Commentary
Metrics Muddle: The Struggle For An Online GRP
- by Adam Kasper , August 4, 2009
OK, I know the issue I am about to tackle is a hotbed of controversy, but I am going to do it anyway. Guess I'm a glutton for punishment or something. It's just that I can't stand by and watch so much being said in the industry, in the press and in conference rooms everywhere without speaking up on the topic of the online GRP.
Much has been written on the subject from many leaders in our industry but yet it is still an undefined concept that lacks consensus. To date, I do not believe the models that have been proposed and published take the concept far enough or are even well developed enough to merit usage. Among other missing pieces, the models that I have seen proposed lack an element that is critical component to the idea: a currency that is common between media.
To me, the idea of the online GRP exists for one reason only: to make the translation from television to online video easy and seamless. There are television dollars that are just waiting to shift into video but won't until planners, buyers and advertisers are comfortable with what those dollars are going to by them online as it relates to what they were buying on TV.
I should note here that I believe in online reach and frequency as a way to gauge the potential impact of a campaign. If the online GRP is a byproduct of that, I have no problem with it. But instituting an online GRP for the sake of having another way to express impressions is redundant. I agree with both Jeff Ramsey and Young Bean-Song in that a GRP gives you the denominator that puts your numerator into perspective, but I believe that problem is better solved by looking at reach and frequency as an evolved measurement of the audience.
Adding to the complexity of developing an industry-wide metric is the fact that, regardless of which is higher or lower, online video and television affect people differently and thus the value of an impression on each is, by nature, different. And yeah, once we start talking about the differences between in-stream, in-banner, pre, mid post and different content types online it becomes an even tougher problem to solve...but I digress.
For the online GRP to work, it needs to have three components:
1. Clean counting methodology. ComScore currently counts both content and ads as separate streams. This will change in time, as they are currently working on a fix for this, but this is a major issue as of today. We can't make any accurate projections for site activity if the numbers we have to rely on are cloudy and misleading.
2. Accountability for TV and online universe differences -specifically, factoring in TV viewers who do not have Onternet access. As I stated above, having an online GRP that exists in a silo only to the online world is a non-starter for me. The currency needs to apply to both universes or we will only be talking to ourselves. Trust me when I say this: Clients want to know that an online GRP is the same (in size) as a TV GRP.
3. Recognition that online video and TV work differently. To create a common currency between two media without accounting for the fact that the value/impact of an impression on one is different from the other is like placing the same value on a Bentley and a SmartCar because they are both cars.
I look forward to seeing what we can all come up with. And yes, before someone posts something asking where my online GRP proposal is, we do have a solid one at the agency. Maybe I'll even share it with you...
Interesting article, but why are we all still stuck on GRPs as a measure. Impressions are more important. A rating point has never purchased anything, only people do. Rating points are not consistant across media. I have seen cases where regional dealership groups use rating points as a way to figure out equal payments, but in reality, it should have been done by impressions.
I agree that there needs to be a common measurement to compare traditional and new media -- and the second we have that I am hoping everyone will (please!) stop to refer to it as old and new media. (Thanks.)
More to the point, though, Virginia is right that impressions are important for measuring impact and we, as an industry, have been vexed by the weight of a TV vs. a Print 'impression' long before the internet came along and threw us another curve ball. But as long as we do not have a better metric, let's make sure to break down GRP (into reach and frequency, that is) and figure out how to apply to the (not so) new media. As long as we're consistent it's at least a starting point.
Great discussion points.
My problem with an online GRP is that it is just a tool for translating TV's flawed counting methodology to online. Surely the TV buyers can see the value of reaching their target audience through online so why can't they use online language?
The payment models for online video are moving towards engagement metrics which ideally leads to sales - surely it is way more important to move in that direction than to try and fit around a broken model.
Great topic. And there's something to be said for you being a glutton for punishment! I'm looking forward to having you share your thoughts on your internal agency GRP methodology. We've been preaching that we're at the ready to fund research to measure the effectiveness of online video, so if clients need the extra boost to make them comfortable, then let's together work to usher them into the medium.
Hey Adam, great article.
I run Video Metrix and am interested to as you to elaborate on the 'cloudy and misleading' aspect of ads and content. Would love to see if I can help clarify this at all...or work out how we can improve.
thanks
Tania
I agree and believe there more reasons why an online GRP is still a long way away.
First it can only work when the audience universe is defined in the same way ie the cross media campaign is all trying to reach women 25-44. Fine but why throw away the ability to create a target universe that is more granular and a better fit with the product? ie health conscious women with an interest in exercise. While this makes the audience smaller (and far more efficient) it means that when we define the reach component of the GRP equation (reach x frequency) we are not comparing apples to apples. Why leave BT aside as a tool just to have comparable GRPs?
Secondly we have look deeper at the frequency part of the GRP equation. The number used is the *average* frequency and as we know the average person puts mascara only on one eyelid. In other words, average is useless. To understand the effectiveness of frequency we need to look at frequency distribution. Two campaigns could have the exact same average frequency but one has a lot of reach at low and high frequency with none in the just right zone while another has very little in low and high frequency. Since television does not have the luxury of actual campaign data with measured frequency distributions (like Artemis does) there really is no point in trying to compare a television GRP to an online GRP. Online should only focus on reach within desired frequency ranges and discount impressions bought and served in the "too low" and "too high" ranges.