SaaS:
  • Open architecture to utilize standard BI (ETL, reporting, etc.) tools

  • Query performance

  • Data load performance

  • Dashboard and reporting sophistication

  • Pricing options
In the end, each of these organizations sees their fit in your business intelligence architecture similarly:
  • For smaller organizations who have not adopted business intelligence just yet, these applications can act as your enterprise data warehouse or a federation of data marts.

  • For larger organizations, these services can act as specialized data marts with superior performance for extraordinarily demanding users, huge volumes of users, or users with heavy exploratory requirements.

  • For all organizations, offering solutions with shorter implementation timelines and less IT involvement.

Applications

Analytical applications in a SaaS model is the other type of genre I have noticed in the marketplace. Since many CRM tools come in SaaS, their reporting applications also come in the sameformat. However, specialized analytical applications have surfaced for specific business processes.

Probably most notable is Omniture as a web analytics provider. Omniture is by far the market leader in web analytics (maybe except forWebTrends and/or NetRatings). They provide management reports, ad hoc query and lately a suite of applications to help test your website design. For instance, if you are thinking of changing your registration page, you can use their suite to perform an A/B (or far more complicated multivariate) test and monitor the results.

Hitwise, Compete, and Coremetrics provide countrywide (and in some cases international) web analytics across organizations. You can use these applications to competitively benchmarkyour website activity against your competitors. You can see if you have spikes at the same time as your competitors or if changes in your website stole page consumption.

To illustrate how important analytics is to on-demand applications, VerticalResponse has recently overhauled the analytical component oftheir core email application. VerticalResponse is one of the leaders in providing marketing technology for small businesses. They now provide benchmarks and scorecards so their customers can evaluatethe effectiveness of their campaigns, offers, creative and overall approach. They can essentially rank their campaigns and strategies in order to increase response rates.

Maybe you can even consider Panorama’s deal with Google Apps and Salesforce.com a hosted analytical application. Panorama has recently embedded their query technology into Google Spreadsheetsto help make Google Apps more Excel-like. Base functionality is free and advanced functionality is only $29/year.

LucidEra is my final solution in this group, offering SaaS BI analytics for ERM and CRM needs, having available pre-built vertical solutions or more general use of their enterprise solution. Diverging somewhat from the prior solutions, LucidEra provides connectors for data integration and targets the ease of setup and use for their solutionto very specific CRM (salesfoce.com) and ERM (Oracle) focuses. Like most of the others, LucidEra’s executive team has a strong pedigree in the business intelligence and CRM arenas, and their direction has strengthened their position in the market.

I think the idea of pre-canned, hosted applications for marketing response, sales targeting, sales analysis or customer analytics may provide a huge value for organizations – especially ifthese applications have plug-ins to popular CRM and marketing automation tools like Siebel, salesforce.com, RightNow and Unica.

Summary

Though there is still hefty concern (for good reason) about the ease of enterprise integration, quick changes and flexibility with business intelligence SaaS solutions, this wave will continueto increase as more customers opt for the model. I think that the infrastructure vendors will need to prove that time to market does decrease, that this is just not a total cost of ownership play andthat they can make their information pervasive throughout the organization (possibly in near real-time) if need be. Regardless, the several recent acquisitions by the BI behemoths (IBM and Cognos,for one) have altered the BI industry landscape, and the SaaS solutions mentioned in this article already have several competitive advantages that will make it difficult for the larger organizationsto catch up in the short term.

The issue with the application vendors is the long-term aspect of integrating their application with the rest of the organization’s data. For instance, for user level behavior, many clients wind up extracting detailed data from Omniture into their corporate BI systems. Is that effort worthwhile or should they bring web analytics completely in house? Point applications on a hosted levelwill need to have easy ways of integrating data with the enterprise so they don’t become just another island of analytics.


  • Larry GoldmanLarry Goldman
    Larry has more than 15 years experience in database marketing, customer relationship management, business intelligence and analytics. A well-known speaker and author, he has been a regular contributor to industry publications for almost 10 years. With experience across multiple industries, Larry helps his clients create new business processes, sales and marketing strategies, analytical plans, contact strategies and customer experiences. And with his extensive technology background, he helps operationalize these strategies by ensuring they can be practically implemented. Larry can reached at 773-456-3996 or larry@amberleaf.net.

    Editor's Note: More articles and resources are available in Larry's BeyeNETWORK Expert Channel. Be sure to visit today!

Recent articles by Larry Goldman

 

Comments

Want to post a comment? Login or become a member today!

Be the first to comment!