Online Advertising

Dated Apr 10, 2019; last modified on Mon, 05 Sep 2022

Targeting Ads Based on Mood

NYT’s nytDEMO launched perspective targeting, an ad product that allows advertisers to place ads based on the sentiment that the content is predicted to evoke . ESPN, USA Today, and Oracle’s Grapeshot have similar products too.

How successful is this approach? Google, the largest online advertiser, does not mention it in their ads targeting help page , but Google is not in the publishing business either. Did not find much success for “mood-based advertising” on Google Scholar.

Available emotions include positive ones (optimistic, inspired, self-confident, amused, adventurous, in the mood to spend, love, interest, hope, happiness, nostalgic, indulgent, competitive, informed) and negative ones too (sadness, boredom, fear, hate). On whether the newsroom will be prodded for more emotional stories, NYT claims that it already produces many emotional stories. Furthermore, the ads are less privacy-intrusive.

Less privacy-invasive, true, but maybe more predatory? Tracking is bad, but at least the user can recognize that the ad is following them around and up their guard.

The major challenge is proving that mood-based targeting works and at scale. In areas like music, mood-based targeting’s effectiveness is more plausible.

Is that I’m more likely to play sad songs when I’m sad, rather than the sad songs making me sad?

Online Advertising is Ineffective

eBay paid $20m to MSN for sponsored links for people that searched for ‘eBay’. Consultants viewed it as a $245.6m return. Asked whether the ad clickers were going to use eBay anyway, the consultants invoked fancy math like Lagrange multipliers. When Prof. Tadelis halted Google ads on ‘eBay’, ‘shoes’, ‘shirts’ and ‘glassware’, the effect was insignificant. However, it’s not in the best interest of eBay’s marketers to know whether its campaigns are profitable.

P&G, Chase and Uber observed similar results: cutting digital ad spending did not affect business outcomes. The vast majority of impressions and clicks come from bot activity.

Warrants more investigation. If ad provider knows that the traffic is a bot, then are they not scummy because they know that the bot won’t buy anything?

Attribution to an ad is easy to measure. The question of incrementality, how many of those “conversions” would have happened anyway, is harder to answer. Google’s Ad division doesn’t seem to have convenient tools for determining incrementality.

References

  1. The New York Times Sells Premium Ads Based on How an Article Makes You Feel. Rick Edmonds. www.poynter.org . Apr 10, 2019.
  2. The New Dot Com Bubble is Here: It’s Called Online Advertising. Jesse Frederick; Maurits Martijn. thecorrespondent.com . Nov 6, 2019.
  3. The New York Times Advertising & Marketing Solutions Group Introduces 'nytDEMO': A Cross-Functional Team Focused on Bringing Insights and Data Solutions to Brands. Danielle Rhoades Ha; Angela He. investors.nytco.com . www.businesswire.com . Feb 15, 2018. Accessed Aug 30, 2021.
  4. Project Feels: How USA Today, ESPN and The New York Times are targeting ads to mood. Lucia Moses. digiday.com . Sep 19, 2018. Accessed Aug 30, 2021.
  5. Want to improve your measurement? Get a grip on incrementality. Avinash Kaushick. www.thinkwithgoogle.com . Jan 1, 2021. Accessed Oct 14, 2021.
  6. When Big Brands Stopped Spending On Digital Ads, Nothing Happened. Why? Augustine Fou. www.forbes.com . Jan 2, 2021. Accessed Oct 14, 2021.