Advertising That Works. Sorta.

Big Data, Analytics and Programmatic Buying Are No Panacea

Harvard Business Review recently devoted much of their March 2013 print issue, including the cover, to digital advertising and the way big data and analytics are transforming the field. You can read an intro to the piece here, but will have to subscribe to get the full article.

The Promise of Smart Advertising

The article discusses how today’s data-centric marketing is not only able to precisely target ad placements based on demographic, psychographic, behavioral, and a host of other data, but can then also reliably track and attribute subsequent consumer behavior to exactly the advertising that triggered it. This insight is then used to further tune campaigns, optimize media spend, targeting, and placement on an ongoing basis.

Usually, all of this is automated. Media is bought via programmatic platforms and campaigns are optimized in real time based on results achieved, driven by powerful algorithms and without human intervention. According to the HBR article – and countless ad-tech vendors’ marketing claims – this approach promises the equivalent of laser-guided precision ad units that hit the mark each time, at precisely the right place and time for maximum effect and minimal cost. HBR describes this new advertising landscape similar to the precision-guided munitions employed by our military. Perhaps an apt comparison, as they seem to share a common flaw: both require a change in approach and neither seems to work quite as promised in less than ideal settings. Allow me to elaborate.

Measuring is no Substitute for Managing

Digital advertising has always claimed to be more measurable than its traditional, offline brethren. While billboards, print, and broadcast media were forced to rely on crude (and often manipulated) metrics like circulation, Nielsen ratings, GRPs, traffic counts and the like, digital media could always point to its inherently interactive nature. And this claim could be substantiated by “hard” data, including click-through rates and paths, engagement, time-on-site and more.

Proper Ad Targeting, Context, and Placement are Crucial

Proper Ad Targeting, Context, and Placement are Crucial
(Image source: imgur.com)

Even better, when a product was purchased online or a lead generated, this action would commonly be credited to the media that drove the consumer to the conversion point, a process commonly referred to as last-click attribution. It seemed the only blind spot for digital media was the inability to clearly attribute offline purchase decisions to media consumed (and ads viewed) online, where both traditional and digital media appeared to be on equal footing. And, given that consumers were increasingly consuming media online and in digital formats, yet brands were still disproportionately allocating spend to traditional, all that remained was for marketers to shift spend so that online media got its fair share and all would be right. Wrong.

For one, it turns out that not everyone peddling eyeballs online is totally honest and a lot of those clicks and views are not by actual consumers but, instead, automated bots that generate phantom traffic. As a result, much of what brands spend on digital ads is wasted on non-existent views and fraudulent clicks. And it’s not just amateurs getting taken here. Even large, presumably sophisticated advertisers like AT&T, BMW, and McDonalds, running what I assume to be sophisticated (and expensive) campaigns, are falling victim to this and, at times, losing up to 80% of their spend to fraud according to one report.

But, click-fraud and bot-traffic aside, there remain other issues with today’s state of digital advertising and it is still far from the perfect, well-oiled machine portrayed in the HBR. For one, online display ads are notoriously weak in building brand awareness and show weakness even for direct-response advertising. A recent study illustrates this and data shows that consumers are statistically more likely to survive a plane crash than click on a banner ad. Of course, campaigns vary greatly in terms of effectiveness, but you get the gist. Beyond display advertising (banners and such), similar issues surface in paid search, social media, and the latest trend to emerge, so-called “native” advertising.

My Reality Belies Market Claims

At the heart of these issues seem to lay poor segmentation and targeting, along with the crude, heavy-handed execution of many of the campaigns. For all the much-vaunted data that Google and Facebook have on consumer behavior and the supposed sophistication of today’s marketers, I continue to be perplexed by the ads I am (mis-)served on a daily basis. Any parent who has had their teenage daughter borrow their computer for 30 minutes and then paid for this with countless ads for One Direction’s new tour or the spring sale at Forever 21 knows what I’m talking about. These re-targeted ads – which occur when you visit a site but then fail to buy (or even when you DO buy, ‘cause, y’know, you might have missed something, right?) – are becoming increasingly popular with marketers, despite a growing backlash by Internet users.

Marketing Still Requires Marketers

This is not to say that digital advertising isn’t the way forward. Nor that smart, targeted ads guided by data-intensive insight don’t hold great promise, for both are clearly true.

It’s just that, for now, they are not yet delivering on their promise of identifying the right consumers and serving them the most relevant ads at the most appropriate time. Big data, analytics and ad-technology will clearly play a key role in getting us to there. But, technology alone will not do the trick. While it is clearly necessary, it is simply not sufficient on its own. As programmatic buying and RTB (Real-Time Bidding) platforms gain traction and big data becomes an even a bigger deal in guiding marketing (hard to imagine in the current hype-cycle, right?), advertisers, their agencies, and consultants will need to ensure that the human component is given equal consideration. Business processes, marketing departments, campaign design, and media spend all need to be adapted to the new reality. Some even claim that the rise of technology-centric marketing calls for an entirely new member of the C-suite, the Chief Digital Officer or CDO. While this is a debate for another time, this much is clear: investments in hiring, training, and consulting services must accompany those made in systems and technology if they are to be effective. Technology, no matter how refined and sophisticated, is never the panacea promised by vendors. We saw this with ERP systems, EAI/BPM, and a host of other initiatives that all required several years and multiple investment cycles in both IT and human capital, along with structural changes to organizations and processes, to deliver on their promise.

Solid marketing fundamentals, effective strategy, strong creative, and execution still matter, and I fear there’s just no app for that. New techniques and tools will require marketing teams and brands to change their thinking, acquire new skills, and develop effective approaches. Same as it ever was, one might quote one of my favorite 80s bands.

6 comments to Advertising That Works. Sorta.

  • Jeff

    Good insights, Dan. I love the potential of re-targeting but agree we need even more data to make it truly effective (but perhaps even more creepy). Not sure if consumers will react positively when the ads are spot-on with what’s happening in their life – reminds me of Target’s profiling efforts a few years back when they knew a woman was pregnant before her family did. We’ll see the next leap forward with integration of ipTV data into the online funnel – could be interesting when your daughter starts seeing ads based on what you watch…

  • I say “www” stands for Wild Wild West because even though there’s a sheriff, the cowboys do most of the policing of the outlaws. It’s an imperfect system but it functions. If you’re a cowboy, the only thing you can do is to stay away from outlaws, but they’re part of the environment and sometimes they can find you.

    Anyone that has gone deep into any website’s analytics can tell you, there is a wide varying degree of values of clicks, visit, conversions, or any other metric for that matter. That’s why the value of a “user” has decreased significantly in my mind. How many email addresses do you use? How many people use your primary IP address? What about the browser on your computer? The problem is that 1 small incorrect assumption made by the algorithm is multiplied a million fold. Every platform is analyzing the same data and creating outputs based on their personal assumptions they use in their algorithm. They then put the altered (and a little more incorrect) data back into the system, allowing it to churn into perpetuity. For example, I decided to attend my MBA school 5 years ago and probably clicked “apply” somewhere back then. Well 5 tears later I’m still getting re-targeted for the school, whom I’ve already given thousands of dollars to! They’ve probably wasted thousands of impressions on me after I enrolled.

    The only way to really know what your data is telling you is to think behind the metrics. Think about who is doing the clicking and why they are doing it. People use the Internet as a tool to find information, simply give them what they are looking for. Everything else is just noise, inefficiencies, and spam.

    • @Chris,
      Excellent comment and insight. My heroes have always been cowboys (http://www.youtube.com/watch?v=31aETl1BESU), so I like how you tied my childhood dreams into your remarks. 😉

      You’re also spot on that even the smallest error in initial assumptions, recycled and amplified by each algorithmic iteration, gets compounded exponentially, leading to campaigns that drift further and further afield without human intervention. It’s not unlike the trajectory of a projectile flying several miles, where even a slight deviation from the ideal flight path near the point of launch gets amplified along the flight path and will mean that it will be hundreds of yards off-target at impact.

      Moreover, while automated programs are great at recognizing patterns (often incorrectly) and shifting spend accordingly, they lack intuition, context and instinct (what we commonly refer to as “common sense”). This is why it’s crucial to have experienced marketers review campaign performance at regular, frequent intervals and adapt KPIs that go beyond impressions and click-through and include downstream measurements like attributable conversions, lift in sales, etc. Unfortunately, many advertisers and media agencies trust programmatic buying too much and treat them as “fire and forget” systems that self-adjust.

      As to you still being targeted by your B-School, perhaps you should give them (or their agency) a quick primer on this topic and show them the error of their ways. I’m sure they would knock a couple of grand off the price of your second MBA in return. 😉

  • Dan,

    Great article and I completely agree with many of the points made. I will add that it is my opinion that the industry has missed the mark as badly as they have by aiming for the wrong target most of the time. The majority of the industry still measures by impression and click data. Agencies should focus on the end business goals as the key KPI’s in a campaign. Conversion centric campaigns that optimize towards the end conversion deliver bottom line performance that not only drive true ROI but also deliver valuable business insights. Until the industry changes it’s focus, that change will be slow to come.

    Shameless plug, my agency, Sq1 (http://www.sq1agency.com) uses a Conversion Optimization methodology to help all of our clients overcome the problems mentioned in this article.

    Great job.

    • Gabe,

      Thanks for the flowers… and the valuable insight. And worry not about plugging Sq1. You’re a marketer, right? It’s what we do.

      I agree that all parties – brands, agencies, technology providers, etc. – need to look beyond their immediate role, identify where the true business value lies, and define BUSINESS goals accordingly. Which is not to say that there shouldn’t be relevant and “hard” KPIs along the way (there should). But, as Chris points out below, site traffic, impressions, and clicks alone are meaningless. Nothing happens until something happens, and marketers often seem to forget this. I have worked at agencies and with clients that seemed so in love with the practice of marketing, they seemed to forget its role: SELL STUFF! I have seen websites that were beautiful to behold, cost millions of dollars, and were technical marvels. But, alas, they had no drive-to (either somebody forgot or there was just no budget left after all that video and Flash development) and thus, despite sweeping every award on offer, they were never seen by more than a few thousand folks, most of them members of the AMA, the IADAS or The One Club. Other projects would lack conversion points, a call-to-action, lead generation or CRM integration and so would fail to deliver anything more than tangible than brand exposure.

      Lately, there is this obsession with data. But data in and of itself is meaningless without context, analysis and insight. Yes, the digital channel offers interaction and the ability to measure reach, engagement and much more, but this must be put into context, and must tie back to clear business and brand objectives.

      I agree the industry still has a way to go to deliver on their promise. Keep fighting the good fight!

Leave a Reply

You can use these HTML tags

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>