Digital Branding: Measuring the Effectiveness of Online Brand Advertising

Posted by staff on November 6th, 2009 at 6:00 pm
Written by Richard Cacciato

In 2008, US measured media was $280 billion, of which $254 billion was offline and $26 billion online.  73% of offline media was brand marketing whereas online, direct response marketing was 76% of the mix.  Part of the skew is attributable to the measurability of direct response, particularly online, and the difficulty of measuring brand marketing.

Connecting online ads to offline purchases is difficult at best.  Comscore published a white paper in 2008 on the dramatic underestimation of display advertising ROI, primarily the result of the fact that any assessment of less tangible brand marketing is qualitative and after the fact.

Online has been held to higher standards than traditional, because everything is measured, and probably because we still need to prove the value of online for branding.  In the past few years we’ve made great strides to prove the efficacy of digital advertising on the consumer funnel.  According to the panelists, we’ve reached a threshold and there are now good technologies that can help us measure.  However, the truth is it’s still difficult to connect the dots between online ads and offline purchases.

The mix on this panel was interesting: one marketer (Andy Markowitz of Kraft), two agencies (Carl Fremont of Digitas and Richard Guest of Tribal DDB) and two technology companies (Dan Beltramo of Vizu and Andy Atherton of brand.net which claim to provide tools to help read the tea leaves).

I’ll confess that I usually put more weight on what a marketer says.  Kraft has been engaged in online for a long time, and they have lots of measurements from soft to hard ROI metrics.  According to Andy Markowitz, the only constant has been this: there’s only tolerance for what sells more product: the old saw of “the proof is in the pudding”.

Online brand campaigns and offline-style extrapolated measurements lead to suspicion and mistrust, as well as fears of inaccurate sampling which make brand marketing an easy budget-cutting target.  Most marketers think about online brand campaigns as a murky immeasurable necessity.  Can we use new technology to close the gap and measure online brand advertising ROI?  Can we directly attribute sales lift or brand lift to the online campaign?

The problem is that the online content environment is very fragmented.  Efficient tools have been developed to run direct response campaigns online but there are no (or few) efficient tools to run branding campaigns online.  There are platforms out there but they are complicated and not well developed.  The challenge is to “operationalize” the efficiency of things like multivariate testing, that is to make the analysis routine and embedded in the process.

The numbers are somewhat misleading since the direct response portion is driven by search and not necessarily display advertising.  There are still a lot of marketers who are not convinced of the efficacy of brand advertising online, and some technologies are being developed.  But the bottom line is, it’s still very qualitative.

Consumers don’t understand the difference between “direct response” and “brand”.  The data suggests that on average the offline sales lift vs. online spending is 140%.  The good news is that with digital, we are starting to have the ability to read the effects in real time, even with the tools we have now.  This means you can see your campaign results in days or weeks instead of months, and adjust accordingly.

The key to doing this right is what panelist Carl Fremont of Digitas calls “closed loop marketing”.  It’s all got to be integrated, the old-fashioned well-structured marketing mix, something I’ve been preaching for more than 10 years…  Maybe I can finally stop feeling like a modern-day Cassandra.

One Response to “Digital Branding: Measuring the Effectiveness of Online Brand Advertising”

  1. sagar says:

    I really enjoyed reading this, I agree and it opened my eyes to things I haven't thought about. thanks