Rock-solid tips for building a profitable database
As the sponsor of a new marketing database, you will be faced with both technical and political challenges. A commitment to profitability will help your project survive scrutiny both now and in the future. These tips should help:
First, don't let your project be hijacked. Your database is a big investment. In order to be profitable it needs to be completed on time, and all the costs associated with it need to support marketing goals. If you allow other departments to add requirements that are operational in nature, like support for inventory management or financial reporting, your project will cost a lot more and take a lot more time to build.
Then, spend the time defining clear objectives and measurable goals before soliciting bids. This may mean bringing in a series of potential providers just to get a clear picture of what's available and how the pieces should work together. Every data source that feeds into a database represents both initial cost and ongoing cost. You should have a pretty good idea of why you need any given data before you include it.
Remember that after a certain number of years, either you will outgrow the solution, the underlying mission will change, or technology will advance in a way that makes a complete redesign appropriate. So, plan around the concept of a useful life. There's no single answer to how many years a marketing database solution should hold up, but don't build the most expensive house on the block if you think you'll need to move in three years.
It's also good to consider a shared environment. Many providers offer this option, which is an excellent alternative for businesses that need an entry-level, low-cost, or quick-start alternative to custom database design. Almost no one fits perfectly into a predesigned box in terms of schema and data model, so some level of customization will always be required. If the cost for custom work is less than 30% of the overall cost to build your database, you may be a good candidate for a shared environment.
Finally, fix data problems at their source. Plan on a critical look at the data destined for your database as part of a data discovery project. You will uncover data issues relating to changing practices in order entry and item categorization, as well as poor upfront data edits. If you don't plan for this, you'll be tempted to have your provider massage the data each time he receives a feed. That adds both initial cost and ongoing cost. More importantly, it leaves the problems to be discovered again in a few years' time.
Terry Fitzgerald is director of client development at Creative Automation. Reach him at firstname.lastname@example.org.