An incremental approach
Posted at 3/18/2008 09:47:00 AM
Here is a real query from hell in the making. One of our customers wanted to prepare a list of customers for a promotion. They wanted the usual selection criteria like geographic and demographic selections and they wanted at least 5,000 good names and not more than 6,000.
With the incremental approach, you ask for one qualification at a time and you can review the results between each step. The marketer who was preparing the promotion was able to select the customers from the area where the promo was going to take place. A quick look showed that there were far more customers there than they wanted to contact so another selection on a demographic code narrowed the result but there were still too many names.
Scanning through some of the results, it was quickly clear that there were many old customers who had not shopped recently. Now the marketer had the option to go after the current customers and focus the promo on retention or go for the old ones to try to get them back. Selecting the old ones resulted in a few more names than were really desired. Another quick scan through the data showed a wide range in average spending. An addition to the query selected the big spenders resulted in the right sized group and the promotion was launched. Total time for the discovery process was less than one hour.
Imagine attempting that with a typical DW! What was the original requirement? It wasn’t really known. If this had to be done as an IT project, it would never start since the requirements definition would take longer then the cycle between promotions.
If the marketer had access and skill in SQL, then the same process could be attempted. However, the probability of having all of the required fields indexed would be near zero. If the exploration and business decisions had taken another direction and they went after active customers, a different set of fields would be required. What are the chances that they would also be indexed?
Without a doubt, there is a query from hell somewhere in this business process.
With the incremental approach, you ask for one qualification at a time and you can review the results between each step. The marketer who was preparing the promotion was able to select the customers from the area where the promo was going to take place. A quick look showed that there were far more customers there than they wanted to contact so another selection on a demographic code narrowed the result but there were still too many names.
Scanning through some of the results, it was quickly clear that there were many old customers who had not shopped recently. Now the marketer had the option to go after the current customers and focus the promo on retention or go for the old ones to try to get them back. Selecting the old ones resulted in a few more names than were really desired. Another quick scan through the data showed a wide range in average spending. An addition to the query selected the big spenders resulted in the right sized group and the promotion was launched. Total time for the discovery process was less than one hour.
Imagine attempting that with a typical DW! What was the original requirement? It wasn’t really known. If this had to be done as an IT project, it would never start since the requirements definition would take longer then the cycle between promotions.
If the marketer had access and skill in SQL, then the same process could be attempted. However, the probability of having all of the required fields indexed would be near zero. If the exploration and business decisions had taken another direction and they went after active customers, a different set of fields would be required. What are the chances that they would also be indexed?
Without a doubt, there is a query from hell somewhere in this business process.
