• by Anto Chittilappilly , Op-Ed Contributor, December 11, 2012

When planning the implementation of an attribution solution within your organization, there are numerous decisions that will affect the length of time required to complete that implementation and begin benefiting from the insights and recommendations driven by your solution.  Your organizational priorities, the amount, condition and location of marketing performance data, and the make-up of your overall marketing infrastructure will all affect these high level and detail-level decisions, as well as the integrity and accuracy of the attribution process output. 

Legacy Data vs. Harvesting Data

One of the most significant decisions you’ll need to make is whether you want to start your attribution implementation by gathering legacy marketing performance data from your various data sources and running it through the attribution process in order to derive insights and resulting recommendations based on that historical performance.  The alternative is to turn on your attribution solution and begin harvesting attribution data beginning on Day 1– either by piggybacking on already existing tags on your digital ads and web pages, or through the implementation of a new attribution-specific tag/pixel on your digital assets.

The trade-offs to consider in making this decision are:

• Consolidating/normalizing legacy data and taxonomies from multiple sources takes time and will incur costs, but will enable you to more accurately compare the performance of each channel and campaign in the long run.
• Limiting data collection to the harvesting approach (beginning on Day 1) requires time to build enough history to derive meaningful findings, but will get you up and running with attribution more quickly.
• Piggybacking on already existing tags to harvest attribution data from multiple sources requires application of business rules/methodologies to stitch together the touchpoint history of your prospects across those multiple sources, but will enable you to avoid the application of an additional attribution-specific tag to your digital assets.
• Relying solely on attribution-specific tags requires time/cost to implement those tags on ads (if you want impression data) and website (if you want click data) and possibly to overcome the objections from your IT department or other internal constituencies, but will enable you to avoid the time and cost involved in implementing business rules to stitch together the touchpoint history that’s been recorded via multiple platform-specific tags.

Customized Attribution for Your Organization or a Fit-All-Sizes Solution?

Another extremely significant decision that you’ll confront is the extent to which you want to customize or configure the attribution solution to match your specific business/marketing model, or whether you wish to adopt a more plug-and-play model based on your attribution provider’s “out-of-the-box” offering.  These customizations/configurations include rules to calculate business-specific metrics or custom categorizations of business units, product lines, sales territories, etc., multiple KPIs to reflect your unique marketing and business objectives, different data views for different job functions/partners who need to access the data, and multiple other business-specific needs.

For example, one of my banking clients requires an attribution solution that assigns the credit of the life- time value (LTV) of their customers. This client calculates LTV based on the funding balance of their customers month by month. As the funding balance changes over time, the attribution credits already given also need to be changed. A solid custom attribution solution will go back and re-apply the new credits to those old conversions.

The trade-offs to consider when selecting a custom or out-of-the-box attribution solution include:

• The more customization and configuration that takes place, the more time it will take before your implementation is complete and insights are recognized, and you will incur more cost during that process. But, it will give you a more accurate understanding of your current performance and opportunities for improvement based on the same success metrics you already use.
• Inherent with any customization and configuration will be a more complex data management, integration and QA process that could provide increased exposure to problems as changes to the structure of the data that feeds your attribution process takes place over time. 
• Implementing an out-of-the-box solution will cost less and get you up and running faster, but may force you to adopt the attribution providers taxonomy, metrics, objectives, business rules, ways of looking at and grouping data that may be not meet your needs.

Remember the old adage “Good. Fast. Cheap. Pick any two.”  Consider the trade-offs in the context of your own environment, and decide what’s best for your organization long-term.