Programmatics 2 types:
ad inventory - we convert web page into money
through AD Generation
Lumascape - constantly changing
buy side - how to convert IO into cash - deliver
Outdoor advertising what was said and agreed uponwho landed on a
Billboard advertising page and if an ad wasn't shown, it expires. it
perishes quickly
combining advertising and marketing
Perishability - 100ms kailangan mabilis magload
yung website kasi may limit
Digital advertising
Digital made everything measurable and After the Dot Bomb
targetable
-not about cpm anymore
-analytics
data fuel ng malala
Go To (called Overture) - they develop the cpc
ROS
dinagdagan nila ng Relevance
Dot-com Era
it all started when the internet came along and
people started to realize its potential Perishabilty - there's 100ms to show an ad to a
user
93- AOL - Free but funded by ads
Ad CEN - company founded to address the
popularized CPM fragmentation of ad networks
95 - Nielsen for tv - no equivalent for digital -helped connect all ad network and get rid of
DoubleCLick was very first ad networks - to liquidity problems by puling all ad networks
measure together to create a central place where people
buy and sell
Ad Network
-siya yung cocontact sa madaming publishers
Price Discovery
they create relationship directly with publishers
and advertisers and will connect them with each -kailangan mo malaman kung ano exactly binibili
other mo
liquidity -when the buyer gives sufficient information for a
seller to know what the market demand looks like
how easily can something be converted into cash
-when the seller gives sufficient information so
that the buyer knows what they're buying
Auction
-sa ad exchange
ad networks are creating liquidity
ad exchanges make advertising like the stock -gathers information on audiences through online
markets or offline sources and compiles it in a single
platform
they create the most efficient price discovery
process in advertising -assists DSPs in providing a better understading of
consumers
Real Time Bidding
-refers to the buying and selling of online ad
Impressions through real time auctions that occur
in the time it takes a webpage to load. Those
auctions are often facilitated by ad chances or ssps Supply-Side Platforms
-type of programmatic advertising -host ad auctions for RTB, maximizing the price for
publishers
*some platforms let publishers sell their inventory
in advance for a fixed price (programmatic direct -pwede mag set ng price si publisher
or programmatic guaranteed) -a piece of software used to sell advertising in an
automated gashion
Demand-Side Platform (DSP) -used by online publishers to help them sell
display, video, and mobile ads
-toolbox you need to buy media
-goal is to maximize the prices their impressions
-getting the best price for media buyers sell at
platform that yu can buid on top of, something -publisher equivalent of
that can be extended with technology
-allow publishers to connect their inventory in
-allows advertisiers to buy impressions across a multi[;e ad exchanges, dsps and ad netowrks at
range of publisher sites, but targeted to specific once. this allows a huge range of potential buyers
users based on info such as their location and to purchase ad space - and for publishers can
browsing behavious maximize the revenue for their inventory
DSPs integrate thier system with ad exchanges and because of this, ssps are sometimes referred to as
other sources enabling publishers sto offer their YIELD-OPTIMIZATION PLATFORM
inventories to advertisers
-the price of these impressions is determined by a
real-time auction, through a process advantage of ssp
-enhanced -were created to combat the risk of driving the
value of inventory down
-they offer the ability for publishers to set "price
floors" which dictate the minimum prices for
which their inventory can sell to specific buyers, or
through specific channels
Data Management Platform (DMP) -ssps can also be used to dictate which advertisers
can and cant purchase inventory
-He;ps evaluate data, reporting, segmentation
-cant have a good dsp w/o a good DMP
YIELD-OPTIMIZATION PLATFORM -market efficiency refers to the degree to which
market prices reflect all available, relevant
“we will help publishers monetize their inventory information
by running an auction"
-as long as may competition exist
2 ways to become a yield optimizer:
-if the price is too high or too low, that wil be fixed
figure out unique ways to obfuscate the inventory out of market efficiency
you're selling
figure put ways to package and enrich
Advertisers has the most data
In any market there's more supply than demand,
DCO - Dynamic Creative Optimization the power is going to graviate towards the
-ginagawa ng SSP demand side
may ad personalization
however, the complexity is more on the BUY SIDE
in terms of dealing w/ data, dealing with targeting
Brand Safety and Ad Quality versus operating a sell side platform that's more
double verifier companies credible.
RTB
Marketplace weakness
Spot market
-transparency of data providers
-it can be costly to store permutations of diff types
of audience to target Deal ID
QPS – Queries for Seconds
Notes:
Quid Pro Quo of the internet – nothing on the
internet is free. The internet is ad-funded.
It's not important for the consumer and data Consumers get access to free content
center to be close together
Cookies
-"100ms" refers to the amount of time it takes for
a publisher to serve an ad to a single consumer -a single ID without any PII (personally Identifiable
when they click into a webpage Information)
-within the 100ms in RTB, millions of impressions PII – kaya ka niya matandaan
are prcessed simutaneously through the auction Cookies used to be able to store everything,
-DSP data centers should be close to exchange narealize nilang maliit lang yung space. place
data centers ti increase efficienct and reduce the horder na lang sila.
rate of timeouts
Cookie exchange table
Market efficiency
When an ad request comes over to a dsp, the dsp
doesn’t see the user’s browser. It doesn’t have
access to the cookies. To solve that every DSP and 6- Bid Factors
ad exchange has to do cookie mapping Line Item – create a permutation of all the
Each ad exchange calls the dsp with their identifier variables you can target on, and then assign a bid
for a user and the dsp then will assign an identifier and budget.
to that user -used to gleam insights on which line item is
Federated exchange going to perform best
Every player in any of the RTB aspects has their This would entail multiple items to be set up and
own cookie table and does their own optimization multiple categories and multiple retargeting
there approaches
There is still the opt-out option wherein the user You wull then chisel it down by sorting them by
chooses not to be tracked. performance and shutting down whatever line
items are not doing well
Exchange term, federating ad request to all by the
side-players *controlled variable – same data na di nagbabago
Federate – to join
Decisioning Bid Factors – takes an opposite approach to line
items
On the exchange side, the challenge is how to
create enough connections Each bid factor has a multiplier.
On the buy side, it’s the processing challenge If A is performing better than B, you can adjust
the A to say 1.5. So if your base bid is a dollow,
In about 7ms, the aggregating platform helps the you’re bidding $1.50 whenver there’s a variab;e
buyer buy the things that are best for them, of A.
manages the frequency, the budget and the
targeting. Pag di siya ganun nagpeperform, babaang ang bid
factor
-pag di kasya yung ad size, di pwede
You can make trillions of [ermutations in one ad
1:24 kapag napakitaan na yung website, di na din group
pwede
Bid and budget shouldn’t be combined together.
WEAKNESS If something works, it’s your bod that’s making it
work and therefore you might be inclined to
*transparency of data providers* increase the budget
In the course of a campaign, there might be six or *mas matagal, mas ibababa yung bid
seven companies you have to deal with.
Integration of these companies would be BID FACTOR = AVG PERFORMANCE / ACTUAL
important. PERFORMANCE
It can be costly to store permutations Para bumaba yung CPC
Data Centers & Consumers Bid and budget
It’s no use having dsp data center close to the There are 2 big factors that drive bidding
consuemer if it is not close to the exchange
Budget
Common data center locations are: East coast US,
west coast US, etc Runway
If there’s a huge budget, bid reducing can be Setting goals and achieving success
approached conservatively. If there’s a small
budget, you can cut stuff out instead of bid Reasons why campaigns don’t work
reducing. -they don’t get full credit for view through
conversions
PIXELS AND DATA COLLECTION -they have the same cpa goal for prospecting and
retatrgeting
Average person has over 20 digital identities
-they think all geo locations are created equal
The cookie works for only a specific domain
-they have a crowded media plan
Pag ibang domain, different data
Actionable Data:
During cookie mapping, the dozens of players in
RTB should be in the same page about who a user
is.
Need pixels to implement for cookie mapping
Floodlight tags – doubleclick container tags
iFrame tags – container for lots of different
piececs of codes
JavaScript tags – code that gets laid down directly
Pixels – image talaga siya, saka dapat image talaga
siya para magwork
Si javascript pag nilagay siya sa iframe, di na siya
makikita kasi nakacontainer na sya
MECHANICS OF COLLECTING DATA
Each company can read its own cookie when it
interacts with the browser. That browser needs ti
be handed off to other parties so that they can
read their own cookie, gtake that data and then
put it onto the url in clear txt si nobody can read it
*kung sino may pinakamadaming data, yun yung
mhga icoconnect lang
Tag types:
Conversion tags – nilalagay siya just to track
Retargeting tags – kumukuha ng data para
malaman if tatargettin ulit sya
Data import
TARGETING AND DATA MANAGEMENT 3RD PARTY DATA
RECENCY & FREQUENCY -nakukuha sa dmp
Recency is the most important variable in -get started with dmp
Retargeting
-put pixels in place
When was the data created
-get your first-party data in the dmp
-do lookalike modeling from pixel activity
DATA ANALYSIS
-look for keywords
-importante na isang beses lang ipapakita yung ad.
1/24 -browse by data type
Data, Data everywhere -test data elements through reports
To understand Quality is to understand the -use predictive audience targeting
SOURCE -exclude irrelevant data
Who is the source USE CASES
How is the data collected -conquesting: serve ads to your competitor’s
What action was taken to qualify customer
-cross pixel for custom retargeting: build a custom
How often is it refreshed
audience based on a custom list of keywords
The bigger the audience, the worse it’s going to bringing search and display targeting
perform
-onboard offline data
-turn first party data into prospecting data
Measures of quality
DMP Features
Reach > Day After Recall (DAR), vCE (validated
campaign essentials) -audience segmentation
Awareness > Brand impact study -lookalike modeling
Influence > BTR, MTA Data elements reporting
Sales Lift > Sales (DLX ROI) Third party audience data partnerships
Viewability > chance na Makita yung ad mo, Data alliance segments
percentage din ng ad mon a Nakita sa screen The data marketplace
Offline to online audience mapping
Pre-Data RFP
DON’T STOP TESTING
3X SHOULD HAVE 3X MORE THAN WHAT IS
EXPECTED
ADVANTAGES OF USING DMP 3) It’s a new era
Data management is added value, it’s an integral Hypotheses are now data driven
part of an effective DSP
2) Intersections of 1st and 3rd party data
DMP and DSP tools are tied together. All dsp
audiences and insights are immediately Finally matter
1) the word data is both cool and controversial
CREATING AN AUDIENCE The word “data” has been outlawed in a division
of comcast
-made up of included and excluded data groyps.
Data vriups are made up of data elements. Here’s
how it works
Targeting & data management
Tracking tag and click retargeting > 2 ways to CREATIVES
retarget
Every format has a diff goals
Dynamic creative optimization – kapag ang user
Contextual – infinite signal pumunta sa shopping ads, di nagpurchase,
pinakitaan ng specific product na chineck ni user.
Ito yon.
10 cool things and forecasting
Playable Ads
-Data and ads are the currency of the web
-user can play a game and can win something
-data and media have different supply level
Expandable ads
-pag clinick, lumalaki yung ads up to whole page
10) data is massively underutilized
-known as rich media
Only small amounts of data are used now
Video Ads
9) proving data’s accuracy and effectiveness is
easier using standards
8) pricing models continue to evolve TAGS
7) in mobile, data is driving What tags are and how to manage them
Data is moving to mobile Piece of code to track data
6) targeting and attribution Use of tags
Find power in location -retarget customers; retarget by categories
Where someone can tell you a lot -measure how customers got to the site
5) offline and online data unite -track what customers buy
4) measurement of ad effectiveness is the best it’s -product category
ever been -specific product-level category
The feedback loop is getting better
Type of tags -leading to growth in programmatic, building
confidence among advertisers
-pixel
Dynamic Parameters
Values in the tag that may pass more data about
the client and the product they’re interested in
Conditional fires
-getting only visibility into the conversions that the
advertiser’s ad server has determined that we
drove as opposed to full visibility on each and
every time that conversion pixel fired on the
website
-finifilter yung data
TROUBLESHOOTING TAGS
Use tools like fiddler or firebug or ghostery to
ensure that the pixel is placed correctly
Marketplace Quality
3 levels of Standards ng tradedesk
*a prevention, not cure’ approach to ad fraud
-Whiteops
*Increasing supply chain transparency
*Insuring brand safety
Always have blacklist
Inventory Quality Standards
Ads.txt – para siyang certification in a text format
Partnerships
Identification
-Media buyers need more choice when it comes to
data and inventory
-as a result, this will increase efficiemcy industry-
wide
