"ery
G
2.7 PROGRAMMATIC DIGITAL ADVERTISIN
Vertising is automat i es place in rea
Togrammatic ad vit ¢ buying takes pla
ti ising is automated ad buying. The buying
UeTammatic advertising isa
™m Ti 'B) model. Bidding here
del. H Wis called Real Time Bidding (RTB) model, ie
adel Hence, ea a
Publisher's w ebpages
tite 40 fot,
M8 a
takes place for aq 4
Xba
a t related to automation j
Programmatic advertising is the most recent developmen
adverusing. In the US, about 82
tm Figure 2.30,
Nn the ¢
ent of the digital ad spending is programmatic ad spend,
per c
Programmatic Ad Spending US, 2014-1018
82% %
78% te
7 73% 20%,
i s 70%,
30 7
55% $ oon
25 atuey 50%
20 $ el 40%
IvaLus 5
18 s 308
10 VALUE 20%,
s .
5 |ivatu x e
: 2014 2015 2016 2017 2018
C1 Programmatic digital Oisplay ad spending (billions)
~#= petwentage of total digit splay ad spending
Source: eMarketer, June 2016
Ficure 2,39 Programmatic ad Spending Us, 20142018
Courtesy: eMarketer, used with permission
Programmatic by:
US assume that y
Team-BHP. 4)
have visited
Ying’ Sivnifies the shift from
OU are a car brand, Contexty
Ulocar India, ete Instead,
ar portal at |
month and booked :
Contextual
‘Atgeting to Mdien
al adverti
ME means
CU target Audion
veting. For evan
i setals suche
Advertising On car portal
advertiser
Mn the last
.
Who may be defined as thee
east once OnE Month, visited finance si vist once in the last
nate! ticket online at teay once
in the last one month, Tig User is
From
® more likely to yy ~
ser
To
—>
ees the 19 audiences
ad. This is called audience ATECLINE Which is hase wa ps ( proxy
on online behaviour, i inf INEM OF the user ant ‘wudoncon Vv wees
targets based on inten buying Figure 34 shows
Programmatic buying
. _-
Ficune 231 Programmatic buying
Scanned with CamScannerpipe averse ED
Programmatic buying reaches the right au
cence at the right time and place at the right price
with the right message to optimise the campaign
and deliver higher ROM (Figure 2.32).
27.1 Process of Programmatic
Advertising
The user requests for a webpage, which means
there is potential to show ad impressions to the
user. Every time a user requests for a page, ad
exchange holds a live auction. On behalf of the
publisher. Supply Side Platforms (SSP) send a
bid request to the Demand Side Platform (DSP)
DSP examines the potential impression in terms FIGURE 2.32 Five key inputs for programmatic buying”
of target audience and context match and retums
4 bid amount based on its value to the advertiser. Winning bid is chosen by the ad exchange and that ad is
shown on the publisher's website. Figure 2.33 shows the working of a programmatic advertising process’.
‘Abid request
‘Third Party Ad Server sent to each DSP
Five Key
Inputs for
Programmatic
2—~ pa exchange
hholds an auction for each
impression
ee]
—_
‘Ad creative is
hosted by third party
server or DSP
shown on publisher
website
5 winning ba
is chosen by Ad Exchange
“Advertisers
DSP enables advertisers to
Publisher
each cual er impression | 4 OsP evatusies each Publisher (websito) makes ad
based on what ic foerm about | impression and roturns a bid amount based Tnwentory avaliable on ad
ae santas la
Figure 2.33 Process of programmatic advertising’
From the process, itis clear that there are many players in the ad) 1g ecosystem, Let ts understand
their role and the value they add in more detail.
Souree: Adapted fromhtips:J/radiumone.com/p-contenoploads/201 6/08/ Whit
om March 20, 2017,
°" Source: Adapted from utps:/thlogs.adobe.com/digital marketing/advertising/|S-years-of-defiaing-programuma
Maren 20, 2017
Paper-Programinatic-ocket-Guide pal Retrieved
(, Reieval on
Scanned with CamScannerEBD disinrrt0re0000
2.7.2 Players in Programmatic Adve
‘consumer
Drie Partes
—— Ll
@ @ io @ @ oO
averse ose | | exenange ee] frown | publsher
—
@ —_aasgemes
Froune 234 Ropers in proprammarcodvering ecosyten™
Publishers
Publishers have ad impressions through which dhey monetise the site afd it is a mg
for them as most ofthe cont India are fre an hence do not have subscript
imagen, Reuicved on March 20,2017
represented by trading
adience and
Scanned with CamScannerpuphvAderiing ED
Table 2.8 Holding Company Trade desk mapping”™
c ae :
‘wer MEDIACON axis |
| Publics RAZORFISH, DIGITAS. ‘VIVAKI ‘]
meson Gan om, rp aeccueN |
iad UNL INITIATIVE MAGNA GLOBAL |
| carat ANNET
Havas Worldwide HAVAS MEDIA AFIPERE |
pressions for as
ofthe webpage, such
stop, Keswords, page URL and the p
behaviour with the DSP and request
(OpenX. Rubicon Project, Double
Taleo shases the cookie ID withthe DSP.
inthe browser ofthe usc. A cookie i @ small
pase URL, browser and deviee information. Wt enbances user experience as when the user visits the webpage
un cookie willbe tettieved by the webpage and it wil know whether th use 83 First time visitor ora
repeat visitor. what items were placed in the shopping exe ele. These cookies ate persistent and are valid for
‘eenai ine. These ae enerypted and can be readonly bythe domain which dropped them.
Demand Side Platform (DSP)
|A Demand-Side Platform (DSP) allows the advertiser to buy the inventory from various ad exchanges and
Iades gs
bid sate,
through DSPs, Some ofthe DSP companies are Double Click for Advenisers, Rocket Fuel, Tura, Appnenus,
‘Meuiamath, Double click, and Audience science.
DSP wets the request for bid from SSP w
webpage. DSPs have audiences which are prima
Since DSPs have good understanding of eonsurnet behaviout, they ete
ass, avid investors based on their online behaviow
wusines sites and stock exchange si
Scanned with CamScannerScanned with CamScannerowned ty big e-commerce a 3 1s AT&T. Third
arty data is From data provides
‘21 Panty ata
Dat Creaion
1 Panty at
Customer Onto
Daialeg,
‘excae
(orate
‘usa
posionarate
3" Paty Data
‘ata Power
FicuRe 2.36 Types of dot suppliers
1y companies that hasts your a and give
tsplay oF ads. tis 8
a know for
these a servers have yoo tec
SPS and publisher sits for slay of as.
Scanned with CamScannerme $$.
EBD vistarrtarrerng
hramically by the creative designe from his laptop.
email or other manual coordination, The creative must be prepare a
: The ereative must be ere heading, call 0 action but
and xml format for child te. The parent file i the template tat identifies where heating, call oaton bu vn
OF image and description will come. The child file is the data that gets pulled nn time and set rules ff
format. Hence, the ereative designer ean update the data inthe child file dynamically any q
customising the ad
th , 1s. Ad serv
Hence, while media optimisation is done by DSP, creative optimisation is done by Ee Tale a
charge a small per cent of media spend for the fees. Unfortunately only 40 per cent adve
third party ad servers as they consider it as cost
ultiple versions through
1 having to upload m v6 \
aaa mevo ile formals: WF file for parent file
Ad Exchange 4
‘Ad exchange brings together buyers and sellers to buy and sell online ad inventory in an automated mane
ike @stock exchange. tea ime auctioning forad spaces on publishers’ websites happens in an ad exchange
Further, ad exchanges are of two types. They are described inthis section.
Open Exchanges Open exchanges, as the name suggests, are ad exchanges open for everybody on the.
web. There is no restriction on the purchase of inventory available in these exchanges. They are run by tech,
Giants such as Google, Yahoo!, OpenX. etc. Over 70 billion ad impressions flow through these exchanges
faily. These volumes have mage policing the genuineness ofthese impressions extremely difficult and this
has given rise to bogus and low quality impressions. Open exchanges allow buyers to purchase media on
an audience basis through cookie-based targeting; achieving scale is usually a higher priority than running
campaigns with specific publishers
Private Exchanges The problem mentioned above led to the ereation of private exchanges wherein a
2roup of publishers pooled up to sell their advertising inventory only through select agencies or to reputed
Companies. This type of sclective selling has helped prevent fraudulent activities as seen in open exchanges,
Also, these exchanges are considered to be premium exchanges as advertisers get better ad placements,
insights, etc., as compared to open exchanges,
Real Time Bidding
Real time bidding (RTB) includes any transaction executed on an individual impression basis and in which
Pricing is determined in realtime. It works in open auctions as well as private auctions,
2.8 ANALYTICS TOOLS
There are some challenges plaguing the display advertising industry, Fortunately some ‘companies have
devised tools to overcome those challenges. Let us look at some ofthese issues and futuristic solutions
2.8.1 Viewability
Per Integral Ad Science & Media Rating Council Data in 2015, 57 percent of display ads were not viewable.
Ad may not be visible due to disruptive human behaviour, such as action taken before ad loads and renders,
multitasking or because of low-quality or compromised pages. It could be because ad loads in area out of the
user's browser, oF frequent page refresh prevents ad load, or web load gives errors,
It should be noted that 50 per cent of ad pixels should be displayed for a minimum of one second in image
ads and two seconds for video ad for an ad to be considered viewable. This condition has heen peserned ty
Media Rating Council in May 2015. Both the pixel and time components must be Measured and must satisfy
the above minimum requirements for an ad to be counted as a viewable impression,
Scanned with CamScannerpieptay Advertising ED
If both the pixel and nse components associated with an acl are greater than zero, but one or both do not
eet or exeeed the above minimum requirements, head mayb a "loaded ad’ It should not be
fesveiated With the {erm “impression
Toovercome this viewabihity issue, tools such as Integral Ad Seience and MOAT provide metrics such as
in-view impressions, oUL-of-view impressions, viewable rate
ferred thy
2.8.2 On Target Reach
The current cookie based methods of targeting overestimate reach and underestimate frequency by a huge
margin. The demographics accuracy is also low. Some tools such as *“Atlas’ is used to measure On-target
reach publisher wise for each campaign, It is based on logged in FB data, Nielsen DAR is another tool, They
do cross device measurement and provide reach over!
lap reports, These tools determine accuracy based on
real people logins across mobile and desktop. They enable people-based marketing and understanding of the
gonsumer joumey.
2.8.3 Ad Fraud
Ad fraud isthe deliberate practice of serving ads that have no potential to be viewed by a human. Perpetrators
of ad fraud are criminal gangs and the problem is costing the ad industry a lot of money. Publishers with
higher cost-per-thousand (CPM) rates are more susceptible to bots, which x
millions of ad impressions that are seen by no one but often get charged to marketers as a viewed promotion.
1, Domain Identity Theft This is a form of fraud where fraudsters add a fake URL into the ad unit,
fooling the auction into believing that an ad will appear on an acceptable site that it will not end up
on.
Click/Impression/Conversion Fraud This is when fraudsters mechanise clicking, viewing or
converting on ads. Click farms exist where people are paid to sit and click, view, or convert on ads
through desktop and mobile devices.
Bot Activity This is when robots carry out similar actions by masking IP addresses through
advanced algorithms. Malware can be installed on computers, giving fraudsters the power to control
what domains a computer visits and the ads it views
4. Adestacking This is when multiple ads are served within the same placement, This method tricks ad
servers, as each of the ads is served; however, they are stacked beneath each other so the advertiser's
message is often not viewable.
5. Pixel Stuffing when an ad of say 300 x 250 pixels is crammed into 1 x 1 pixel size which makes it
appear on page but is invisible to human eye.
enerally create millions upon
2.8.4 Brand Health
The process and controls implemented to protect advertisers from having their ads appear alongside content
that could damage their brand or reputation (even if by accident), Examples of brand safety issues are:
1. Ads appearing on sites publishing illegal content. For example, adult, illegal downloads
2. Ads appearing alongside user generated content, For example, controversial opinion
3. Ads presented in a negative context due to content on the site. For example, flights alongside an air
disaster,
Brand safety is ensured by blacklisting websites having controversial content; using keywords to not
Show ads in that context and by labeling content for adult use, etc, Integral Ad Science tool assigns brand risk
Scores to prevent brand health from gelting compromised.
Scanned with CamScanner