While many in our industry were focusing on of retail media measurement during these past few weeks, Google was busy thinking about how to measure media effectiveness as they deployed a new pay-per-action / cost-per-acquisition advertising experiment. Since we're frequently forced to use measurements of reach and recall as a proxy for ad effectiveness (since that's often all that we have), it would seem like the search giant's experiment could be a more accurate and direct way of calculating ROI for a particular message or campaign, which is really what we need to be looking at, in my opinion. Google has a massive database of information, thousands of advertising partners and clients, and gazillions of dollars in the bank to throw at any problem that they find interesting. But for all of that, the thing I find most compelling about the new technology being tested in Google's Labs is its elegance and simplicity. Instead of billing by impressions (as Yahoo! and others did in the old days), or clickthroughs (as Google and most others do now), they'll instead bill by actions. Participating websites will display contextually relevant ads similar to what Google has used in the past. Only now, these AdSense partners will not receive payment for an impression or clickthrough unless the visitor completes some critical action, whether it be filling out a form, completing a purchase, or something else altogether. In this way, Google's clients - the advertisers - only pay for ads that generate revenue or conversions for them, essentially creating a gigantic affiliate marketing network, like ValueClick's Commission Junction on steroids. (In fact, one writer theorizes that Google CPA spells the demise of ValueClick.) While this technology will likely become commonplace on the Internet fairly soon, I'm also excited by the possibility of eventually seeing a large-scale system like this deployed for various in-store media, like kiosks and digital signs.
Tracking a shopper's motions online is reasonably straightforward (and culturally accepted) these days, but how would such a system work in the real world? A great number of approaches have been tried in the past, with varying rates of success, but up to this point all have had some caveat or another. For example, there have been several experiments in the past where a sign (print or digital) would feature some password or offer number, which, when mentioned at checkout time, would garner you a discount. While effective at establishing a relationship between seeing something in-store and completing a purchase, this technique has several shortcomings. First of all, there's no causality: how do we know if the person saw the sign and thus was compelled to buy the product, or instead, was intending to buy the product from the get-go and only happened to notice the sign as a way of saving money on a purchase they were going to make anyway? In fact, if primed to seek out a particular brand or product, a shopper will be more likely to notice branding and signage for that object, further obfuscating the "true value" of the promotion. Second, the technique relies on an individual not only to notice the advertisement and purchase the product, but also to remember to tell somebody about it at checkout time (in order to realize the value of the prize or discount offer). So, there could be any number of people who are influenced by the ad but aren't counted because they forgot to redeem the offer. Third, by using a reward to stimulate the offer, you've done something to measure reach and recall, but nothing to identify whether it was the presence of the message or the size of the offer that drove the purchase. A multivariate experiment using different sign placements, different messages and different redemption values would go a long way towards answering that question, but also adds a lot of complexity to the project.
Another group of marketing experiments rely on a product or campaign's social value as a tracking marker. For example, advertising agency Crispin Porter + Bogusky developed a novel approach in using eBay to measure the social value of their marketing campaigns for Burger King (with their Subservient Chicken campaign), and now with VW and their "Fast" character. Following a few weeks of non-stop TV coverage, staffers at CP+B placed props related to the campaign (and not necessarily the product being marketed) up for sale on eBay. After a number of successful auctions, they started to correlate both the reach and effectiveness of the campaign with the number of bids received for each object. The major caveat here is that the technique is geared towards a particular audience, and requires a really exciting and innovative ad campaign, so it probably wouldn't work for every application. Also, it's much more effective in the awareness, recognition and recall phases of the buying process, which is why it's well suited to broadcast media (like TV, radio and national print campaigns), but probably less suited towards narrowcast marketing techniques.
Then there are those solutions that rely on some kind of tracking technology to determine store traffic patterns, linger times, and even glance and gaze durations. Some of these techniques are getting sophisticated enough to actually identify an individual's race, gender and age, determine where they are looking, and track their exact motions throughout a venue. This sort of thing is a bit too Orwellian for me, but with the advent of cheap RFID and fast computers, more retailers are starting to question whether or not they should be tracking customer movement to this degree. I'd expect that at some point in the near future, we'll be at the point where a shopper can walk into a store, pick up an RFID-tagged basket or shopping cart, move about in a camera-tracked environment, and check out using a private credit card or loyalty card, giving the retailer an authoritative and complete picture of how every bit of in-store marketing, from floor layout to POP displays to lighting made a difference. I know of a number of experiments going on in the industry right now that employ technology from one company or another working towards this goal, and if and when I get any concrete examples of performance, I'll certainly pass them along. While quickly gaining in popularity, tracking technology is still very much on the cutting-edge, and if I were a retailer, there would be a number of legal and ethical questions that I'd like answered before getting involved.
For now, I'll continue to watch Google and their brute-force approach to advertising management. Those clever guys and girls have demonstrated that they're willing to take risks when determining if technology can be used to improve some part of the online advertising process, and their results in the past have certainly had implications for the offline world. Since in-store media has gained a higher profile and kind of bridges the online and offline spaces - with web-enabled kiosks delivering interactive applications while digital signs pull data over the Internet to deliver timely and relevant offers - I know that keeping an eye on Google gives us an idea of what the future holds for our market, and the whole out-of-home advertising industry as well.