Welcome to TDWI FlashPoint. In this issue, Jill Dyché prognosticates BI and DM developments in 2008.
- FlashPoint Snapshot
Predictive Analytics: Extending the Value of Your Data Warehousing Investment
In Store for ’08: Some New Year Prognostications
- FlashPoint Rx
Don't Assume that EAI Can Replace the Data Warehouse
FlashPoint Snapshots highlight key findings from TDWI's wide variety of research.
What Does Your Group Do with the Models It Creates?
Based on 166 respondents selecting multiple answers.
Which Best Describes Your Group’s Approach to Model Management?
Based on 164 respondents.
Source: Predictive Analytics: Extending the Value of Your Data Warehousing Investment TDWI Best Practices Report, Q1, 2007. Click here to access the report.
Based on 833 total respondents. Rounding and multi-choice questions are responsible for percent totals that do not equal exactly 100 percent.
In Store for ’08: Some New Year Prognostications
Jill Dyché, Baseline Consulting
It’s no surprise we’re all a bit suspicious of New Year predictions. Britney Spears and Barry Bonds are easy targets. Forecasting the demise of Fidel Castro has become positively trite. And I finally stopped predicting the federal government’s comprehensive Y2K plan sometime around 2005. But we’re seeing developments in business intelligence and data management played out in our client work that may indeed represent coming attractions if not emerging trends. Here are 10 trends I’ll be watching this year.
Prediction Number 1
The emergence of “data as a service.” We’ll forgo the unwieldy acronym here, but data as a service has legs. As corporate purchases of packaged applications continue, the need for the data these systems generate has become more urgent. When combined with the continued confusion around service oriented architecture (SOA), data as a service becomes an elegant and practical way of distributing data to the applications that need it. Data as a service can take the form of standardizing a data cleansing platform available via Web services to multiple applications. Or, more likely, it will take the form of master data management, in which data creation and access services are managed in a centralized and rules-driven way.
Prediction Number 2
One step forward, two steps back for operational reporting. We’ve watched many companies turn their collective head toward more right-time, on-demand BI. Your company is probably one of them. Several of our clients have allocated additional ’08 funding to beef up operational reports so they can respond to business events in a more timely and relevant way. But just when we thought it was about performance, we realize that the real bugbear of operational reporting is data quality. We predict that companies will postpone increasing investments in operational BI projects until they are more comfortable with the quality and integration of their data.
Prediction Number 3
Executives endorse enterprise data management. Companies talk a lot about data management. Many now practice it, but usually at the application or project level. This is about to change. The economies of scale promised by centralizing a data management function within IT—ideally reporting directly to the CIO—means not only that fewer specialists practice data management, but they also do so using formalized processes, established data modeling and metadata conventions, and sanctioned business rules. We’ve watched early adopter companies quantify the “people costs” of isolated data management efforts, then proceed to recoup these costs by introducing a formal enterprise data management organization, cleaning, integrating, and deploying better data faster in the bargain.
Prediction Number 4
Data warehouse appliances make the cut. We’ve written before about the double-edged sword of approved vendor lists. On the one hand, they make the entire procurement process a lot easier. On the other hand, they limit a company’s negotiating power by narrowing the vendor field and may actually inhibit a “best of breed” choice by omitting key vendors from the list. The latter situation has begun to erode in the data warehousing world, with upstart data warehouse appliance vendors being given increasing consideration as companies look to avoid costly database upgrades and mitigate their reliance on “the big guys.” Vendors such as Netezza, Paraccel, and Dataupia have shown up on the radars of some of the most risk-averse companies we know. Data warehouse appliances are here to stay, and they’ll be more ubiquitous in 2008.
Prediction Number 5
Data governance gets in gear. When it comes to data governance, 2007 was the year of the tire-kicker. People wanted to see it, peer under the hood, and maybe even take it for a test drive. In 2008 companies will accelerate their data governance efforts, but in low gear. Challenges of finding an executive sponsor and assembling a council will give way to the realization that data governance is collaborative and systemic. Managers who see an obvious need for data governance will go beyond forming a committee and getting to the root of organizational dysfunction and process breakdowns. Data governance is neither project nor program, committee nor competency center. It’s a process that needs to be retrofitted into the development lifecycle, driven by the business, and executed by IT to address business issues that won’t go away anytime soon.
Prediction Number 6
CIOs get behind BI for business-IT alignment. This is the year that CIOs—normally encumbered with the demands of operational systems and technology-wary executives—understand that business intelligence is their ticket out of the “IT as commodity” corner they’ve been painted into lately. For one thing, CIOs need the career boost that BI can provide. Delivering a more enhanced understanding of customer behaviors and preferences, streamlining the supply chain via faster data deployment, or predicting financial risks can make a name for even the geekiest CIO. Furthermore, successful BI projects are fundamentally requirements-driven, thus inviting a dialog between the business and IT that may be absent from the workaday development activities of transaction-based systems.
Prediction Number 7
Master data management breeds success stories. When I wrote The CRM Handbook back in 2002, it wasn’t hard to find companies that had redesigned their customer-facing business processes and reaped the rewards to talk about their ROI. (Think Verizon, Harrah’s, and Eddie Bauer, to name a few.) A few years later, we had to scrounge around to find companies that had even planned for MDM let alone delivered it. The case studies that made the cut for our Customer Data Integration book were from early-adopter companies, some of which had to create their own MDM infrastructures in advance of the market’s hype. In 2008 we’ll begin hearing stories about operational data integration via MDM in action, and the resulting business successes that have accompanied them.
Prediction Number 8
Data governance drives IT governance. I’m surprised by how many of my firm’s data governance projects are with clients that have no incumbent governance. Data governance is more often than not the company’s first foray into the structured and deliberate management of a corporate asset. We define data governance as a policy-making framework, and as companies watch data governance work they will likely lift their heads and see the potential for governance mechanisms to address other assets, such as the IT portfolio.
Prediction Number 9
BI drives innovation. With the economy yawning and the cry for shareholder value continuing unabated, executives will be refocusing on strategic value which, we believe, rests with superior business intelligence. With the inexhaustible supply of information and tools with which to exploit it, executives will enlist their best and brightest to make new discoveries about their customers, their products and their packaging, their corporate risk, and their competitive differentiation. The resulting breakthroughs will hinge on the smart use of good data. That’s simply great BI at work.
Prediction Number 10
In this age of BI market consolidation, it would be wrong not to predict at least one joint venture, so here it is: Starbucks and Netflix will team up to deliver coffee to your door in a red postage-paid envelope. You can say you heard it here first!
is a partner and cofounder of Baseline Consulting
, a technology and management consulting firm specializing in data integration and business analytics. Jill is the author of three acclaimed business books, the latest of which is Customer Data Integration: Reaching a Single Version of the Truth
(John Wiley & Sons, 2006), coauthored with Evan Levy. She will be teaching her popular workshop, BI from Both Sides: Aligning Business and IT, at the TDWI World Conference in Las Vegas on February 20, and is the co-chair of the Master Data Management Insight conference
, to be held in Savannah on March 2–4. Visit the Web site
to qualify for attendance.
FlashPoint Rx prescribes a "Mistake to Avoid" for business intelligence and data warehousing professionals from TDWI's Ten Mistakes to Avoid series.
Ten Mistakes to Avoid When Considering Data Warehouse Alternatives
Mistake 4. Assuming that EAI Can Replace the Data Warehouse
By Evan Levy, Baseline Consulting
I recently sat down with a team of IT architects to lead a review of a new application infrastructure. The company was relying on several packaged application systems to support their business operations and wanted to figure out if there was a simple and elegant way of migrating data into and out of these systems. The business users wanted quick access to data to support their needs; call center staff, store-based customer support representatives, and management staff wanted operational reporting details; and IT wanted to provide them a data access solution that could leverage their existing infrastructure in a cost-effective manner.
The team had successfully leveraged enterprise application integration (EAI) technology to connect their packaged applications. EAI allows for the intelligent migration and sharing of data between applications. When an order for a new customer was placed, information was passed between the CRM and ERP systems to ensure customer discounts, invoicing, sales commission, and inventory were all processed correctly. The EAI solution was able to handle the management of workflow and effectively move the data between the different systems.
On paper, this solution is practical and straightforward. EAI is focused on managing application messages that are geared to passing discrete transactions that help move information between different applications. These interfaces are usually engineered to move large numbers of transactions in an OLTP environment. Such functionality sounds like an absolute windfall in a vendor presentation, but the devil is in the details.
When considering EAI for operational reporting, it’s important to understand the volume of data that is to be processed. While a query that retrieves a record or two isn’t a big deal, the EAI interfaces aren’t typically set up to manage the migration of thousands or even millions of records in an efficient manner.
And business intelligence queries typically require the analysis of large data sets and can be very process intensive. Which brings us to the other key question: can the operational system or OLTP database support the actual query being submitted? Once again, if it’s a small number of records (e.g., your airline reservation), EAI is a good choice. If it’s a large number of records (every flight I took out of Atlanta last year after 6 p.m.), there could be a problem. It’s not that the EAI solution can’t move large volumes of data—it’s that the architecture was likely engineered to support moving small volumes of data (as in less than 1,000 bytes) instead of large volumes of data (as in 1 gigabyte). Once the data volumes exceed the design, the process will become highly inefficient—or just plain slow.
Is there a middle ground? The key with EAI as a data movement alternative is to focus on the targeted application functionality and the amount of data that is migrating between systems. Small volumes are likely to work effectively. Bulk data migration or large answer sets are likely to stress a system like EAI that was principally designed to support smaller-sized messages.
This excerpt was pulled from the Q1, 2005, TDWI Ten Mistakes to Avoid series booklet, Ten Mistakes to Avoid When Considering Data Warehouse Alternatives by Evan Levy.