How do I cut my warehouse costs? I need general tips to reduce my warehouse costs.


Sid Adelman’s Answer: I think the goal of cutting data warehouse costs is the wrong way to look at the data warehouse. It makes more sense to consider the opportunities to increase revenue, improve customer service, deliver higher quality products, negotiate better with suppliers, etc. However if pressed to minimize your costs you might consider the following:

  1. Every project must be cost justified. Measure costs and benefits post- implementation.
  2. Charge the users for everything they do; they will be less likely to waste resources. In fact, they may not use the data warehouse at all which could reduce your costs even more.
  3. Use existing software as much as possible. Buy new software only if you absolutely need it.
  4. Accept longer implementation schedules (this could displease management).
  5. Fully understand user requirements and only supply data that has been identified as required.
  6. Minimize historical data (this will limit your ability to perform trend analysis).
  7. Only clean up the data that absolutely requires it.
  8. Accept low levels of service for availability, performance and support (now I’m gagging).

While some of these ideas are universally appropriate, if you follow all these suggestions, the best you can deliver is a mediocre data warehouse. Are you sure that’s what you want?
Doug Hackney’s Answer:

  1. Leverage low cost, integrated platforms, i.e., SQL Server
  2. Think about maintenance and sustenance now rather than later. This will help you to define and implement the processes around the DW system early, which opens the door to build them in a cost effective manner.
  3. Keep your scope small. Build the smallest possible scope that will still be politically meaningful.
  4. Hire people who have experience in building and sustaining a DW system. The most expensive mistake you can make is one that has been made before.
  5. Look at packaged solutions. If a turnkey solution is available for your operational system, look there first. You will save significant money if a packaged solution will meet your needs without extensive modification. (Be careful of the hidden long-term costs of non- compliant architectures and maintenance of heavily modified systems.)

Clay Rehm’s Answer: Since you did not provide more specifics on what problems you are facing, and since I am not familiar with your data warehouse, the people who use it, the people who support it and the processes to maintain it, all I can provide are generic suggestions.

Tools: Are you using the right tools? Have you considered an extract, transform and load (ETL) tool instead of using procedural languages? Are your query and reporting tools easy to use and easy to create reports with?

Processes: Are there redundant tasks and processes? Can you document the current environment and try to spot where there are redundancies and overlap of processes, functions, procedures, etc.?

People: Does your DW have the appropriate numbers and types of people supporting it? Are they fully educated and trained on the aspects of the areas of expertise?

Data: Are you spending a lot of time cleaning and transforming data, that could have been cleaned up front by having proper edit checks on the OLTP source systems? Does an easy-to-use data dictionary (meta data) exist so users and IT alike can understand the data?

Scott Howard’s Answer: Do you really mean to cut costs or control costs? Cutting costs could itself be an expensive proposition because it most likely involves retooling the environment with lower cost ETL, RDBMS and analysis tools, but if you’ve already purchased your tools and software ... enough said. There could be potential savings on software if there is a cost of ownership consideration beyond your original purchase, assuming that you don’t have a long term maintenance or support contract to consider. If you’re running on an expensive hardware platform, you can rehost to a lower cost environment, but again, what do you do with the now idle equipment that you may already own ... enough said.

The right way to control costs is by implementing an architected DW. A well- designed architecture in which you only introduce models and data required by the current and active data marts will require you to only pay for the resources you currently need. It’s the closest to a "pay as you go" approach as one can get. This leads to a methodical incremental approach in which the warehouse grows to meet your demands. Costs are controlled because you will always need to financially justify DW expansions that require the introduction of new models reusing that which already exists in your enterprise DW or first- tier data mart required by new projects or business process changes.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access