ArticlePDF Available

The effectiveness of public health advertisements to promote health: a randomized-controlled trial on 794,000 participants


Abstract and Figures

Online advertising: healthier ads promote healthier living People who see specific health-promoting messages after searching online for weight-related terms are more likely to subsequently search for information on weight loss interventions. A team led by Elad Yom-Tov from Microsoft Research Israel in Herzeliya conducted a randomized trial involving 794,000 users of the Bing search engine who queried terms related to weight, diet, and exercise. Randomly chosen subjects were shown advertisements designed to promote healthy living, while all other users were shown standard ads. The researchers found that 48% of people exposed to the health-promoting advertisements made searches within the next month for weight loss information, compared with only 32% of those in the control group. The findings suggest that targeted online messaging can help change population health behaviors.
Content may be subject to copyright.
The effectiveness of public health advertisements to promote
health: a randomized-controlled trial on 794,000 participants
Elad Yom-Tov
, Jinia Shembekar
, Sarah Barclay
and Peter Muennig
As public health advertisements move online, it becomes possible to run inexpensive randomized-controlled trials (RCTs) thereof.
Here we report the results of an online RCT to improve food choices and integrate exercise into daily activities of internet users.
People searching for pre-specied terms were randomized to receive one of several professionally developed campaign
advertisements or the status quo(ads that would otherwise have been served). For 1-month pre-intervention and post-
intervention, their searches for health-promoting goods or services were recorded. Our results show that 48% of people who were
exposed to the ads made future searches for weight loss information, compared with 32% of those in the control groupa 50%
increase. The advertisements varied in efcacy. However, the effectiveness of the advertisements may be greatly improved by
targeting individuals based on their lifestyle preferences and/or sociodemographic characteristics, which together explain 49% of
the variation in response to the ads. These results demonstrate that online advertisements hold promise as a mechanism for
changing population health behaviors. They also provide researchers powerful ways to measure and improve the effectiveness of
online public health interventions. Finally, we show that corporations that use these sophisticated tools to promote unhealthy
products can potentially be outbid and outmaneuvered.
npj Digital Medicine (2018) 1:24 ; doi:10.1038/s41746-018-0031-7
Hundreds of millions of dollars are spent on traditional public
health advertisements annually.
In theory, public health
advertising can save money and lives by encouraging behaviors
that prevent disease before it happens.
While the objective of
advertising investments (e.g., encouraging people to quit smok-
ing) differs from those of private advertisers (encouraging people
to purchase a good or service), the central idea is the same: to
change behaviors.
Before online advertising, it was only possible to empirically test
public health campaigns by randomizing small numbers of
participants and to examine a few outcome measures.
makes it difcult to test to whom different forms of advertisement
are best targeted.
Humans vary greatly with respect to both their biology and
their beliefs. Medical researchers use predictive analytics to mine
databases of genetic information in order to target treatments to
individuals who are more likely to respond to them. Similarly,
private advertisers use predictive analytics to mine multiple
sources of sociodemographic and behavioral data to better target
individual consumers with the goal of changing their behavior.
However, precision public health interventions have largely sat on
the sidelines both due to the large sums of money required for
targeted advertising and due to ethical concerns.
Ethical concerns arise for a number of reasons. First, participant
data are collected without informed consent.
Second, many in
public health feel uncomfortable with the idea of manipulating
individual behaviors, preferring instead to work with anonymous
means to attempt to change behavior more generically.
concerns have largely pre-empted the use of precision public
health advertising, leaving only private rms to employ these
In the private sector, Google, Microsoft, Facebook, and other
internet-based companies provide online services for free in
exchange for the information that drives precision advertising
using big data analytics. Online ads targeted using data analytics
can inuence emotions and behaviors.
First, advertisers can make educated guesses or small-scale tests
about who might respond most to a given advertisement based
on common search terms by topic. Then, advertisement can be
randomized to be shown to users of search engines that search for
such terms. Randomization provides a gold standardtest of
efcacy. Randomization can also provide causal information on
how different sub-groups (e.g., young women) respond to an
advertisement relative to others. Information on the experimental
responses of different architypesof individuals can then re-
tested with newer, more effective advertisements. This incre-
mental approachtargeting, rening, and testinghas the power
to produce online ads that affect beliefs and behaviors.
Big data companiessuch as Facebook, Google, and Microsoft
conduct tens of thousands of randomized-controlled trials
(RCTs) on their users every year.
These results are invariably kept
inside these companies, but the general process for evaluating
advertisement efcacy is likely similar across companies.
Search advertisements are typically presented as textual
advertisements that appear on a search results page coupled
with a click through link to the advertisers site. More advanced
versions include images in addition to (or instead of) the text.
While it is rare that users click on ads, online advertisements have
Received: 28 February 2018 Revised: 30 March 2018 Accepted: 9 April 2018
Microsoft Research Israel, 13 Shenkar st., 46875 Herzeliya, Israel;
J. Walter Thompson, 466 Lexington Avenue, New York, NY 10017, USA and
Global Research Analytics for
Population Health and the Department of Health Policy and Management, Mailman School of Public Health, Columbia University, 722 West 168th St., New York, NY 10032, USA
Correspondence: Elad Yom-Tov (
Published in partnership with the Scripps Translational Science Institute
been shown to increase sales both for online display ads and in
brick-and-mortar stores.
Display and search ads are believed to
produce similar impacts, and about 75% of this increase in trafc
from the cited study was from those who never clicked through.
In this paper, we take the reader through this advertising
process, and conduct, to our knowledge, the rst fully randomized
online public health communications campaign which tracks not
only click response to ads, but also the searches made prior to and
post advertisement display. Professionals from J. Walter Thomp-
son (JWT), a leading advertising rm in New York City, developed a
series of online ads aimed at improving exercise and eating
behaviors among search engine users (users) who are over-
weight. We then experimentally tested these ads using a series of
10 RCTs, each for a different textual advertisement paired with
unique click through content. Finally, we explored the impact of
the ads on changing health behaviors as measured by future
health-promotion searches.
Descriptive analysis
During the month of the RCT, the experiment ads were shown
265,279 times and clicked 1024 times. Of these displays, the ads
were shown 3108 times to 2996 users who could be tracked in
their queries before and after. Additionally, during the month of
the RCT, a total of 505,693 non-exposed users made at least one
query such as the ones which triggered a campaign ad in the
treatment population.
The majority of users were between the ages of 35 and 64, and
females were more likely to see the ads than males (Supplemental
Fig. 1). Of those over the age of 65, males and female users were
about equal in number. A total of 36 tracked users clicked on the
The tracked users and users who were not tracked were both
successfully randomly assigned (Table 1).
Click through rate
The click rate is congruent with the click rate in other advertising
As shown in Supplemental Fig. 1, females were more
likely to use terms which triggered the campaign ads, but there
was a trend toward males having a higher click through.
Exposure to textual ads and future target searches
A model predicting future target searches from prior interest in
target searches and from exposure to campaign ads reaches an R
or 0.314 (p<10
). As shown in Table 2, prior interest in target
searches increases the likelihood of future target searches by 52%
(slope =0.52, standard error [SE] =0.001; <10
). However,
exposure to campaign ads signicantly increases the likelihood
of future target searches by 15% (slope =0.15, SE =0.01; p<
), especially in the absence of prior target searches. Stated
differently, 48% of people who were exposed to the ads made
future target searches, compared to 32% of the controls (a 50%
increase). This difference is even greater when observing the
population which did not have past target searches (30% vs. 15%).
Predictive analytics
We constructed a model to predict future target searches in the
treatment population. Using only respondent characteristics (both
behavioral and demographic) produced a model with an R
0.414. When only previous query topics were added, the R
0.410. When both were added, the R
was 0.491.
Cox hazards analyses
Table 3shows the hazards ratios for the likelihood of future target
searches for the sociodemographic and contextual characteristics
of the ads. Recall that we correct p-values for the number of
comparisons within each category. We discuss statistically
signicant results here. While the number of previous searches
for keywords is associated with a very slight change in the HR for
future keyword searches, the average person tends to make a
large number of searches. None of the ads were particularly more
effective than other ads in evoking a future keyword search.
However, exposure to more than one ads increases the chance of
a keyword search by 11% (HR =1.11; 95% CI =1.03, 1.20). Females
were much less likely than males to perform a future search for
keywords when exposed to an advertisement HR =0.84 (95% CI
=0.76, 0.91).
We randomized Bing users to receive a professionally designed
public health advertisement or to receive control (status quo)
advertisements. We found that people who view online health
promotion advertisements are much more likely to perform
searches related to health promotion than those who were
assigned status quoadvertisements. The experimental effect
sizes were large, with 48% of those with exposure to the text
messages (and in some cases, the landing pages) more likely to
perform future health-related searches while only 32% of the
matched control group performed such searches.
At the population level, searching for specic health behaviors
is associated with performing these behaviors in the physical
For example, the number of people searching for
information about cannabis is highly correlated with the known
number of users of cannabis,
the number of people searching
for specic medicines corresponds to the number of prescriptions
and the distribution of birth events, as inferred from search
queries, is extremely well aligned with the distribution provided
by the Centers for Disease Control.
We show that it is possible to
Table 1. Percentage of users by age group and by gender among all
people exposed to the ads and among the tracked population
All users Tracked users
By age group Control Treatment Control Treatment
1317 1.0 0.9 1.3 1.2
1824 5.3 9.1 6.5 8.8
2534 10.0 12.7 11.0 12.5
3549 28.7 32.4 29.3 32.4
5064 44.9 37.4 43.6 37.5
65+10.1 7.5 8.4 7.6
By gender
Male 31.0 30.8 26.9 30.4
Female 69.0 69.2 73.1 69.6
None of the differences are statistically signicant (χ
, two-sided, p> 0.05)
Table 2. Model of the likelihood that a user will make future target
# Parameter Slope (SE) t-stat p-value
1 User made past target searches 0.525 (0.0007) 699 <10
2 User is in the treatment
0.149 (0.0106) 14 <10
3 Interaction of (1) and (2) 0.202 (0.0143) 14 <10
The effectiveness of public health advertisements
E Yom-Tov et al.
npj Digital Medicine (2018) 24 Published in partnership with the Scripps Translational Science Institute
alter the behavior of those with enough interest to conduct a
search online, and show that it is possible to test such behavioral
changes experimentally. With online advertisements, it is no
longer necessary to stab in the dark with public health
advertisement design. Nor is it necessary to guess who will
respond to those advertisements. Rather, it is possible to
systematically target users with advertisements to which they
are most susceptible, thereby eliciting behavior change. Our
identication strategies can, in theory, be used to continuously
rene, randomize, and test the targeting algorithms on different
user types. For instance, it is not only possible to target based on
the usersage, race, and location, but also on their characteristics
as dened by their internet searches, shopping preferences, and
even email content. The targeting algorithms can use the
information to be stepped upuntil there is evidence that the
user changes his or her behaviors.
Our study was susceptible to few limitations, including those
typically inherent to RCTs. Perhaps the most important limitation
is external validity since we ran the campaign on only one
platform. Second, it is difcult to quantify the impact of the
counterfactual advertisements that were shown to users. The
counterfactual could be health promoting (e.g., gym member-
ships), neutral (e.g., vitamin supplements), or negative (e.g.,
unhealthy foods or products targeted toward high-risk groups).
It is therefore possible that an ad with no efcacy could appear
efcacious if the bulk of counterfactual advertisements discour-
aged future keyword searches.
Experimental ofine advertisements have shown that it is
possible to motivate health behavior change with traditional
advertising modalities, such as associating behaviors with those of
desirable social groups.
The only published RCT of online health
promotion advertisements we are aware of demonstrated that it is
different audiences respond very differently to a given advertise-
For example, empowering advertisements were generally
more effective at inducing future searches on smoking cessation
than those that emphasized the negative health impacts of
smoking. But this varied dramatically by the demographic
characteristics of the viewers.
This also raises signicant concerns for health departments and
other agencies that could greatly benet from access to the data
underlying the health advertisements. Most large information
technology rms provide services to users for free in exchange for
access to user data. To best understand how to target users and
change their behavior, it would be useful for them to have access
to identied data that could be linked across multiple sources of
big data. Like academic institutions, most public agencies require
approvals that are difcult to obtain in part because institutional
review boards have not adjusted to the nuances of big data
research and partly because there is no clear opt in for users of
online services. Clearly, cooperation between ethicists, big data
companies, governmental bodies, and academia has great
potential to advance population health. We show that it is not
only technically possible to launch an online campaign that
effectively improves health behaviors, but also that corporations
that promote an unhealthy diet or a sedentary lifestyle can
potentially be outbid or outmaneuvered.
Our RCT was conducted by Microsoft during April 2017 using the Bing Ads
system. In this system, advertisers bid to place the ads when specic
keywords are searched by users of the Bing search engine. Internet users of
the Bing search engine who were logged into a Microsoft account and
searching for pre-specied keywords were selected for this study. Eligible
users were randomly exposed to JWT ads (treatment) or any other ads
served up by the system (control). We then followed both the treatment
and control usersfuture search queries, and retrospectively examined past
queries to build and interpret predictive models.
This study was approved by the Microsoft Institutional Review Board and
was declared exempt by the Columbia University Institutional Review
Board under the understanding that the Columbia University researchers
would not have access to the data in any form other than the tabular
results presented in this paper, and further that they would not seek
funding for the study.
This trial was registered on February 2018 with,
registration number NCT03439553.
User selection
Those who are motivated to change their behaviors are more likely to do
so. As a result, advertisers often attempt to target individuals with some
motivation to change. In this study, we attempt to improve the viewers
diets and increase their levels of physical activity. The goal of the user
selection process was therefore to identify individuals who were motivated
for behavioral change due to social stigma or disease, and then present an
advertisement that suggests a behavioral change that is within reach given
their lifestyle. We therefore selected users who used search terms
associated with social stigma or diseases related to poor diet or low
levels of exercise.
The Bing Ads system is designed to randomize advertisements for
experimental purposes. In this study, we selected users for inclusion if they
(1) were using the Bing search engine; (2) logged into a Microsoft account;
and (3) typed any of the following combination of terms:
1. (Weight, Overweight, Obesity, lose weight) AND (<none>, Hyperten-
sion, High cholesterol, High blood pressure, Exercise, Diabetes,
Table 3. Cox proportionate hazards ratios (HR) and 95% condence
intervals (CI) associated with searcher characteristics
HR 95% CI
Number of past searches by the user 1.0001
Number of past target searches by the
Number of past ads shown to the user 1.1146
Hour of the day 1.0062 [0.9995241.012954]
Title Burn calories sitting0.9820 [0.8885911.075501]
Title Chores: The New Workout0.9362 [0.8164411.055910]
Title Drink Water Lose Pound1.0311 [0.9263411.135788]
Title Find a Hairy Partner0.9642 [0.8062211.122249]
Title Laugh Your Calories Off0.9754 [0.8164401.134264]
Title Lose weight watching TV0.9840 [0.8324131.135643]
Title Pimp Up Your Snack1.0125 [0.7936171.231302]
Title Your kids are an exercise0.9117 [0.7279721.095497]
Title 'Swalty' Snacks are best0.9774 [0.7973001.157544]
Page number 0.9561 [0.7667381.145376]
Page position: Top 1.0094 [0.9106391.108114]
Page position: Right 0.9014
Page position rank 1.0191 [0.9981091.040139]
Age group 1824 1.0912 [0.9195501.262805]
Age group 2534 1.1736
Age group 3549 1.1160 [0.9800891.251910]
Age group 5064 1.0937 [0.9574001.230037]
Age group 65+1.1636 [0.9852471.341941]
Gender: Female 0.8369
Ad clicked? 1.0461 [0.6822351.410057]
Denotes variables statistically signicant at p< 0.05. Variables with the
prextitlerefer to specic advertisements
The effectiveness of public health advertisements
E Yom-Tov et al.
Published in partnership with the Scripps Translational Science Institute npj Digital Medicine (2018) 24
2. Hypercholesterolemia, Fat, BMI, Body fat, Big gut, Big and tall
clothing, Easy exercises, Healthy diet, Easy workout, Plus size, Weight
loss pill, Diet pill, Weight loss surgery, XXL
The vast majority of Bing search engine users who typed the above
keywords (n=283,716) were excluded based on a missing Microsoft
account pre-randomization (the account is needed for user demographic
data and analytics). Users who had a Microsoft account were further
analyzed (Fig. 1). Among users with a Microsoft account, those with
incomplete demographic data (age, gender, zip code) were also excluded,
leaving 2996 treated participants and 505,693 control participants. The
CONSORT diagram (Fig. 1) and the age and gender characteristics of the
users (Table 1) show no threats to internal validity. There were no
statistically signicant differences in the demographic characteristics of the
treatment and control users (χ
test, p> 0.05).
In addition to the above four criteria for inclusion of users, campaign ads
also had to competitively bid for an ad on the search results page.
Keyword demand differs by advertisers, and so does the associated
maximum bid for each keyword. To account for the differences in keyword
demand and have a similar baseline for all keywords, we set the Bing Ads
system to automatically adjust the bids for each of the campaign terms
listed above to be high enough for our ads to be as competitive as control
ads (i.e., those of other advertisers), but no more than US$1 per click.
User characterization
We extracted all queries made on Bing by treatment and control users in
our trial, from 1 month before the rst advertisement was shown through
until 1 month after the last ad was shown.
For each query, we registered an anonymized user identier, the time of
the query, the US county from which the user made the query, and the text
of the query. The query was further classied (using a proprietary classier
developed by Microsoft) into one or more of approximately 60 topical
categories. These categories were encompassed broad topics, such as
commerce, travel, and health.
Users exposed to ads were further characterized by their self-reported
age, gender, and the county-level poverty as inferred from the county from
which they made the query.
JWT developed both the textual ads initially displayed to treatment users,
as well as the content on the landing page shown when the user clicks on
a textual link. The advertisements were grounded in the Fogg Behavior
In this model, three elements must come together at the same
time: motivation to change, ability to change, and a trigger for change. The
ads were designed to be hot triggers,
designed to prime highly
motivated users with content that is easy and actionable in order to nudge
a behavior change toward more positive health habits.
All treatment users that enter the keywords above are exposed to the
textual ads. The associated landing pages contained information on how
the subject might improve health behaviors through nutrition or exercise.
However, the vast majority of users will only view the textual advertise-
ments. The JWT text ads fall into three categories: (1) suggestive of healthy
behavior change (Laugh your calories off,Chores: the new workout,
Your kids are an exercise,'Swaltysnacks are best); (2) related to
nutrition or exercise, but not to behavior change (Burn calories sitting,
Lose weight watching TV,Pimp up your snack); or (3) unrelated to both
behavior change and nutrition or exercise (Find a hairy partner). Each
textual ad was designed to motivate users to click on the ad. Both the
textual and click through advertisements were explicitly designed to avoid
stigmatizing obesity.
The landing page ads focused solely on nudging users toward changing
their behaviors with suggestions for incorporating small amounts of
exercise or easy dietary changes into day-to-day activities. These were
accompanied by an animated image meant to reinforce the message of
the advertisement (see the Supplemental Appendix). Users were also
provided links to additional content developed by professional health
organizations or the Centers for Disease Control and Prevention if they
wanted more information.
Fig. 1 CONSORT 2010 ow diagram. CONSORT ow diagram template courtesy of
The effectiveness of public health advertisements
E Yom-Tov et al.
npj Digital Medicine (2018) 24 Published in partnership with the Scripps Translational Science Institute
Outcomes and predictor variables
Our primary outcome measure was the likelihood of a future search using
a set of pre-specied keywords. These keywords were selected by
identifying common weight-related search terms among Bing users. The
terms fell into categories that suggest that the subject either (1) desires a
deeper understanding of obesity (fat; nutrition; calories; body mass index;
BMI; body weight; body mass) or (2) wishes to change their behavior
(weight loss; weight watcher; weightwatcher; losing weight; and lose
We were interested in exploring differences in outcome measures for
treated and control users overall, by demographic characteristics, and by
advertisement characteristics (content, placement, etc.). We were also
interested in building predictive analytics that could identify which user
types are most likely to respond to a given advertisement.
We used the following covariates to operationalize demographic and
advertisement characteristics:
1. Past user behavior:
a. Number of past searches by the user
b. Number of past target searches by the user
c. Number of past ads shown to the user
2. User demographic:
a. Age group (categorized into six groups: 1317, 1824, 2534,
3549, 5064, or 65+years)
b. Gender (female or male)
3. Advertisement information:
a. Hour of the day ad shown (integer between 0 and 23 h)
b. Advertisement title (categorized into 10 groups)
c. Was the ad clicked? (yes/no)
d. Search page number on which the ad is displayed (integer
between 1 and 100)
e. Search page position (indicator variable for whether the ad was
placed on the top or the right-hand side of the search page)
Statistical analysis
Given the large sample size, we specied an effect size of greater than or
equal to 10% to be meaningful.
We explored the likelihood of future target searches given user
exposure to ads, controlling for past searches. We used ordinary least
squares regression to model the association between variables:
where yis an indicator of future searches, and xthe predictors of the
Because previous searches predict the probability of subsequent
searches, we were also interested in the interaction term between the
probability of a subsequent search given exposure to the treatment (the
interaction between the coefcient of conducting a previous search and
being in the treatment group).
We then developed a predictive model. In this model, each user was
proled prior to the ad campaign with respect to demographic
characteristics and previous topical searches. With respect to topical
searches, we explored whether the user had performed searches that
relate to one of 60 pre-specied categories of interest. These included
broader topics, such as shopping, travel, and health. By adding a term to
the above equation that includes previous searches in each of these
categories, it becomes possible to examine the inuence of inclusion of the
topic on the models predictive value, as measured by the models
These models included all 10 of the covariates listed above. These
covariates are used for predictive purposes, and regression is conducted
on a cohort that has already been randomized. This way, it becomes
possible to make predictions based on treatment response when only
treatment status introduces non-random variation.
Next, we used Cox proportionate hazards models and explored 32
predictors of future searches:
HR ¼exp X1β1þ¼þXNβN
where HR is the hazard ratio, Xthe predictors of the model, and βtheir
corresponding model coefcients.
The predictors fell into ve broad characteristics of the user and their
exposures: previous searches, exposure to our advertisements, advertise-
ment placement characteristics, age, gender, and poverty. We used
Bonferonni correction for the number of categorical variables within each
of these broader categories. We examined the HR for future searches for
various user and ad characteristics.
In a secondary analysis (see Supplementary Materials), we used
propensity score matching of users meeting inclusion characteristics,
who were matched to unexposed users based on age, gender, and zip
code, and analyzed using the above characteristics. This analysis allows for
a low-noise, low sample size analysis in which it becomes possible to
obtain very conservative assurances that there are statistically signicant
differences by treatment status, rather than relying on clinically meaningful
effect sizes (as in the parent analysis).
Data availability
The data that support the ndings of this study are available from
Microsoft, but restrictions apply to the availability of the data. Specically,
all aggregate advertising data are available from the authors on reasonable
request. Individual-level search data are available from the authors on
reasonable request and with permission of Microsoft.
The authors wish to thank Nicholas Orsini, Zeynep Cingir, Javier Pinol, Yudi Rojas,
Gustavo Tezza, Valerie OBert, Vaibhav Bhanot, and Pritika Mathur for their help in
designing the ads.
P.M. devised the study. S.J. and S.B. designed the ads and landing pages. All authors
decided on the keywords. E.Y.T. ran the advertising campaign, collected the data, and
analyzed it. All authors were involved in writing the paper. This work was carried out
as part of the authors salaried employment, with no specic funding.
Supplementary Information accompanies the paper on the npj Digital Medicine
website (
Competing interests: E.Y.T. is an employee of Microsoft, owner of the Bing search
engine. The authors declare no competing interests.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims
in published maps and institutional afliations.
1. Atlantis, E., Salmon, J. & Bauman, A. Acute effects of advertisements on childrens
choices, preferences, and ratings of liking for physical activities and sedentary
behaviours: a randomised controlled pilot study. J. Sci. Med. Sport 11, 553557
2. Berger, J. & Rand, L. Shifting signals to help health: using identity signaling to
reduce risky health behaviors. J. Consum. Res. 35, 509518 (2008).
3. Snyder, L. B. Health communication campaigns and their impact on behavior. J.
Nutr. Educ. Behav. 39, S32S40 (2007).
4. Snyder, L. B. et al. A meta-analysis of the effect of mediated health commu-
nication campaigns on behavior change in the United States. J. Health Commun.
9,7196 (2004).
5. Witte, K. & Allen, M. A meta-analysis of fear appeals: Implications for effective
public health campaigns. Health Educ. Behav. 27, 591615 (2000).
6. Nutbeam, D. Health literacy as a public health goal: a challenge for contempor ary
health education and communication strategies into the 21st century. Health
Promot. Int. 15, 259267 (2000).
7. Mathieson, S. A. DH doubled ad spending to £60m. The Guardian https://www.
advertising-spending-60m (2017).
8. Rice, R. E. & Atkin, C. K. Public Communication Campaigns. (Sage, Thousand Oaks,
CA, 2012).
9. Grady, C. Enduring and emerging challenges of informed consent. N. Engl. J. Med.
372, 855862 (2015).
The effectiveness of public health advertisements
E Yom-Tov et al.
Published in partnership with the Scripps Translational Science Institute npj Digital Medicine (2018) 24
10. Kramer, A. D., Guillory, J. E. & Hancock, J. T. Experimental evidence of massive-
scale emotional contagion through social networks. Proc. Natl. Acad. Sci. USA 111,
87888790 (2014).
11. Zuboff, S. Big other: surveillance capitalism and the prospects of an information
civilization. J. Inform. Technol. 30,7589 (2015).
12. Andreu-Perez, J., Poon, C. C., Merrield, R. D., Wong, S. T. & Yang, G.-Z. Big data for
health. IEEE J. Biomed. Health 19, 11931208 (2015).
13. Ruggeri, K., Yoon, H., Kácha, O., van der Linden, S. & Muennig, P. Policy and
population behavior in the age of Big Data. Curr. Opin. Behav. Sci. 18,16 (2017).
14. Kohavi, R., Crook, T., Longbotham, R. Online Experimentation at Microsoft. Third
Workshop on Data Mining Case Studies and Practice Prize. Proceedings of the 13th
ACM SIGKDD international conference on Knowledge discovery and data mining
(Association of Computing Machinery (ACM), San Jose, CA, 2009).
15. Lewis, R. A. & Reily, D. H. Online Ads and onine sales: measuring the effects of
retail advertising via a controlled experiment on Yahoo! QME-Quant. Mark. Econ.
12, 235266 (2014).
16. Yom-Tov, E., Muennig, P. & El-Sayed, A. M. Web-based antismoking advertising to
promote smoking cessation: a randomized controlled trial. J. Med. Internet Res. 8,
e306 (2016).
17. Yom-Tov, E. Crowdsourced Health: How What You Do on the Internet Will
Improve Medicine. (MIT Press, Cambridge, MA, 2016.
18. Yom-Tov, E. & Lev-Ran, S. Adverse reactions associated with cannabis con-
sumption as evident from search engine queries. JMIR Public Health Surveill. 3,
e77 (2017).
19. Yom-Tov, E. & Gabrilovic h, E. Postmarket drug surveillance without trial costs:
discovery of adverse drug reactions through large-scale analysis of web search
queries. J. Med. Internet Res. 15, e124 (2013).
20. Fourney, A., White, R. W., Horvitz, E. Exploring time-dependent concerns about
pregnancy and childbirth from search logs. 33rd Annual ACM Conference on
Human Factors in Computing Systems, 737746 (Seoul, Republic of Korea, 2015).
21. US Bureau of the Census. Census 2010.
cen2000.html (2010).
22. Fogg, B. J. Fogg Behavior Model.
KATA_Files/FBM.pdf (The author, 2007).
23. Fogg, B. J. A behavior model for persuasive design. Proceedings of the 4th
International Conference on Persuasive Technology, Claremont, CA (Association of
Computing Machinery (ACM), New York, NY, 2009).
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons license, and indicate if changes were made. The images or other third party
material in this article are included in the articles Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not included in the
articles Creative Commons license and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly
from the copyright holder. To view a copy of this license, visit http://creativecommons.
© The Author(s) 2018
The effectiveness of public health advertisements
E Yom-Tov et al.
npj Digital Medicine (2018) 24 Published in partnership with the Scripps Translational Science Institute
... To naturally apply the memory model in a web environment, we focused on web advertisement such as behavioral targeting. Several studies have indicated the potential of behavioral changes toward healthy behavior through this type of online media (Kramer et al., 2014;Yom-Tov et al., 2018). In our system, the visited product images are always presented in the right region of a web page. ...
... The contribution of the present study is to extend the principle of affective computing by including a computational cognitive modeling of memory. This model differs from the previous behavioral model (Fogg, 2009) used in web advertisement (Yom-Tov et al., 2018) in that it includes internal memory processes. Although there are several options for modeling emotion and memory [e.g., Friston (2010), Schmidhuber (2010), as mathematical models of curiosity], we claim that including the ACT-R cognitive architecture provides another theoretical basis of implicit-prompting systems designed to adapt the emotional states of users based on an academic field with a long history. ...
Full-text available
Even though the web environment facilitates our daily life, emotional problems caused by its incompatibility with human cognition are becoming increasingly serious. To alleviate negative emotions during web use, we developed a browser extension that presents memorized product images to users in the form of web advertisements. This system utilizes the cognitive architecture Adaptive Control of Thought-Rational (ACT-R) as a model of human memory and emotion. A heart rate sensor attached to the user modulates the ACT-R model parameters, and the emotional states represented by the model are synchronized (following the chameleon effect) or counterbalanced (following the homeostasis regulation) with the physiological state of the user. An experiment demonstrates that the counterbalanced model suppresses negative ruminative web browsing. The authors claim that this approach, utilizing a cognitive model, is advantageous in terms of explainability.
... To naturally apply the memory model in a web environment, we focused on web advertisement such as behavioral targeting. Several studies have indicated the potential of behavioral changes toward healthy behavior through this type of online media (Kramer et al., 2014;Yom-Tov et al., 2018). In our system, the visited product images are always presented in the right region of a web page. ...
... The contribution of the present study is to extend affective computing by including a computational cognitive modeling of memory. This model differs from the previous behavioral model (Fogg, 2009) used in web advertisement (Yom-Tov et al., 2018) in that it includes internal memory processes. Although there are several options for modeling emotion and memory (Friston, 2010;Schmidhuber, 2010), we claim that including the ACT-R cognitive architecture provides another theoretical basis of implicit-prompting systems designed to adapt the emotional states of users based on an academic field with a long history. ...
Full-text available
Even though the web environment facilitates daily life, emotional problems caused by its incompatibility with human cognition are becoming increasingly serious. To alleviate negative emotions during web use, we developed a browser extension that presents memorized product images to users, in the form of web advertisements. This system utilizes the cognitive architecture Adaptive Control of Thought-Rational (ACT-R) as a model of memory and emotion. A heart rate sensor modulates the ACT-R model parameters: The emotional states of the model are synchronized or counterbalanced with the physiological state of the user. An experiment demonstrates that the counterbalance model suppresses negative ruminative web browsing. The authors claim that this approach is advantageous in terms of explainability.
... This exemplified the fact that physicians may encounter barriers due to public opinion, which may be in part shaped by critical media publications such as this production. It reinforced the importance of health advertising in public health, as seen in other campaigns [29]. Interestingly, non-statin lipid-lowering therapy remained steady in the same period, suggesting that compromise and a shift to alternative forms of therapy may be an effective mechanism to address patient concerns, at least temporarily. ...
Full-text available
Background and Objectives: Statins have been extensively utilised in atherosclerotic cardiovascular disease (ASCVD) prevention and can inhibit inflammation. However, the association between statin therapy, subclinical inflammation and associated health outcomes is poorly understood in the primary care setting. Materials and Methods: Primary care electronic health record (EHR) data from the electronic Practice-Based Research Network (ePBRN) from 2012–2019 was used to assess statin usage and adherence in South-Western Sydney (SWS), Australia. Independent determinants of elevated C-reactive protein (CRP) were determined. The relationship between baseline CRP levels and hospitalisation rates at 12 months was investigated. Results: The prevalence of lipid-lowering medications was 14.0% in all adults and 44.6% in the elderly (≥65 years). The prevalence increased from 2012 to 2019 despite a drop in statin use between 2013–2015. A total of 55% of individuals had good adherence (>80%). Hydrophilic statin use and higher intensity statin therapy were associated with elevated CRP levels. However, elevated CRP levels were not associated with all-cause or ASCVD hospitalisations after adjusting for confounders. Conclusions: The prevalence and adherence patterns associated with lipid-lowering medications highlighted the elevated ASCVD-related burden in the SWS population, especially when compared with the Australian general population. Patients in SWS may benefit from enhanced screening protocols, targeted health literacy and promotion campaigns, and timely incorporation of evidence into ASCVD clinical guidelines. This study, which used EHR data, did not support the use of CRP as an independent marker of future short-term hospitalisations.
... weight loss) by 50%. 30 Social media also offers opportunities for media campaigns. Social media messages recorded by health professionals before the winter holidays in the United States led to a significant reduction in holiday travel and subsequent COVID-19 infections. ...
Background It is unknown whether online information about the benefits and harms of surgery contains an accurate description of evidence. Objective To describe the proportion of webpages containing information about surgery for spinal pain (decompression and fusion) that accurately described the evidence on the benefits of surgery, described harms, and provided quantitative estimates of these harms. Methods We performed a content analysis of webpages containing information about spine surgery. Two reviewers identified webpages and extracted data. Primary outcomes were the proportion of webpages that accurately described the evidence on the benefits, described harms, and provided quantitative estimates of these harms. Results We included 117 webpages. Only 29 (25%) webpages accurately described the evidence on the benefits of spine surgery, and more webpages on decompression accurately described the evidence compared to webpages on fusion (31% vs 15%, difference in proportions = 16%; 95% CI: 2%, 31%). Harms of surgery were described in most webpages (n = 76, 65%), but a much smaller proportion of webpages (n = 18, 15%) provided a quantitative estimate for the mentioned harms. Conclusions Most webpages failed to accurately describe the benefits and harms of decompression and fusion surgeries for spinal pain. Unbiased consumer resources and educating the public on how to critically evaluate health claims are important steps to improve knowledge on the benefits and harms of spine surgery.
... [7][8][9][10] Further, determining what messaging appeals to which groups of people remains critical. 11 For example, a review of vaccine communication found participant characteristics could be moderators of the effect of goal-framed versus loss-framed messaging (eg, perceived risk or loss), but again findings are inconsistent. 12 The existing literature does reinforce, however, that adaptation is needed to fit local contexts. ...
Full-text available
The public’s need for timely and trusted COVID-19 information remains high. Governments and global health agencies such as the WHO have sought to disseminate accurate and timely information to counteract misinformation and disinformation that has arisen as part of an ‘infodemic’—the overabundance of information on COVID-19—some accurate and some not. In early 2020, WHO began a collaboration with Google to run online public service announcements on COVID-19, in the form of search ads displayed above results of Google Search queries. Web-based text ads can drive online searchers of COVID-19 information to authoritative COVID-19 content but determining what message is most effective is a challenge. WHO wanted to understand which message framing, that is, the way in which ad information is worded for the public, leads searchers to click through to WHO content. WHO tested 71 text ads in English across four COVID-19 topics using a mix of message frames: descriptive, collective, gain, loss, appeals to values and emphasising reasons. Between 11 September 2020 and 23 November 2020, there were 13 million views of the experimental WHO text ads leading to 1.4 million click-throughs to the WHO website. Within the set of 71 ads, there was a large spread between the most effective and least effective messages; for messages on COVID-19, the best performing framings were more than twice as effective as the worst performing framings (18.7% vs 8.5% engagement rate). Health practitioners can apply the messaging tactics WHO found to be successful to rapidly optimise messages for their own public health campaigns and better reach the public with authoritative information. Similar collaboration between big technology companies and governments and global health agencies has the potential to advance public health.
... A randomised controlled trial demonstrated that passively collected data from social media could be used to personalise health promotion messages for improved efficacy. 24 It has also been demonstrated that during pandemics, when the public acquires abundant disease-related information via social media, consuming social media can influence disease prevention behaviours. 25 Overall, these types of non-traditional data are applicable to disease surveillance and prevention. ...
Full-text available
Introduction Cardiometabolic diseases, including cardiovascular disease, obesity and diabetes, are leading causes of death and disability worldwide. Modern advances in population-level disease surveillance are necessary and may inform novel opportunities for precision public health approaches to disease prevention. Electronic data sources, such as social media and consumer rewards points systems, have expanded dramatically in recent decades. These non-traditional datasets may enhance traditional clinical and public health datasets and inform cardiometabolic disease surveillance and population health interventions. However, the scope of non-traditional electronic datasets and their use for cardiometabolic disease surveillance and population health interventions has not been previously reviewed. The primary objective of this review is to describe the scope of non-traditional electronic datasets, and how they are being used for cardiometabolic disease surveillance and to inform interventions. The secondary objective is to describe the methods, such as machine learning and natural language processing, that have been applied to leverage these datasets. Methods and analysis We will conduct a scoping review following recommended methodology. Search terms will be based on the three central concepts of non-traditional electronic datasets, cardiometabolic diseases and population health. We will search EMBASE, MEDLINE, CINAHL, Scopus, Web of Science and Cochrane Library peer-reviewed databases and will also conduct a grey literature search. Articles published from 2000 to present will be independently screened by two reviewers for inclusion at abstract and full-text stages, and conflicts will be resolved by a separate reviewer. We will report this data as per the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews. Ethics and dissemination No ethics approval is required for this protocol and scoping review, as data will be used only from published studies with appropriate ethics approval. Results will be disseminated in a peer-reviewed publication.
... Beyond online search engine data, which are already being used to influence digital health interventions [14,15], the relevance of footprints captured by likes, comments, and shares on social media platforms, including Facebook, Twitter, and Instagram, remains largely unanalyzed and unexplored. Compared with traditional mass media channels, the targeted advertising tools (TATs) available through such sites are already being used by some researchers to recruit study participants [16], create representative samples [17], identify people with particular characteristics [18,19], and obtain public health insights in the United States [20]. ...
Full-text available
Although established marketing techniques have been applied to design more effective health campaigns, more often than not, the same message is broadcasted to large populations, irrespective of unique characteristics. As individual digital device use has increased, so have individual digital footprints, creating potential opportunities for targeted digital health interventions. We propose a novel precision public health campaign framework to structure and standardize the process of designing and delivering tailored health messages to target particular population segments using social media-targeted advertising tools. Our framework consists of five stages: defining a campaign goal, priority audience, and evaluation metrics; splitting the target audience into smaller segments; tailoring the message for each segment and conducting a pilot test; running the health campaign formally; and evaluating the performance of the campaigns. We have demonstrated how the framework works through 2 case studies. The precision public health campaign framework has the potential to support higher population uptake and engagement rates by encouraging a more standardized, concise, efficient, and targeted approach to public health campaign development.
Full-text available
As mensagens publicitárias são dotadas de argumentos persuasivos capazes de influenciar crenças, opiniões, atitudes e comportamentos impactando profundamente a saúde pública de uma determinada geografia. Este trabalho, neste sentindo, objetiva mapear as pesquisas sobre publicidade e propaganda e saúde pública na Coleção Saúde Pública da base de dados da SciELO (Scientific Electronic Library Online) por ser um repositório de acesso aberto que disponibiliza conteúdo mundial com predomínio no contexto ibero-americano. Os trabalhos analisados contemplaram todos os estudos disponíveis até o ano de 2018, sendo os termos utilizados na busca publicidade e propaganda em português, inglês e espanhol cruzados com o termo Saúde Pública. Os resultados demonstraram uma pequena quantidade de publicações ao longo dos últimos 35 anos, uma média de 3,14 artigos publicados por ano. Encontramos também uma clara predominância dos estudos relacionada às formações das Ciências da Saúde dos pesquisadores, o que demonstra que a temática selecionada não tem atraído as atenções dos pesquisadores das ciências sociais aplicadas. Ademais, os temas trabalhados em estudo têm recaído invariavelmente sobre questões da regulamentação publicitária, das mensagens persuasivas (análises do seu potencial persuasivo) e mercado de consumo, sobretudo, de produtos e substâncias que ocasionam mal à saúde ou que promoveriam a saúde. Por fim, vimos que a maioria dos estudos estavam concentrados em periódicos brasileiros e foram realizados por pesquisadores brasileiros, apontando um claro interesse do tema concentradamente neste país.
Full-text available
Search advertising, a popular method for online marketing, has been employed to improve health by eliciting positive behavioral change. However, writing effective advertisements requires expertise and experimentation, which may not be available to health authorities wishing to elicit such changes, especially when dealing with public health crises such as epidemic outbreaks. Here, we develop a framework, comprising two neural network models, that automatically generates ads. The framework first employs a generator model, which creates ads from web pages. These ads are then processed by a translation model, which transcribes ads to improve performance. We trained the networks using 114K health-related ads shown on Microsoft Advertising. We measure ad performance using the click-through rates (CTR). Our experiments show that the generated advertisements received approximately the same CTR as human-authored ads. The marginal contribution of the generator model was, on average, 28% lower than that of human-authored ads, while the translator model received, on average, 32% more clicks than human-authored ads. Our analysis shows that, when compared to human-authored ads, both the translator model and the combined generator + translator framework produce ads reflecting higher values of psychological attributes associated with a user action, including higher valence and arousal, and more calls to action. In contrast, levels of these attributes in ads produced by the generator model alone are similar to those of human-authored ads. Our results demonstrate the ability to automatically generate useful advertisements for the health domain. We believe that our work offers health authorities an improved ability to build effective public health advertising campaigns.
Background Men who have sex with men (MSM) increasingly use internet-based websites and geospatial apps to seek sex. Though these platforms may be useful for public health intervention, evaluations of such interventions are rare. We sought to evaluate the online behavior of young MSM of color in Philadelphia and the effectiveness of using ads to link them to, where users can order free condoms, lubricant, and sexually transmitted infection test kits delivered via the U.S. postal service. Method Data collection and analyses were conducted in two phases. First, we performed keyword research and analyzed web browser logs using a proprietary data set owned by Microsoft. Subsequently, we ran a Google Ads campaign using the keywords identified in the preliminary phase, and directed targeted users to the condom or test kit ordering pages. Results were analyzed using MATLAB 2018. Results Test kit advertisements received 5,628 impressions, 157 clicks, and 18 unique conversions. The condom advertisements received 128,007 impressions, 2,583 clicks, and 303 unique conversions. Correlation between the click-through rate and the conversion rate per keyword was ρ = −.35 ( P = .0096) and per advertisement was ρ = .40 ( P = .14). Keywords that directly related to condoms were most effective for condom ordering (42% conversion rate vs. ≤2% for other classes), while keywords emphasizing the adverse effects of unprotected sex were most effective in test kit ordering (91% conversion rate vs. 13% and 12% for other classes). Conclusions Online advertisements seemed to affect real-world sexual health behavior, as measured by orders of condoms and test kits, among a group of young MSM living in the same community.
Full-text available
Background: Cannabis is one of the most widely used psychoactive substances worldwide, but adverse drug reactions (ADRs) associated with its use are difficult to study because of its prohibited status in many countries. Objective: Internet search engine queries have been used to investigate ADRs in pharmaceutical drugs. In this proof-of-concept study, we tested whether these queries can be used to detect the adverse reactions of cannabis use. Methods: We analyzed anonymized queries from US-based users of Bing, a widely used search engine, made over a period of 6 months and compared the results with the prevalence of cannabis use as reported in the US National Survey on Drug Use in the Household (NSDUH) and with ADRs reported in the Food and Drug Administration's Adverse Drug Reporting System. Predicted prevalence of cannabis use was estimated from the fraction of people making queries about cannabis, marijuana, and 121 additional synonyms. Predicted ADRs were estimated from queries containing layperson descriptions to 195 ICD-10 symptoms list. Results: Our results indicated that the predicted prevalence of cannabis use at the US census regional level reaches an R(2) of .71 NSDUH data. Queries for ADRs made by people who also searched for cannabis reveal many of the known adverse effects of cannabis (eg, cough and psychotic symptoms), as well as plausible unknown reactions (eg, pyrexia). Conclusions: These results indicate that search engine queries can serve as an important tool for the study of adverse reactions of illicit drugs, which are difficult to study in other settings.
Full-text available
Background: Although hundreds of millions of dollars are spent each year on public health advertising, the advertisement content, design, and placement are usually developed by intuition rather than research. Objective: The objective of our study was to develop a methodology for testing Web-based advertisements to promote smoking cessation. Methods: We developed 10 advertisements that varied by their content (those that empower viewers to quit, help viewers to quit, or discuss the effects of smoking). We then conducted a series of Web-based randomized controlled trials that explored the effects of exposing users of Microsoft's Bing search engine to antismoking advertisements that differed by content, placement, or other characteristics. Finally, we followed users to explore whether they conducted subsequent searches for smoking cessation products or services. Results: The advertisements were shown 710,106 times and clicked on 1167 times. In general, empowering advertisements had the greatest impact (hazard ratio [HR] 2.6, standard error [SE] 0.09 relative to nonempowering advertisements), but we observed significant variations by gender. For instance, we found that men exposed to smoking cessation advertisements were less likely than women to subsequently conduct smoking cessation searches (HR 0.2, SE 0.07), but that this likelihood increased 3.5 times in men exposed to advertisements containing empowering content. Women were more influenced by advertisements that emphasized the health effects of smoking. We also found that appearing at the top right of the page (HR 2.1, SE 0.07) or at the bottom rather than the top of a list (HR 1.1, SE 0.02) can improve smoking cessation advertisements' effectiveness in prompting future searches related to smoking cessation. Conclusions: Advertising should be targeted to different demographic groups in ways that are not always intuitive. Our study provides a method for testing the effectiveness of Web-based antismoking advertisements and demonstrates the importance of advertisements that are tailored according to specific demographics.
Full-text available
This paper provides an overview of recent developments in big data in the context of biomedical and health informatics. It outlines the key characteristics of big data and how medical and health informatics, translational bioinformatics, sensor informatics and imaging informatics will benefit from an integrated approach of piecing together different aspects of personalized information from a diverse range of data sources, both structured and unstructured, covering genomics, proteomics, metabolomics, as well as imaging, clinical diagnosis, and long-term continuous physiological sensing of an individual. It is expected that recent advances in big data will expand our knowledge for testing new hypotheses about disease management, from diagnosis, to prevention to personalized treatment. The rise of big data, however, also raises challenges in terms of privacy, security, data ownership, data stewardship and governance. This paper discusses some of the existing activities and future opportunities related to big data for health, outlining some of the key underlying issues that need to be tackled.
Conference Paper
Full-text available
We study time-dependent patterns of information seeking about pregnancy, birth, and the first several weeks of caring for newborns via analyses of queries drawn from anonymized search engine logs. We show how we can detect and align web search behavior for a population of searchers with the natural clock of gestational physiology via proxies for ground truth based on searchers' self-report queries (e.g., [I am 30 weeks pregnant and my baby is moving a lot]). Then, we present a methodology for performing additional alignments, that are valuable for learning about the concerns, curiosities, and needs that arise over time with pregnancy and early parenting. Our findings have implications for learning about the temporal dynamics of pregnancy-related interests and concerns, and also for the design of systems that tailor their responses to point estimates of each searcher's current stage in pregnancy.
Policies are large-scale interventions that typically aim to influence behaviors and decision-making across entire populations to obtain a desired outcome. With the rapid increase in Big Data and its growing influence on policy, there is an emerging opportunity to produce meaningful and efficient mechanisms for improving public policy outcomes. However, there are still considerable gaps between existing theories in the behavioral sciences and evidence generated by Big Data, including the representation of key groups within the population. We outline the need for replicating established behavioral insights through Big Data that should coincide with clear ethical standards for implementing such approaches through evidence-based policymaking.
This chapter presents an overview of the recent literature on the persuasive effects of public communication campaigns. The scope of the review is substantial, ranging from traditional media to new technologies and from US settings to developing countries. The campaign topics primarily deal with health promotion, along with prosocial behavior and environmental reforms. The chapter examines key theoretical concepts, processes, and strategic guidelines, including campaign design, evaluation (formative, process and summative), types of effects (direct and indirect), messages (prevention vs. promotion vs. informational vs. persuasive, and appeals), message sources, mediated communication, and quantitative dissemination factors. The chapter then illustrates these guidelines with three campaign foci: drug use, smoking, and risky drinking.
This article describes an emergent logic of accumulation in the networked sphere, ‘surveillance capitalism,’ and considers its implications for ‘information civilization.’ The institutionalizing practices and operational assumptions of Google Inc. are the primary lens for this analysis as they are rendered in two recent articles authored by Google Chief Economist Hal Varian. Varian asserts four uses that follow from computer-mediated transactions: ‘data extraction and analysis,’ ‘new contractual forms due to better monitoring,’ ‘personalization and customization,’ and ‘continuous experiments.’ An examination of the nature and consequences of these uses sheds light on the implicit logic of surveillance capitalism and the global architecture of computer mediation upon which it depends. This architecture produces a distributed and largely uncontested new expression of power that I christen: ‘Big Other.’ It is constituted by unexpected and often illegible mechanisms of extraction, commodification, and control that effectively exile persons from their own behavior while producing new markets of behavioral prediction and modification. Surveillance capitalism challenges democratic norms and departs in key ways from the centuries-long evolution of market capitalism.
The author summarizes emerging standards for informed consent as the underpinning of ethical research in humans.
A randomized experiment with 1.6 million customers measures positive causal effects of online advertising for a major retailer. The advertising profitably increases purchases by 5%. 93% of the increase occurs in brick-and-mortar stores; 78% of the increase derives from consumers who never click the ads. Our large sample reaches the statistical frontier for measuring economically relevant effects. We improve econometric efficiency by supplementing our experimental variation with non-experimental variation caused by consumer browsing behavior. Our experiment provides a specification check for observational difference-in-differences and cross-sectional estimators; the latter exhibits a large negative bias three times the estimated experimental effect.