About the author : Louis Thibault holds a Ph.D in cognitive neuroscience in the field of attention and conscious visual processing. He is the co-founder of the Sentimens Research Group and heads the business-intelligence research branch, where he where he develops systems for automating human-centered processes.
Marketing will forever be frustrating because of the rift between the conceptual simplicity of a problem (e.g. getting 600 leads with >90% conversion rate) and the tentacular nightmare of implementing a working solution. More generally: everything in marketing is “easier said than done.”
Consider web marketing, for instance. The practice of embedding adverts in the margins of a web page has steadily decreased over the past 10 years (at least!) While this practice still enjoys a few niches, usually on the less hospitable fringes of the internet where malware is a prominent feature, most respectable sites have gradually moved onto forms of advertising that are less-intrusive, more palatable and – we argue – more valuable to the end-user.
A recent blog post by the developer and entrepreneur Troy Hunt entitled Ad Blockers are Part of the Problem provides an excellent case study for introducing the notion of end-user value in modern web advertising, as well as a glimpse of the forces at play in data-driven marketing. Generalizing from this case, we argue two things:
- The “sticker marketing” approach to advertising (defined below) is an unreliable approach to generating revenue.
- The problem Mr. Hunt is facing is best understood as an economics problem with implications for marketing, namely: native advertising and strategic automation
Naturally, we would like to disclose that native advertising and strategic automation constitute our line of work, but the irony of doing so is that it constitutes an example of producing end-user value – or so we hope. We hope that our opinions – which we try very hard to derive from our domain expertise – are useful enough that our readers accept that we introduce ourselves. Our move is to offer some advice in exchange for permission to explain who we are and what we do. With that, we invite the reader to interpret this as an opinion-piece and to consider sharing the article with his/her peers.
How does this relate to Mr. Hunt’s article? A few months ago, Mr. Hunt decided that removing adverts from his website would be a net improvement in UX. Importantly, he deliberately reasoned that he could remove advertisements from his site without interrupting the revenue stream it generates. As such, he made a very reasonable move: replace graphical advertisements with textual accreditation to sponsors and donors. The public, Mr. Hunt thought, would be sure to applaud the decision and investigate the sponsors.
However, Mr. Hunt faced an unexpected turn of events. Someone, somewhere, flagged the HTML element containing the accreditation text as an advert on EasyList, and as a result, this text is absent from all browsers running an ad-blocker. Uh-oh.
Let’s take a moment to appreciate the paradox: trading objectively more obtrusive ads for less obtrusive text had the net effect of perturbing ad-generated revenue. This is contrary to everything we know about contemporary advertising! Users want to be annoyed?! Well, no, but that’s not (yet) the point. The point is rather that Mr. Hunt engaged in a completely reasonable, principled approach to web advertisement with respect to UX: remove annoyances to better engage with users. We contend that removing annoyances is necessary but not sufficient in modern web-advertisement. A marketing strategy need also add value. Mr Hunt’s web-marketing problem is simply that the user is not rewarded for his reading of the sponsor’s message. Worse, this reading incurs an attentional cost which makes viewing the site more effortful, and therefore less rewarding. In this case the user can employ two strategies to mitigate his inconvenience:
- Orient his attention in order to filter-out the irrelevant information. (N.B.: cognitive neuroscience tells us that the deployment of endogenous attention is voluntary and effortful in such cases)
- Right click on the element, select Block element.
Now the economics problem reveals itself: the user is incentivized to block the ad simply because it’s irrelevant.
A corollary: embedding elements on a page like they’re stickers is counter-productive. We refer to the practice of embedding rewardless semantics into a medium “sticker marketing”. It is a pejorative term 🙂 Simply slapping a sticker onto a product is not a viable marketing strategy! (Though, thankfully, Mr. Hunt’s strategy does not hinge on this piece of text being viewed by masses of readers).
Still, this reveals a piece of crucial insight. The formal web-marketing problem can now be refined to ask “how can we replace intrusive ads with non-intrusive sponsorship messages with the constraint that these messages provide value to end-users?” One good approach for Mr. Hunt is also a very popular one: he can syndicate content on behalf of his sponsors and label the individual publications as “sponsored content”. This fundamentally changes the dynamic. Since the readers of troyhunt.com presumably trust the publisher enough to consider his opinions, they can trust the syndication process by extension. That is, they consider that Mr. Hunt will do a good job of ensuring that the interests of the sponsors and themselves align, and that the promotional content is informative to them. Users exchange their trust for rewarding (i.e. relevant) content.
This principle is central to native advertising methodology, and an interesting argument follows:
- EasyList’s blocking of troyhunt.com is user-feedback. It supports the notion of two groups of users: those who are annoyed by the promotional HTML element, and those who don’t care. Those who are annoyed flag the element as spam. Those who don’t care don’t bother to whitelist troyhunt.com.
- In the ad-blocking game, the costs are asymmetrical. There will always be more users ready to circumvent non-rewarding content than resources to defeat them.
- System failures should be treated as market data and leveraged for future advantage rather than actively opposed.
Point 3 is particularly interesting as it allows one to consider EasyList as quality-assurance test. To illustrate, consider a website that produces content, which is quickly flagged as intrusive advertising. Rather than seeing EasyList as an adversary, one might first consider it to be an objective standard of performance: is the sponsored content rewarding enough to pass the “EasyList test”?
But can we go further? We can consider the passing and failure of this test as a measure, or even a cost we would like to minimize. AB testing is a good approach to minimizing a cost function (show two versions of the site: the version that produces the best result wins), but it requires careful coordination with existing development and operational processes in order to be effective. Likewise, advertisers can reverse the dynamic – bringing interested users to them – with strategic automation tools. This is the approach being taken by one of our Native Pricing projects. The principle of operation is that users are incentivized to provide reliable markers of purchasing-intent in exchange of sales on relevant items. This is achieved through gamification of said measures.
It may be tempting to conclude that serving advertisements on commercial websites is no longer a viable business model. In our experience, however, this is not the case. Rather, the market demand has shifted from “adverts that are not annoying” to “adverts that have added-value”. The problem of how to implement this kind of content-curation or production at scale is best left to a separate article in which we will outline the business cases of strategic automation.
Undress.us is looking for partners and sponsors. Please contact us if you are interested: firstname.lastname@example.org