Commentary
Audience Must Be The Atomic Unit Of Any New Cross-Media Measurements
- by Dave Morgan , Featured Contributor, July 19, 2012
Cross-platform ad campaigns are coming, and so is the desire for real cross-platform measurement. Unfortunately, no one yet knows what that measurement will look like and which metrics will rule it.
More and more marketers today talk about the need to buy and measure advertising campaigns across multiple media and devices in a coordinated way, especially all electronic media, including Web-delivered ads on computers, tablets and mobile devices or spots on television and radio. It’s a compelling and sensible vision, but making it work in practice will require that all market participants make some tough choices about which specific metrics will drive cross-platform advertising.
At this point, I don’t know which metrics will win out. However, I recently had a conversation on this topic with my colleague Jack Smith, our chief product officer here at Simulmedia and founder of the Media Innovation Group, and former head of product strategy at 24/7 Real Media. Not surprisingly, Jack has some very strong opinions on this topic. Specifically, he believes that audience will become the atomic unit and primary driver of all cross-media measurements, and that the federated (anonymous) identity of a consumer of content will be fundamental to cross-media measurement and monetization. Here are his reasons:
advertisement
advertisement
People are not commodities. There are a finite number of people in the universe. If the audience universe is finite, the value of any "individual audience" can be determined if you can reduplicate this audience across media.
Impressions are commodities. Measuring impressions is easy. They are different silo by silo (e.g. TV, Web, tablet and mobile), and that impression measurement can and must exist somewhat differently in each and every silo.
Reach is the baseline of media value. Unique (or unduplicated) audience reach is the baseline for media value (e.g., who and how many did you reach?).
Essential for cross-media de-duplication. De-duplicating audience across media/devices isn't possible until and unless audience is the central unit of measurement and made consistent across all media.
All other measurements follow. Once the cross-media unique reach problem is solved, additional media-specific measures can be layered on top of reach. For example, social gestures, likes or Twitter mentions are important metrics for social campaigns. Show ratings are critical in TV. Time spent is important in magazines.
What do you think? Will “audience” be the atomic unit of cross-media measurement?
But what will be the Higgs Boson?
Audience is certainly a critical metric because it's a starting point. You must have an audience. However, conversion rate is also consequential since that represents the percentage of audience members who reacted the way an advertiser hoped they would act. And finally, the average transaction value identifies what each conversion is worth. Hence, you need all three metrics to project and optimize revenue. In fact, one can improve revenue by increasing one or all of these metrics.
Bravo, Dave. Publishers and networks are already scrambling to incorporate audience measurement into their inventory management systems, since increasingly TV cum Digital Video buyers are demanding to be billed on actual TRP/GRP instead of raw impressions. This is a huge shift away from the tonnage model that has plagued digital for so long. As I (and others) have said, it's not that brand dollars will come online in a vacuum, rather it's TV - with all its measurement foibles and scarcity based valuations - that will come online itself.
I feel compelled to respond to comments since Dave is quoting me. (So, Dave, you elegantly delegated this one.)
@michael People buy things. They make social gestures such as share content or like a page. They click ads. They watch TV shows. Each of these actions (buy, share, like, click, watch) could be considered a part of a set behaviors that people do. Conversion rate is a calculation of the number of people who perform a single behavior: conversion (assuming for simplicity a sequence of behaviors is unimportant). The context in which this occurs (online, offline, app, TV, wherever) is another set of variables. While you could improve revenue by increasing one or all of the metrics you describe, you would have difficulty doing so predictably over time without first establishing a framework which places people at the center and relates contexts, behaviors etc to people. In this sort of framework, you can more easily understand the effect that the change in any one variable such as cost of traffic, Revenue Per Thousand Pages (RPM), CPM floor, click rate, etc has on the other variables in a problem like yield optimization.
@hobo LOL. No idea. I'm just a media & product guy. Our Chief Scientist is the only Physicist we've got. You'll have to ask him.
Gosh, this is exactly what I was told in the media training program at Benton Bowles in 1981. I just explained this to 20 people in a training session earlier this morning.
Yes, advertisers are interested in the audience that lives on the other side of all the cool stuff the media industry seems to prefer to talk about.
The audience is the only asset that matters. GRPs along with Reach and Frequency models are an elegant way to keep track of the audience.
The sad thing is that our industry as so sorely lost touch with its foundations that Mr. Smith's brilliant observations seems like an epiphany.
I'll be your Huckleberry, Jack Smith from Simulmedia, First, you either can improve revenue or you can't improve revenue by manipulating the three metrics. Since, second, the size of the audience matters, in terms of measurement, insofar as if you could increase the size of it you would increase revenue. if, however, that was not an option you could still increase one or both of the remaining variables and still increase revenue. Third, predicting someone's intention to buy is easier than trying to predict actual behavior because many situational factors cause one not to engage in a specific behavior. Dave's on the right track. The audience is the key, but if atomic unit of measurement means "end all" or "perfect" measurement I would tend to disagree.
Great discussion. May I add another metric in the mix that could be a predictor of revenue? ROI is the ultimate measure of success. For most cases it is difficult to connect the money invested in media campaign with revenue generated as a result. So two out of Michael's 3 metrics are difficult to measure unless the advertiser is only selling online and can track which buyers came from a media campaign. We believe that "engagement time" with the brand is almost as close to purchase intent as searching for a product on Google. De-duplicating or not de-duplicating audience is an interesting conversation but if there is a way to by-pass that and find a media metric that is a better predictor of revenue....like time spent with the brand, I would certainly advocate for that. Social media campaigns can be looking at minutes of engagement compared to seconds for TV. Thoughts?
Good points Michalis about ROI being a core objective, which is another reason audience is the critical atomic unit, since whether the "return" sought is engagement, awareness, interest, advocacy, response, sale, repeat sale, etc., the best way to measure it is on a per audience basis.
Michalis, measuring engagement time will certainly get a marketer closer to learning what a buyer intends to buy. Let me ask you, are you measuring engagement in the ACTION stage of AIDAS because the intent is already there - they have a problem and your product or service solves it? If so, Key word conversion rate is a good way to measure buyer intent here; Or are you targeting people in the earlier INTEREST stage of AIDAS when there unfamiliar with your brand? Those people rarely convert right away, but it's a great place to build your brand.
Dave and Michael thank you for your responses to my comment. Michael to answer your question we are attempting to measure engagement in the earlier stages of AIDA. Our core objective is to generate leads that will lead to sales thus we are working on awareness first, and then try to create interest and desire. We follow the statistic that a prospect needs to be touched by the brand or offer at least 7 times before they buy :). If time spent at "touchpoint" grows each time then it is a good indication that interest and desire are developing which will hopefully lead to action.