Sample Blending Best Practices – Questions You Should Be Asking

In the old days of sampling, meaning more than five years ago, industry-wide best practice was generally considered to be leveraging a single sample source – the philosophy being that it created consistency.  However, as the industry has consolidated, sample providers have built APIs with each other, and programmatic sampling has become more prevalent.  It has become commonplace for sample providers to utilize non-proprietary sampling methods.  Coupled with the fact that most panels aren’t able to fulfill most studies, along with particularly low incidence rate studies or tracking studies with a non-reuse rule, most sample providers have pushed a blending strategy out of both necessity and convenience. 

If you encounter an online panel company that says they are experts in sample blending, here are a few questions to ask to prove their credentials:

1. How are you blending your sample? What kind of formula or criteria are you using?
2. How do you select the sample sources for your blends?
3. Are you blending in an intentional and controlled manner?
4. How do you ensure that your blending does not lead to any accidental skewing of the data due to panel bias?
5. Do you use a standard blend or do you custom blend providers based on the research goals and project criteria?
6. Are you giving any preferential treatment to any sources in your blends? (i.e., focus on your own panels?)
7. Are you able to keep this blend consistent over time?

At EMI, our tradition has always been blending multiple sample sources and we feel that our method – building sample solutions without any bias – is the ideal way to provide sample to your surveys. We do this with Intelliblend®, our unique methodology of blending sample sources in a controlled and intentional manner, to deliver the best data. For more information or to learn more about Intelliblend®, visit our website at www.emi-rs.com.