Fashion is not my strength.  When it comes to shopping for clothes, especially shoes, I’m a pragmatist.  I tend to buy the shoes I can wear often and really get my money’s worth.  This does not please my fashion focused better-half.  She’s constantly urging me to splurge for the expensive and trendy shoes.  On occasion I break down, spend the big money and head home trying to imagine under what circumstances I will actually wear these fancy shoes.  I realize this is boring and I’m missing some critical DNA element that makes me cool, but I’ll get over it :).

Each time I visit my closet those expensive shoes (still in their box) stare up at me, reminding me how expensive they were and cause me to think about all of the other ways those funds could have been spent.

Recently I read the excellent blog post from my colleague Taylor about keeping your email house file clean and this is what started me thinking about shoes and data integration. 

You see, for organizations, it’s literally impossible for all data to come from a single source, or be stored in a single repository that can support the creative and strategic ideas they have to engage supporters and constituents.  Ergo the need for data integration, in some form or fashion is always necessary.  But “total” integration is not only incredibly difficult, expensive and resource consuming - and almost always unnecessary.  A better approach is to consider replicating your shoe buying habits.  Determine which data sets are most valuable and will best support your priority strategic goals and spend your integration resources on developing the most efficient synchronization of these data sets in a deliberate order.  In other words… be an integration pragmatist, don’t try and achieve 100% replication between systems, think about which pieces of data are most valuable and focus there first.  Start with the most obvious constituent profile fields like name, email, address, etc. and then expand to transactional data, interest values, group membership, etc.  If you prioritize the data that will be used the most, you’ll never have that wretched feeling of buyer’s remorse.

Hey techies, the next time a boss from your organization requests the integration of two data sources, ask the following questions:

• Specifically how are you going to use the combined data?
• Will this integration increase our results?  How?
• To achieve these results, will you need recurring synchronization or would a one-time or infrequent sync suffice?
• Once combined, do you have the tools you need to segment this data or will you be needing my help with that as well?

Understand, I’m not advocating data silos and a closet full of dirty old slippers, quite the contrary, in order to succeed in multi-channel (online, direct mail, etc.) marketing, integration is imperative.  But just as with my shoe shopping, there is a cost and we must always rationalize our investments with sound reasoning based on the results we expect to achieve.