The Tom Sawyer Effect of Ad Widgets

futurelab default header

by: Ilya Vedrashko

Wired ran a story yesterday by Dave Demerjian on the future of ad technologies. I've got a chance to share a few thoughts for the article (ad robots and such), but naturally not everything we talked about found its way into the print. Here are a couple of questions that have come up that I thought I'd share. (image credit)

On "widgetization":
I think that the single most important effect of "widgetization" of advertising is that it has become much easier for small publishers (and by publishers I mean everyone who "publishes" something, be it videos on YouTube, blogs and other such things) — to participate in the advertising economy.

Some do it for fun — hey, check out this cool ad; others do it for profit. Amazon with its affiliate system is a great example of how to put an army of small publishers to profitable — for everyone involved — use. Instead of publishing and sending out expensive printed catalogs, they let others to pick whatever products they like and do the advertising job for Amazon for a slice of revenue. Publishers get paid and Amazon knows exactly what it is paying for, down to each individual unit. This is what I call the Tom Sawyer effect — remember how Aunt Polly asked him to whitewash the fence but instead he made it look attractive enough for other boys to line up to do the work for him?

[Jeff Chausse wrote more about Amazon on Hill Holliday blog.]

Do you agree that technology has fostered a new kind of acountability in advertising:

It is inaccurate to argue that advertisers were somehow not accountable for the past 100 years and only have become so with the advent of new technology. What happened, I believe, is that technology has brought about new thinking and raised expectations about how advertising can and should be measured.

Historically, there has been a gap between the moment an ad is seen and the moment the corresponding product is bought. Because of that, we have largely relied on indirect means to measure and predict the extent of causation and advertising success — for example, brand recognition change after an ad campaign would be a proxy indicator of future sales. It worked well for established products that had some marketing track record where future results could be inferred from past performance, and not so well for new products with no such record to rely on.

Television and mail-order catalogs — direct-response ads — have narrowed this time gap considerably, and Internet brought the two moments closer together still, which gave us a way to measure (internet) advertising success directly through sales, while often disregarding the effect of other media. Here we have two problems:

1. Internet advertising raised expectations of measurability for other media that may be equally effective but for technical reasons are impossible or very expensive to measure with the same precision.

2. Internet advertising sometimes is given credit for transactions that seem to occur as a result of a customer seeing and clicking on an internet ad, but in effect may be rooted in "brand recognition" that was due to something the same customer had seen on TV earlier.

Original Post: