1. The document describes a publisher solutions company that helps publishers optimize visitor engagement, content delivery, and advertising revenue through four connected services: AudianceConnect, ContentConnect, YieldConnect, and DataConnect.
2. YieldConnect uses data passing and real-time bidding to optimize ad revenue by determining the winning bid based on factors like floor prices, previous user data, and minimizing burn risk to bidders.
3. DataConnect collects visitor data through the bidding process and across multiple visits, which is then used for audience targeting and attribution modeling by the publisher solutions company and advertisers.
The QSR Media Dispersion: Pre, Mid & Post Pandemic – By the Numbers
Putting the publisher in the quarterback spot
1. 2013 Publisher Solutions
ContentConnect:
Smart CSM and ETL We Help Publishers:
YieldConnect:
1. Bring in more visitors to…
AudianceConnect:
PUBLISHER A SSP 1
eCPM
(AudianceConnect)
D
AD Unit SSP 2
C • Early, Look
eCPM
P
C
• Early, Filtration S
SSP 3
2. Optimize content delivery
• Early, Data E
T
R
A
> R
V
eCPM
SSP 4
(ContentConnect)
F Dsad asdasd asd asdadsdas E eCPM
F
I
dasdasdasd asdas a asdas asd a
asd adas asdas ad asd asd ada
asfad dfdsfd sdfd sdf sdfds sd
R SSP 5 3. Optimize AD revenue
C
eCPM
(YieldConnect)
We use data to drive the
DataConnect: other three things.
• Actionable Visitor Data. (DataConnect)
• Reverse Re-Targeting.
2. 2013 Publisher Solutions
One Visitor PUBLISHER P
U
Entry = Google AD
B
“Football Scores” • Web
Unit A
D
• Mobile Phone
S
• Mobile Tablet E
R
• TV? V
E
R
3. 2013 Publisher Solutions
YieldConnect
STANDARD DATA PASS TO BIDDERS; GEO, IP, Browser, Referring URL, DMA, ETC…
One Visitor PUBLISHER S
P
U
Entry = Google AD
U T
P A
B Direct DSP DSP SSP SSP AdNet
“Football Scores” Unit E G
A
& 1 2 1 2
• Web R Floors RTB RTB HIST HIST FIXED
D
• Mobile Phone
S
• Mobile Tablet E
R
• TV? V
E
R
4. 2013 Publisher Solutions
YieldConnect
STANDARD DATA PASS TO BIDDERS; GEO, IP, Browser, Referring URL, DMA, ETC…
One Visitor PUBLISHER S
P
U
Entry = Google AD
U T
P A
B Direct DSP DSP SSP SSP AdNet
“Football Scores” Unit E G
A
& 1 2 1 2
• Web R Floors RTB RTB HIST HIST FIXED
D
• Mobile Phone
S
• Mobile Tablet E
R LAST LOOK
• TV? V
E
R
CALC EST EST
?? $15 $1 ~$2 ~$3 ~$2
$8 NIKE.COM FORD.COM
2nd Price Auction. Floors, Winner, Data Pass.
Auction / Choose Winner (Assume never seen before):
Determine ‘LAST LOOK’ bid entry based on data trading logic. (Other common terms; ‘Soft Floors’, ‘Dynamic Floors’, etc…)
• Look for previous user data. If yes evaluate rules on direct sold. If no proceed with price optimization and reverse re-targeting.
• If $14.99 = DSP 1 as winner, pays full $15, max rev, but has long term ‘burn’ risk to bidder relations.
• If nothing bid = DSP1 as winner, pays $3.01, min rev
• If $8 = DSP 1 as winner, pays $8.01, uplifted rev, bidder is remains under max and ‘burn’ risk is mitigated. SELECT AND EXECUTE AD.
5. 2013 Publisher Solutions
YieldConnect
STANDARD DATA PASS TO BIDDERS; GEO, IP, Browser, Referring URL, DMA, ETC… Messaging and Data: Publisher Owned
One Visitor PUBLISHER Data
S
P
U Post trade messaging and data activity
Entry = Google U T
B
• Send bidders auction results and logic string
AD P A Direct DSP DSP SSP SSP AdNet
“Football Scores” Unit E G
A
& 1 2 1 2
• Web R Floors RTB RTB HIST HIST FIXED
• Cookie visitor and append/add data:
D
• Mobile Phone
S
•
• (STANDARD PARAMS = TOP 300X250
Mobile Tablet E
R LAST LOOK
• TV? V
E
R DATE, TIME, GEO, ETC.)
• (SEARCH RETARGET = ‘NFL SCHEDUEL’)
CALC EST EST
?? $15 $1 ~$2 ~$3 ~$2
$8 NIKE.COM FORD.COM
• (PAGE CONTEXT = ‘FOOTBALL’S WEEK TOP
2nd Price Auction. Floors, Winner, Data Pass.
Auction / Choose Winner (Assume never seen before):
Determine ‘LAST LOOK’ bid entry based on data trading logic. (Other common terms; ‘Soft Floors’, ‘Dynamic Floors’, etc…)
• Look for previous user data. If yes evaluate rules on direct sold. If no proceed with price optimization and reverse re-targeting.
STORIES’)
• If $14.99 = DSP 1 as winner, pays full $15, max rev, but has long term ‘burn’ risk to bidder relations.
• If nothing bid = DSP1 as winner, pays $3.01, min rev
• (AD = HIGH PROB BRAND
• If $8 = DSP 1 as winner, pays $8.01, uplifted rev, bidder is remains under max and ‘burn’ risk is mitigated. SELECT AND EXECUTE AD.
RETARGET, SportsWear, RTB BID=$15, PAID CALC
2ND=$8, RTB2(Auto)=$1, SSP1=$3, SSP2=$2, NET=$
Messaging and Data:
Post trade messaging and data activity Publisher Owned 3)
• Send bidders auction results and logic string
• Cookie visitor and append/add data: Data • IF OTHER ADS ON THE PAGE SIMILAR DATA
• (STANDARD PARAMS = TOP 300X250 DATE, TIME, GEO, ETC.)
• (SEARCH RETARGET = ‘NFL SCHEDUEL’)
• (PAGE CONTEXT = ‘FOOTBALL’S WEEK TOP STORIES’) • And append that same data for each page of the user’s
• (AD = HIGH PROB BRAND RETARGET, NIKE, RTB BID=$15, PAID CALC 2ND=$8, RTB2(Ford)=$1, SSP1=$3, SSP2=$2,
NET=$3)
• IF OTHER ADS ON THE PAGE SIMILAR DATA
experience and for each return visit until the user clears
• And append that same data for each page of the user’s experience and for each return visit until the user clears cache.
cache.
• Other 1st and 3rd party data may be appended to.
• On winning transactions on or off network other data would be collected, including DR and CPA results.
• In the case of DR and CPA results a true value score would be applied to the individual, and subsequently used to identify look-a-
likes.
6. 2013 Publisher Solutions
YieldConnect
STANDARD DATA PASS TO BIDDERS; GEO, IP, Browser, Referring URL, DMA, ETC… Brand Sales
AlphaBird and/or Publisher find an
One Visitor PUBLISHER S
P advertiser who is interested in
U buying against the collected data.
Entry = Google AD
U T
P A
B Direct DSP DSP SSP SSP AdNet In an attribution model.
“Football Scores” Unit E G
A
& 1 2 1 2
• Web R Floors RTB RTB HIST HIST FIXED
D
• Mobile Phone
S
• Mobile Tablet E
R LAST LOOK
• TV? V
E
R
CALC EST EST
DPM
?? $15 $1 ~$2 ~$3 ~$2 Attribution Modeling
$8 NIKE.COM FORD.COM Data collection source is
determined and other data is
contributed:
• AB/Pubs data
2nd Price Auction. Floors, Winner, Data Pass.
Auction / Choose Winner (Assume never seen before):
Determine ‘LAST LOOK’ bid entry based on data trading logic. (Other common terms; ‘Soft Floors’, ‘Dynamic Floors’, etc…)
• Look for previous user data. If yes evaluate rules on direct sold. If no proceed with price optimization and reverse re-targeting.
• Buyers data
• Neilson data
• If $14.99 = DSP 1 as winner, pays full $15, max rev, but has long term ‘burn’ risk to bidder relations.
• If nothing bid = DSP1 as winner, pays $3.01, min rev In most cases this is where attribution
• If $8 = DSP 1 as winner, pays $8.01, uplifted rev, bidder is remains under max and ‘burn’ risk is mitigated. SELECT AND EXECUTE AD. modeling and predictive analytics
would occur.
As ads are displayed and actions
recorded all systems receive
Messaging and Data:
Post trade messaging and data activity Publisher Owned feedback loop data. Attributes
grow in depth, breadth, and begin
• Send bidders auction results and logic string
• Cookie visitor and append/add data: Data to achieve value.
Pricing and segments are modified
• (STANDARD PARAMS = TOP 300X250 DATE, TIME, GEO, ETC.) according to results.
• (SEARCH RETARGET = ‘NFL SCHEDUEL’)
• (PAGE CONTEXT = ‘FOOTBALL’S WEEK TOP STORIES’)
• (AD = HIGH PROB BRAND RETARGET, NIKE, RTB BID=$15, PAID CALC 2ND=$8, RTB2(Ford)=$1, SSP1=$3, SSP2=$2,
NET=$3)
• IF OTHER ADS ON THE PAGE SIMILAR DATA
• And append that same data for each page of the user’s experience and for each return visit until the user clears cache.
• Other 1st and 3rd party data may be appended to.
• On winning transactions on or off network other data would be collected, including DR and CPA results.
• In the case of DR and CPA results a true value score would be applied to the individual, and subsequently used to identify look-a-
likes.
7. 2013 Publisher Solutions
YieldConnect
STANDARD DATA PASS TO BIDDERS; GEO, IP, Browser, Referring URL, DMA, ETC… Brand Sales
AlphaBird and/or Publisher find an
One Visitor PUBLISHER S
P advertiser who is interested in
U buying against the collected data.
Entry = Google AD
U T
P A
B Direct DSP DSP SSP SSP AdNet In an attribution model.
“Football Scores” Unit E G
A
& 1 2 1 2
• Web R Floors RTB RTB HIST HIST FIXED
D
• Mobile Phone
S
• Mobile Tablet E
R LAST LOOK
• TV? V
E
R
CALC EST EST
DPM DSP /
?? $15 $1 ~$2 ~$3 ~$2 Attribution Modeling
$8 NIKE.COM FORD.COM Data collection source is
determined and other data is
contributed:
Bidder
(machine and attribution functions
• AB/Pubs data
2nd
Auction / Choose Winner (Assume never seen before):
Price Auction. Floors, Winner, Data Pass.
Determine ‘LAST LOOK’ bid entry based on data trading logic. (Other common terms; ‘Soft Floors’, ‘Dynamic Floors’, etc…)
• Look for previous user data. If yes evaluate rules on direct sold. If no proceed with price optimization and reverse re-targeting.
• Buyers data
• Neilson data
could exist here in some cases)
Bidder sets starter criteria and
• If $14.99 = DSP 1 as winner, pays full $15, max rev, but has long term ‘burn’ risk to bidder relations.
pricing and bids across many
• If nothing bid = DSP1 as winner, pays $3.01, min rev In most cases this is where attribution
publishers looking for these same
• If $8 = DSP 1 as winner, pays $8.01, uplifted rev, bidder is remains under max and ‘burn’ risk is mitigated. SELECT AND EXECUTE AD. modeling and predictive analytics
users. And users that have these
would occur.
same qualities, (look-a-likes).
As ads are displayed and actions
recorded all systems receive
Messaging and Data:
Post trade messaging and data activity Publisher Owned feedback loop data. Attributes
grow in depth, breadth, and begin
• Send bidders auction results and logic string
• Cookie visitor and append/add data: Data to achieve value.
Pricing and segments are modified
• (STANDARD PARAMS = TOP 300X250 DATE, TIME, GEO, ETC.) according to results.
• (SEARCH RETARGET = ‘NFL SCHEDUEL’)
• (PAGE CONTEXT = ‘FOOTBALL’S WEEK TOP STORIES’)
• (AD = HIGH PROB BRAND RETARGET, NIKE, RTB BID=$15, PAID CALC
2ND=$8, RTB2(Ford)=$1, SSP1=$3, SSP2=$2, NET=$3)
• IF OTHER ADS ON THE PAGE SIMILAR DATA
• And append that same data for each page of the user’s experience and for each return visit until the user clears cache.
• Other 1st and 3rd party data may be appended to.
• On winning transactions on or off network other data would be collected, including DR and CPA results.
• In the case of DR and CPA results a true value score would be applied to the individual, and subsequently used to identify look-a-
likes.
8. 2013 Publisher Solutions
$ $ $ $$
YieldConnect
Brand Sales $ $ $ $ $ $$
STANDARD DATA PASS TO BIDDERS; GEO, IP, Browser, Referring URL, DMA, ETC… $$ $ $
One Visitor PUBLISHER S
P
U
AlphaBird and/or Publisher find an
advertiser who is interested in
buying against the collected data.
$ $$$
$ $$ $ $ $
Entry = Google U T
$ $ $ $$
AD P A
B Direct DSP DSP SSP SSP AdNet In an attribution model.
“Football Scores” Unit E G
A
& 1 2 1 2
• Web
$ $$$ $
R
D Floors RTB RTB HIST HIST FIXED
$
• Mobile Phone
$
S
• Mobile Tablet E
$
R LAST LOOK
• TV? V
E
R
CALC EST EST
DPM DSP / Approved
?? $15 $1 ~$2 ~$3 ~$2 Attribution Modeling
$8 NIKE.COM FORD.COM Data collection source is
determined and other data is Bidder Off Network
contributed:
2nd
Auction / Choose Winner (Assume never seen before):
Price Auction. Floors, Winner, Data Pass.
Determine ‘LAST LOOK’ bid entry based on data trading logic. (Other common terms; ‘Soft Floors’, ‘Dynamic Floors’, etc…)
• AB/Pubs data
• Buyers data
(machine and attribution functions
could exist here in some cases) Publishers:
• Look for previous user data. If yes evaluate rules on direct sold. If no proceed with price optimization and reverse re-targeting. • Neilson data
• If $14.99 = DSP 1 as winner, pays full $15, max rev, but has long term ‘burn’ risk to bidder relations.
• If nothing bid = DSP1 as winner, pays $3.01, min rev In most cases this is where attribution
Bidder sets starter criteria and
pricing and bids across many
(Many)
publishers looking for these same
• If $8 = DSP 1 as winner, pays $8.01, uplifted rev, bidder is remains under max and ‘burn’ risk is mitigated. SELECT AND EXECUTE AD. modeling and predictive analytics
users. And users that have these
would occur. Each ad delivered would return data
same qualities, (look-a-likes).
similar what AB is capturing.
As ads are displayed and actions Would be missing auction results
recorded all systems receive unless AB was also the SSP for that
Messaging and Data:
Post trade messaging and data activity Publisher Owned feedback loop data. Attributes
grow in depth, breadth, and begin
pub.
• Send bidders auction results and logic string
• Cookie visitor and append/add data: Data to achieve value.
Pricing and segments are modified
Where an ad was served there
• (STANDARD PARAMS = TOP 300X250 DATE, TIME, GEO, ETC.) would also be returned to all
according to results. systems a record of
• (SEARCH RETARGET = ‘NFL SCHEDUEL’)
• (PAGE CONTEXT = ‘FOOTBALL’S WEEK TOP STORIES’) CLICK, DR, CPA type data. This
• (AD = HIGH PROB BRAND RETARGET, NIKE, RTB BID=$15, PAID CALC 2ND=$8, RTB2(Ford)=$1, SSP1=$3, SSP2=$2, type of data would then result in
NET=$3) true value scoring on the individual
• IF OTHER ADS ON THE PAGE SIMILAR DATA user. And would inform look-a-like
• And append that same data for each page of the user’s experience and for each return visit until the user clears cache. methods.
Justin Manes • Other 1st and 3rd party data may be appended to.
•
COO On winning transactions on or off network other data would be collected, including DR and CPA results.
• In the case of DR and CPA results a true value score would be applied to the individual, and subsequently used to identify look-a-
AlphaBird likes.
“Football Scores”