Track Opinion
Download the insights industry's #1 strategic planning tool

Talking Ad Research With Joel Rubinson

Joel Rubinson shares his views on the way advertising research has changed since the rise of digital and what it means for the future of the practice.

Editor’s Intro: Joel Rubinson and I have been friends for a long time.  We’re both Long Islanders, and we meet occasionally for coffee and pastry at a local diner. When we’re not discussing important topics like the Yankees, we’ll talk about market research. Below is part of our recent conversation about the direction advertising research is headed (edited so it sounds we actually speak in complete sentences :-)).

Larry: The traditional linear TV ad research model had two parts.  On the attitude/message side, ads were pretested and then tracked after launch.  MMM helped with media allocation and sales impact.  How relevant is that model today?

Joel: Increasingly, marketers are no longer TV first. When TV is a big part of the plan, pretesting still makes sense because the cost of production and media placement for TV commercials is very high, so pre-testing for risk reduction is needed.  However, let’s put this in proportion. While linear TV ad sales are stable, shares have declined substantially as digital has now passed it by. For digital advertising, the equation changes and pre-testing doesn’t make as much sense.  Most digital is just run and analyzed using attribution approaches or surveys triggered by ad exposure within the first few weeks. If it doesn’t work, it gets killed. Very Darwinian! Campaign tracking is getting replaced by attribution approaches like Marketing Evolution where you get tracking, media optimization, and even trading desk-like activation all at once.  About Marketing Mix, it helps with broad channel level media spending but less so with any kind of in-campaign pivoting.  In an era when media was pre-scheduled that was okay but now, with programmatic and addressable media, marketers can pivot faster and would benefit from quick response measurement approaches.

Larry: A mantra I hear is “deliver the right message to the right person at the right time”.  How close are we to achieving this?  What will it take for it to become reality?  What is the research role?  What are the organizational impediments to achieving it?

Joel: I very much believe in this. Basically, this says that targeting is the most important part of the ROAS (return on ad spend) equation and my white paper called “The Persuadables” has proven this. RM/RP/RT can deliver $30 ROAS (30 dollars of sales per dollar of advertising) and nothing, not even a man riding a horse backwards or a Geico gecko can do that! However, there are impediments.

  • Culture: The old advertising model has a “mad men” belief system where creative rules and it is still in our DNA as an industry. I find spectacular lack of understanding of targeting all the time. The worst is when a marketer creates a conquesting segment and then complains that the media company isn’t delivering lift. Duh! ROAS is never high when you target non-users (see my white paper)!
  • Legislative: The new GDPR, in force since May 25, 2018, has created havoc for marketers and publishers. It offers greater protection and control to consumers but at a price of them seeing less customized offers because in some cases, their data are no longer directing relevance. Personally, I love seeing ads for the Yankees and don’t want to see ads for boats or feminine protection products, so I hope the ecosystem sorts this out soon.

Larry: What will “ad research” look like in five years?

Joel: I would not want to be in the copy testing or marketing mix modeling businesses as they existed in the 1990s and early 2000s. MMM providers are embracing MTA (multi-touch attribution) and creating integrated models.  Copy testing is a little more interesting when it has biometrics included, but it is an area where marketers will manage expenses down. For me, in a campaign, near real-time feedback is a more interesting area… more math, more actionable.  Also, it can be highly automated, so the cost structure will be attractive, even with TV as there are a few providers who automatically generate report cards on all new creative.  That is more interesting than one-off $20-30,000 tests.

If you really want to drive ROAS, start testing the segment.  The more valid a targetable segment is (e.g. are those tagged as auto intenders really in the market for a new car?) the more responsiveness you will find to advertising.  I have partnered with some others to offer this segment validity testing service called “Jolt!”

Larry:  Speaking of Jolt, I need some more coffee.  How about splitting another apple turnover with me?

Please share...

One response to “Talking Ad Research With Joel Rubinson

Join the conversation