The market for extraction, transformation and loading (ETL) tools is booming, but companies are paying too much and taking too long to see return on investment. Alan Raubenheimer, CEO of Embarcadero Africa, analyses why this should be the case and suggests an alternative approach.
Data quality problems cost US businesses more than $600 billion a year, according to the Data Warehouse Institute. Yet ironically, given that we are living in the Information Age, management seems to be blissfully unaware of the problem and its extent.
Certainly, if management is aware, there`s little indication of it. Indeed, there is a laissez faire attitude that is a little bewildering. For instance, the Data Warehouse Institute reports that nearly half of businesses believe their data is of such a high quality that they have no need to initiate a data-cleansing programme.
Most executives would seem to accept that poor quality data is a fact of life, almost a consequence of being in business, and something that can`t and doesn`t have to be addressed. Yet this is just not true. Many business initiatives depend for their success on clean data: business intelligence, customer relationship management (CRM), and e-business, for example. If your data is dirty or contains an unacceptable percentage of error or duplication and redundancy, all subsequent insights and decisions will be flawed.
It seems as if, after five decades of business computing, we still haven`t grasped the absolute truth of the old chestnut, "garbage in, garbage out". The Data Warehouse Institute couldn`t be clearer: poor quality data leads to a multitude of problems:
* Extra time needed to reconcile data;
* Delays in deploying new systems;
* Loss of belief in systems;
* Lost revenue;
* Extra costs;
* Customer dissatisfaction; and
* Compliance problems
As a corollary, companies which focus on data quality report significant benefits:
* Less time spent reconciling data;
* Greater confidence in analytical systems;
* Single version of the truth;
* Increased revenues;
* Reduced costs; and
* Increased customer satisfaction.
The return on investment with regards to data quality is easy to define. Meta Group notes: "Research indicates that incremental data quality improvements (especially with customer data) lead to significant business performance gains in each phase of the customer lifecycle."
The market has gone through a phase of "Satori" (Japanese for a-ha!) over the last 24 months, as successive technology waves have failed to deliver their desired and promised value. For instance, CRM has as a core requirement the ability to capture data about ALL customer interactions and transactions, irrespective of channel. Yet Gartner reports that 90% of companies lack a single, integrated view of their customers.
From the business intelligence perspective, Gartner warns that half of all data warehousing projects will fail by 2005, due to denial regarding data quality.
The need, simply stated, is for companies to be able to extract data from any source, transform it into a standardised format, and then load it into any target for further management, analysis or manipulation. This is ETL. Part of the problem, corporates will tell you, is the cost and complexity of current data quality products, which inhibit industry-wide takeup. Many companies in this country, for instance, have been made aware of the problem, but simply cannot cost-justify the extreme investment needed to apply the technology.
Scan the headlines of the trade papers, and you`ll see that a typical ETL project runs into the millions of rand. Try and get that expense accepted by cost-sensitive boards in today`s economic climate!
The second issue is the skills and deployment issue. ETL in its current iteration requires significant skills and can take months - even years - to deploy. It is, therefore, not easily and rapidly delivering the value corporates sought when they signed off the investment.
Then there`s the thorny issue of annual maintenance, which is typically calculated on a sliding scale, for a product that essentially needs no looking after. And all this in dollar-denominated pricelists!
The solution must be a lower-cost, more easy-to-use and -learn approach to ETL; one that has the features customers require but does not cost millions, installs rapidly and does not have onerous annual maintenance bills. Such an approach would make ETL available to the broad base of corporates, and bring it into the mainstream, improving business overall and advancing business intelligence and customer satisfaction.
Share
Editorial contacts