Developing Your Marketing Database -- A Blueprint for SuccessJust as you wouldn't start construction without a blueprint and an end result in mind, you shouldn't build a marketing database without a blueprint for success firmly in place.
There are four major concerns that must be addressed before you start to talk about sexy stuff like neural net models and permission marketing.
Start with your strategy or blueprint. There is an old saying that if you do not know where you are headed, no road will take you there. The strategic phase of database development is called the business requirements phase. This is where you thoroughly define your requirements before beginning development, just like the architect who will not start building your house without understanding the size of your family and what you want in the house. There are several questions that you need to ask to determine the ultimate usage of the database:
* Do you want to build predictive models? If so, what type of results are you trying to predict -- responsiveness, usage, profitability or the likelihood to buy additional products? What data elements will you need to gather in order to make accurate predictions? You will have difficulty predicting profitability if you have not established a way to calculate each customer's profitability.
* Do you want to use the database to track the results of your direct marketing programs? If so, you need to determine the processes currently in place to track the results. A marketing database will not, by itself, solve the problem of not being able to track the inbound calls or direct mail responses. You also will want to consider how you track multiple campaigns over time. Being able to monitor to whom and how often you communicate with your customers and prospects is very important to building relationships.
* Do you want to automate your customer enrollment process? If you haven't documented your current process and uncovered existing areas of difficulty, the database will not solve this problem. As in any computer project, the automation will expose the weak point in your processes, not solve them.
At this time, you should have addressed staging the database development. This means that you should prioritize the need-to-have vs. the nice-to-have elements of the database. Most experts will tell you that you should demonstrate some success within the first six months and then build on that. Your priority list will help you to establish where you should start.
You must also address the organizational issues that will determine the success of your database project. You will need an executive sponsor who will commit to the database project and see it through several funding cycles. This will be important because it will take time before you begin to realize a return on your investment. Typically, tangible returns will not be evident for six to 18 months.
Once you have identified and secured an executive sponsor, you must identify your users, which should be outlined in your business requirements document. Are they primarily analysts or marketing communications managers? What tools and business processes do they currently use? What questions do they routinely need to answer? What reports do they routinely generate? What type of ad hoc requests for reports do they have to fulfill? What questions do they receive that they cannot answer now? What additional information or analytic capability would allow them to answer those questions?
What experience does your IS staff have in developing and maintaining marketing databases? Do they have time for this project or are they committed to Y2K or other projects? Do you have corporate standards for databases and development tools? If not, don't try to reinvent the wheel. Go outside to a qualified organization, which has experience developing databases. The additional dollars spent will be realized in long-term savings vs. trying to develop it from within and the timeframe will likely be much shorter.
These answers will drive decisions about whether to build or buy and whether you maintain your marketing database inhouse or at a service bureau.
You can begin to address the technology issues. Once you know your requirements and who the users will be, you can decide what tools you will use. Certain types of analytic users will be interested in multi-dimensional tools which allow an analyst to slice and dice the data numerous ways. This also will require certain types of data structures and extraction capabilities. Marketing communications users on the other hand, may not find much use for those tools. They may be more interested in campaign management and reporting tools. Which way you go will lead you to decisions about the technical architecture of the database.
Finally, consider the benefits of integrating your database into your Web site. Until the last year or so, the Internet was primarily a one-way street from a true marketing perspective: You got hits on your site, and if you were lucky enough the user would leave their name and address for a follow-up. With the introduction of newer technologies and integration of marketing databases and Web site with communication tools, you can establish a true one-to-one relationship with your customers. Not only will they be able to update important information about themselves through the Web, but also you will be able to communicate with them and address their specific areas of interest on an individual basis.
Determine what the ultimate outcome that you desire (strategy or blueprint).
* Establish the need-to-have priorities first.
* Make sure that you have addressed your organizational issues.
* Determine whether you can use inhouse resources or you need to get outside help.
* Define how the Internet will fit into the mix.
* The answers to the above will provide the answers for the technology and data questions.
* Finally agree on realistic expectations with associated timeframes and funding.
Martin G. Agius is senior vice president of business development at Marketing Resources of New York Inc., Orchard Park, NY, a full-service direct marketing agency. His e-mail address is firstname.lastname@example.org.