Dun & Bradstreet is a sponsor of sfBIG, and a couple weeks ago we participated in “Prove It! Attribution and Analytics.” Thank you Cory Treffiletti, James Beveridge and the rest of the panel for some great insights. Here are some of the highlights:
Attribution should be like a religion: My recommendation is to pick at data model and stick with it (similar to how you commit to a religion). As Cory said, at least this way you get “consistent inconsistency.”
James commented on the fact that attribution is a data point that flows through multiple technology platforms. Marketers live in multiple systems, and deciding on an attribution model and seeing it through all platforms is the only way to close the loop. There are some cross-functional and cross-technology implications to this, but suffice it to say, attribution does not live with one group or one technology.
Long-term vs. short-term: The panel consisted of quantitative analysts and an undercurrent was certainly their need for a more long-term approach, especially when compared to what marketers seem to demand. As a marketer, I can say that we are often still obligated to only think short term. What are your monthly market-qualified leads? What quarterly pipeline forecast did you drive? Monthly and quarterly figures don’t allow the ‘number guys’ to look at their statistical models, and demanding answers will only give marketers a snapshot that might not be accurate.
Interestingly, the whole panel agreed that emotions play a role in analytics. While they all live and breathe the numbers, they emphasized that instinct often makes you ask the questions that give you more insight. In fact, Cory said, “Experimentation is the road between art and science.” I love it! Another reminder for marketers to trust their gut.
Data driven vs. data informed: One final take-away for me was about how we use the data. To be blunt, sometimes data is wrong or can be misinterpreted. Or, even worse, because of silos within the marketing organization (an example is how creative often is in its own little bubble) people who could potentially interpret the data are not included in the conversation. Having data inform your decisions rather than drive them could be the best course of action.
I thought the panelists were all great, and I am sure I was not the only one who wanted the conversation to continue longer!