When working with your database, you may have valuable information that is sparse (i.e. less than half of your records have this element). The first thing to do is decide whether the sparse demographic is truly useful or just an obselete one that is taking up space. If the later purge it and spend your time on better things. If it is truly useful, then you need to decide is it sparse because while it is is an excellent indicator, it simply is sparse. It could alos be a new element that has not yet had sufficient time to be gathered, or third there could be a flaw in your data gathering that this demographic is not gathered in all places it should. You need to understand the reason and develop an appropriate plan.

Just because a demographic is sparse doesn’t mean it can’t be used to select by. A list is usually composed of several different groupings of data elements (a segment). A sparese demographic should never be used in a single segment data pull Instead, it should always be it’s ow separate segment along with other segments that don’t have the demographic. In this way, you can use the specificity of the sparse demographic yet not have it cause your list to overly limited.

A first reaction is to keep all data in your active systems. However, even with today’s powerful processors and databases this can lead to sever performance problems. Even if you have a small customer database that is much less than 100,000 records, your order transactions will most likely run into the millions. Further, those transactions may represent products that are no longer and customers who have not bought in many years.

eInformation Strategies recommends that you keep 5 years of active data and summary records for transactions older than 5 years. If the customer has not purchased or had other activity within 5 years we recommend moving their records from the activate database to an archival file that is not queried by your main system.

It should be noted that the 5 years could be different based on your business cycle. For many businesses in the education market this timeframe is appropriate, but for example if your involved in an adoption cycle of 3-5 years, then clearly 5 years is to short, capturing only 1 business cyle. In that case, it would be prudent to go back 10 and probably 15 years.

Knowing the impact of your campaigns is absolutely critical to ensuring cost-effective marketing and ensure you are maximizing your revenue. Classically campaign analysis uses promotion codes, special URL’s and in days gone past bingo cards to measure how effective a campaign was. While these all have their place, relying on these techniques alone in these days of multiple channels of distribution and sales can result in significant errors in your analysis.

We recommend you expand your analysis to do include a historical trend analysis. Use the historical sales figures of a product or the customers to be targeted. Then as sales come in, an analysis can be done to compare what historicaly the sales would have been compared to what they actually were. This will give you a channel independent look at the success of the campaign.

The promotion codes, specific URL’s and similar channel identifiers are then useful to help you understand how your customers respond to a campaign

It’s far easier to have data going into your marketing database be clean and correct than it is to try and clean it up after the fact. Besides limiting the use of free-form fields, it useful to use web services and database lookups to complete as many fields as possible. For example, to get correct school information ask your users to enter a zip code, then provide them with a list of schools in that zip code. Once they choose a school you can fill in the address fields for them, ensuring the data is correct and at the same time linking the record to an NCES or Market Identifier such as a MDR pin.

Depending on the construction of your database, this can be done by a lookup to a database you have on site, either your customer database or a vendor leased database or through a web service to a database at another location. This same technique can be used to capture customer information reducing the number of duplicate records for individual customers.

This strategy works for both customer entry on the web and for telemarkeing, saving the data entry person numerous keystrokes and greatly reducing the error rate.

Markets change and your data needs evolve with them. A demographic you gathered along time ago may no longer be relevant or gathered, or possibly the values that are current are different than the values you were gathering a couple of years ago. But the ability to target your customers can be negatively impacted if you do not pay attention to evolving demographics.

About once a year it is good to profile your demographics looking at what values are in your database for each demographic and the counts of those values. First, it may point out values to you that are no longer needed. Second, as you review your values you may decide a value or demographic has outlived it’s usefullness.

At that point you have three choices. First, to map those values to a new value, this is not always possible, but is often enough that it is worth pursuing. Another option is to decide to discontinue gathering the demographic, but because of it’s density decide to keep the demographic or value for at least a year. The third option is to purge the demographic or value from the active database.

The second option to reevaluate in a year should be a conscious decision, not a default. Having obselete data leads to bad decisions since if you continue to select by it,the number of records selected will be reduced, also if you continue to select by it, even though you deem it obselete it is giving you the false impression of being very selective in your targets. Stay focused. Clear out the garbage and focus on the good stuff!

For direct mail, the costs of not doing address hygience are readily apparent, every piece mailed to a bad address is a wasted piece and postage. Bad Addresses can be people who have moved, so you need to have your database run through NCOA to ensure you are getting to your customer’s where they now live. Another service that is useful is the USPS’ ACS which will provide you back whether the mail piece is deliverable or with a corrected address. ACS will provide the information in an electronic file complete with your customer number so that it is simple matter to apply it to your database (much much easier than having someone use the yellow labels on a returned mail piece!). All of these will reduce your mailing costs. In addition, when you have applied this information to your database, it allows you to get more accurate accounts of the number of pieces you can mail from database. Instead of deciding how many pieces you are going to mail and then scrambling to try to get to that number from your database, you can set up your target, and then have that number printed

An additional benefit is that getting your addresses CASS standardized will allow for better matching, thus reducing the number of customers who have multiple accounts.

When designing web forms, sometimes the best approach seems to be just provide empty fields for the customer or prospect to enter information. The problem with this is the data becomes very difficult to select buy since users can find so many different ways to enter data. To gather useful and selectable data, it pays to take the time to provide a selectable list of options to the user and to prefill in as much data as possible for the user.

This will result in the cleanest and most usable information possible. Even fields like phone numbers should be structured so that user is forced into an appropriate format and can only enter numbers (we have seen phone number fields that were used as note fields). If you are worried about leaving an important option or value off, you can always use “other”, but it then behooves you to review entries there and classify them according to your needs or the data will be unusable.