Published: June 2011 | By Yahoo!, comScore
Digital
really is the most measurable medium, and as such, it continues to
produce metrics, metrics, everywhere! This constant influx of new
metrics may sometimes be difficult to navigate through, specifically as
it relates to how they relate to measuring the effectiveness of
marketing campaigns. We already have traditional brand metrics (brand
awareness, brand favorability and purchase intent), that have a clear
role in campaign evaluation. However, there are new digital-specific
behavioral metrics, such as trademark searches, site visits, and page
views whose roles in campaign measurement have yet to be clarified.
Methodology
Yahoo!,
in partnership with comScore, looked at over 100 campaigns across 5
categories to see how changes in behavioral metrics relate to changes in
brand metrics and near-term purchases. All 100 campaigns were measured
using behavioral, brand, and when available, purchase metrics.
For
our analysis, we looked at three of the most common behavioral metrics
(trademark searches, site visits and page views), four of the most
common brand metrics (brand awareness, brand favorability, likelihood to
recommend, and purchase intent), and for categories where short-term
purchase data was available (purchase within 1 week of exposure to the
campaign), we included this data.
The
three behavioral metrics were highly correlated with each other (.80
and higher). Campaigns which drove an increase in search behavior also
drove an increase in website visitation and page views. As such we were
able to group behavioral metrics when analyzing their correlation to
traditional metrics.
With our 3 behaviors combined, we performed
multivariate regressions (R2) to understand how well changes in our
behavioral metrics predicted changes in our brand and purchase metrics.
Key Findings
- Depending on the category, changes in behavioral metrics can be predictive of:
- Behavioral
metrics are predictive of lift in brand metrics within CPG: Health
& Beauty and Computer & Technology categories: (Strong
prediction = a R2 .40 and higher)
- Behavioral metrics are predictive of purchase within Retail and CPG: Food & Grocery categories, but not of brand metrics.
- Behavioral metrics are predictive of brand AND purchase within the Quick Service Restaurant category.
Conclusion
Depending on the category, changes in behavioral metrics can be predictive of different traditional metrics. An understanding of these relationships, and how they vary by category, may allow behavioral metrics to be used as a proxy for brand metrics and purchases.
Download this article in PDF format.









