Skip to Content Skip to Footer

Why the Common Wisdom About Native Advertising Is Wrong

Pengyuan Wang, Guiyang Xiong and Jian Yang

In the early days of the Internet, banner ads and pop-ups received eye-popping click-through rates due to their novelty. As Internet users became savvier, however, they avoided ads and installed ad blockers. Click-through rates (CTRs) on these ads fell precipitously and now hover around 0.05% on average. Companies clearly needed another way to connect with consumers.

Enter native advertising, a new twist on an old strategy. Advertorials, infomercials, branded media, and product placements are just a few examples of how advertising and other forms of content have been blurred in the past. With digital media, native ads (also called sponsored content or streaming advertising) inserted into a web stream are designed to look like their surrounding non-sponsored content. This strategy includes advertising consumability as users click, read, and share this content more readily than other forms of advertising, although it can create confusion and annoyance when users realize they are reading ads. In 2019, spending on native advertising is slated to reach $41.1 billion, or 61% of all display advertising.

Advertisement

A new study in the Journal of Marketing is among the first to examine how native ad effectiveness changes across serial positions. Our research team analyzed 120 distinct native ads randomly selected from a leading web portal headquartered in the U.S., encompassing about 180 million page views. To avoid selection bias caused by targeting, we focused only on non-targeted ads. For each ad, we also created separate “natural experiment” studies to compare its performance as its serial position changed. Subsequently, we conducted meta-analyses to generate the results across all separate studies.

Native ads are inserted in between organic articles of the web stream. For example, in our dataset, the topmost (rank #1) ad is placed in the third position of the web stream after two organic posts. After that, there are four organic articles in between every two neighboring ads.

Key findings include:

  • Native ad serial position has vastly asymmetric effects on publishers’ metrics (click-based) versus advertisers’ metrics (conversion-based).
  • As serial position lowers (i.e., from rank #1 to a lower rank), there are only modest changes in publishers’ metrics, but drastic reductions in advertisers’. This pattern is unique to native ads and has not been indicated by prior research on ad serial position, which focused primarily on search ads.
  • If rank #1 is the baseline and given a score of 100%, the conversion ratio (CVR) of rank #2 is only 15.9% of rank #1, whereas its CTR is 97.0% of rank #1.
  • Cross-rank change in native ad effectiveness varies across viewer groups when considered for such variables as gender and age. The CVR change from rank #1 to rank #2 is larger for females (85.24%) than males (83.30%).

The vastly asymmetric effects of serial position on publishers’ versus advertisers’ metrics suggest that native ad advertisers might overpay for lower-rank positions. This is related to the ad ranking and bidding system in the digital ad industry.

Currently in the online advertising industry, ad platforms for both search and native ads share very similar bidding and ad ranking systems. They typically utilize CTR rather than CVR for ad ranking and do not allow advertisers to pre-select the rank position for the ad. Although these systems might be fair to search ad advertisers, they place native ad advertisers at a disadvantage because a native ad’s value to the advertiser (CVR) decreases disproportionately faster than does its cost as rank position lowers.

For advertisers, the ideal bidding system would allow them to track and maximize conversions. One possible way is to place a separate bid for each rank position, with the ideal bid at each rank proportional to the expected CVR.

We recognize that these findings may create a shakeup in the native ad advertising industry as bidders strive to get the value they are paying for.

The findings are also of value to publishers. For example, our findings suggest that publishers should not eliminate lower-rank ad positions, because they can be an important source of ad revenue. They should also customize the density of native ads based on user characteristics.

Under the current bidding system used in the industry, revenue per impression (RPI) remains relatively stable across rank positions, in which RPI for rank #2 measured at 95.2% and RPI for rank #3 measured at 91.7%.

Our study also found that marketers must also adjust their practices to target specific viewer groups, accounting for such factors as gender and age. For example, if advertising is targeting females and younger customers, it is imperative to get ads with the top-rank position to optimize conversion and ROI (compared to males and older customers). Publishers may consider increasing the density of native ads in the upper portion of the webpage for males and older viewers. These implications are readily applicable because of the increasing feasibility of precision targeting.  

Read the full article.

From: Pengyuan Wang, Guiayang Xiong, and Jian Yang, “Serial-Position Effects on Native-Advertising Effectiveness: Different Results Across Publisher and Advertiser Metrics,” Journal of Marketing, 83 (March).

Go to the Journal of Marketing

Pengyuan Wang is Assistant Professor, University of Georgia.

Guiyang Xiong is Assistant Professor in Marketing, Whitman School of Management, Syracuse University.

Jian Yang is Senior Director, Oath Inc.