3. 数据概论

 

Secondary
+ $↓ + time↓ + effort↓		- unknown accuray - ill fitting for the problem - outdated？ - control↓
Internal	Sales/patronage results (outcomes) Mkt activity (inputs) Cost info Distributor reports/feedback Customer feedback
External	Published data (Electronic/Printed)	Gov. Trade associations Periodicals Newspapers Books Annual reports Private studies
	Standardized sources (syndicated sources)	Store audits Consumer puchase panels Scanner data Single source data Media exposure data
	Internet
Primary
Non-experimental	Qualitative	FGI
	Quantitative	Survey
Experimental	Field
	Laboratory

Secondary 

Standardized Sources

-Motivation

• Pipeline effect in distribution channels: difficult for manufacturers to monitor sales because of retailer's behavior 

• Unknown competitor actions

• Unknown store behavior (i.e. promotion)

 

-Store Audits

• Sales=Beg Inv + Net Purchase from manufacturers or wholesalers - End Inv

• i.e. Nielsen retail index

• - Who? Freq? Brand switching? Repeat purchase?

 

-Consumer Purchase Panels

• Home audit: Net purchase = End Inv + Consumption - Beg Inv

  • NPD - Complete Kitchen Audit

• Mail diary

  • i.e. National Purchase Diary Panel (NPD), National Famliy Opinion (NFO)

• + can provide info on personal level (上面店铺审计不能提供的信息)

• - selection bias(对incentive敏感的人才参加), mortality effect(move, illness, refusal), testing effects(discard 1st month records)

 

-Scanner Data

• stated CB (survey) -> actual CB

• Retail level

  • i.e. ScanTrack Services (AC Nielsen), InfoScan (IRI)

  • -Who? Freq? Brand switching? Repeat purchase?

• Household level——Scanner Panel

  • i.e. Homesacan - AC Nielsen's Consumer Panel, InfoScan's Combined Outlet Consumer Panel (IRI)

 

-Single-source Systems

• data fusion 数据融合

  • scanner data+panel+TV measurement

  • ->buyer behavior+demographics+ promotional message exposure

  • ->causal relationships between sales and promotions

• in self-contained communities with their own newspaper and cable TV, and are roughly representative of the demographics of the country

• experiment-like setting, control of

  • ad. exposure in TV, newspapers

  • couponing

  • buyer behavior

• + exclusive pre-test records, immediate results, A/B test, control settings

• i.e. Behavior Scan (IRI), World Wide Consumer Panel Services (AC Nielsen)

 

-Media-related Standardized Sources

• TV: Nielsen TV index(NTI)

• Radio: Arbitron Diary Panel

• Print: RoperASW's Starch Readership Service

• Multiple Media: Simmons Study of Media and Markets

 

Primary

• Communication -> stated preference (versatility, speed, cost↓)

• Observation -> revealed preference (objectivity, accuracy)

• degree of structure↑ —> simple admin, simple analysis

• degree of disguise ↑ —> difficult interpretation