18. Data Economy
傳統-> 數位經濟學
Internet Economy
REACH High
RICHNESS
High
Low
Traditional Economy
(quality)
(quantity)
Attr. vs. behavior
Base of
targeting
22. Data-driven Performance
資料豐富度
(The power source of behavioral
forecasting)
Reach
Richness
High
High
Low
使用者接觸量(Reach of UU)
Range
High 使用者情境(The audience affiliate
of whole context)
Performance (CTR, CVR,
CPI)
of AdNet and DSP
Problem-solving
Thinking
Data
Algorith
ms
Tools
43. Internet as a mass media
“Half the money I spend on advertising is wasted;
the trouble is I don‘t know which half.”
-- John Wanamaker, ~ 1875
44. Current Challenges
How to identify?
Find the "best match" between a given user in a
given context and a suitable advertisement.
-- Dr. Andrei Broder and Dr. Vanja Josifovski, Standford University
Limited Info.
Budget? Creative?...
Bid Price?
45. Channels play different roles in the
customer journey
Source: http://www.thinkwithgoogle.com/
46. Advertiser Utility: The Value
Funnel
CPM campaign:
Revenue = N/1000 ⋅CPM
CPC campaign:
Revenue = N ⋅ CTR ⋅ CPC
CPA campaign:
Revenue = N ⋅ CTR ⋅ CVR⋅ CPA
52. Bid Landscape Forecasting
Only reference market price by base price
Imp, UU, Click, Conv.
DSP, per campaign, per targeting criteria…
the advertisers’ targeting profiles the winning
bid value
Pacing
Source: Ying Cui, Ruofei Zhang, Wei Li, Jianchang Mao(2011), Bid Landscape Forecasting in Online Ad Exchange Marketplace, Yahoo! Labs
53. Traffic forecasting
An impression on Jeremy Lin BBS post of MiuPTT
Two product ads
A: Linsanity T-Shirt
B: Baseketball shoes
Not optimized if only bid for highest price
B bid higher than A
Inventory A is much fewer than inventory B
55. Bid Landscape Forecasting
-- Bid Star Tree Expansion
*
*C
1
*
*C
2
*
B1
C1
*
B1
C2
*
*
*
*
*
*
*
*
C1
*
*
C2
*
B1
*
• Remove few-imp path for
not too sparse
• Easily to target all
Source: Ying Cui, Ruofei Zhang, Wei Li, Jianchang Mao(2011), Bid Landscape Forecasting in Online Ad Exchange Marketplace, Yahoo! Labs
Target Attribute
56. Bidding Price Calculation
Bidding Price = F (base price, CVR ) • In the same campaign, the conversion value is the
same
= base price * φ • φ = CVR / avg CVR
φ = p(c|u, i) / Ej [ p(c|u, j) ]
• I, j : inventory (on Web/App)
Ej [ p(c|u, j) ] = Σj p(c|u, j) p(j) = p(c|u)
φ = p(c|u, i) / p(c|u)
φ = p(c|s, i) / p(c|s)
• All inventories in the same segment are the same,
• i could also be inventory cluster
57. Bidding Price Model Building
• Cold start
• Training feature of segment and inventory respectively
• Do not train combined feature for preventing over fitting
on few training data φ = p(c|s, i) / p(c|s)
• Cold start
• Training feature of segment
DSP cross-campaigns:
bid = BasePrice(s, a) * p(c|s, I, a) / p(c|s, a)
AUC: the area under the ROC curve(TP/FP)
Lift: target response divided by average response.
58. Bidding Price Calibration
Forecasting
Sampling & learn
On-line adjustment
Feedback control & Re-learn
Loss reason
Prior Probability Shift
• Budget, Freq. cap, …
Competition
63. If only attach importance to quantify the business
model, it will not have the ability to find a potential
growth opportunities: "The pursuit of quantifying the
biggest problem is that people ignore the context of
the behavior generated, detached from the context of
the event, and have not been included in the model
ignores variables effectiveness. "
企業若只重視量化模式,
將無法擁有尋得潛在成長
契機的能力:「追求量化
最大的問題在於,忽略人
們產生行為的脈絡,把事
件從情境中抽離,且忽略
沒有被納入模式中的變數
效力。」
- Roger Martin
Rothman School of Management, Toronto
64. 3R:Reach+Richness+Range
大數據經濟學
資料豐富度
(The power source of behavioral forecasting)
Reach
Richness
High
High
Low
使用者接觸量(Reach of UU)
Range
High 使用者情境(The audience affiliate of whole context)
67. Takeaway ~
RTB, SSP, AdX, DSP
Data Scientist as CEO of Data
Big Data Pricing Engine
Scalable Big data infrastructure
Spark, Kafka, Docker, HDFS, Couchbase, …
Bidding Strategy & Design of Pricing Engine
Reach, Richness, Range
Reach:audience span, base of segmentation
Richness:relatedness(contribution) to conversion (target)
Range:affiliation with audience
Integrated, all media, full context engaging factors