Copyright 2011 AiMark Is the classic purchase funnel dead? Sales impact of classic and new funnels Koen Pauwels @ Ozyegin University Joris Merks @ Google AiMark Media Efficiency lab project
Mar 31, 2015
Copyright 2011 AiMark
Is the classic purchase funnel dead?
Sales impact of classic and new funnels
Koen Pauwels @ Ozyegin UniversityJoris Merks @ Google
AiMark Media Efficiency lab project
Copyright 2011 AiMarkCopyright© 2012 AiMark
Classic funnel or new decision journey?
Copyright 2011 AiMarkCopyright© 2012 AiMark
Why does this matter to Business?
• “Walmart made a common mistake repeated hundreds of times in the business world: they relied on what customers said in a survey versus what they actually do in the stores (and online)”, Phil Terry, Founder of The Councils (2011)
• “You need to capture a customer and keep them engaged more than ever. The funnel doesn't account for this in a robust way (…) i.e. people talking about your brand during their own exploration, driving awareness and preference at what would be mid-points to the funnel.” (Cooperstein 2011).
• “As marketers we still need to effectively allocate dollars across multiple touchpoints as not everyone in every industry is living/engaging full in the digital space” Camille (2011)
Copyright 2011 AiMarkCopyright© 2012 AiMark
With Dutch Google, GfK & Alterian we ask:
1) How much do classic (offline survey-based) and new (online behavior) funnel metrics predict brand sales?
2) Does it differ in high vs. low-involvement categories?3) How can managers affect funnel metrics that matter
most to weekly sales?• We study 36 brands in 15 categories: durables (cars),
services (internet, energy, travel, lodging, insurance), food (cheese, butter, snacks, candy, dairy, soft drinks, beer) and non-food (tissue, napkins) packaged goods
Copyright 2011 AiMark
Dynamic system of Sales, Marketing, Classic & New Funnel
Accounts for short and long
term effectsVolume sales
Environment (temperature, economy)
Marketing
Price Promotion
Distribution Advertising
Classic funnel
Loyalty
Preference
Awareness
Consideration
Intention
New funnel Website
Search
Social media
Clicks
Copyright 2011 AiMarkCopyright© 2012 AiMark
Empirical Findings
1) New funnel explains sales beyond classic funnel 2) Classic funnel often driven by new funnel metrics3) Among classic funnel metrics, awareness most
important, before consideration and preference4) Classic funnel predicts best for FMCGs5) New funnel metrics are key in high-involvement
categories, such as durables and services 6) After online marketing, TV most powerful in driving
new funnel, outperforms online in engagement
Copyright 2011 AiMarkCopyright© 2012 AiMark
Adjusted R2 similar across funnels in FMCGs, much better for the new funnel in durables and services
FMCG Durables&Services0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
26%
47%
24%
39%
25%
47%
New onlyClassic onlyHybrid
Copyright 2011 AiMarkCopyright© 2012 AiMark
Out-of-sample Forecast Error lower for classic funnel in FMCGs, for new funnel in durables and services
FMCG Durables&Services0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.260.23
0.19
0.33
0.26
0.37
New onlyClassic onlyHybrid
Copyright 2011 AiMarkCopyright© 2012 AiMark
Awareness matters more to sales than preference
Awareness Consider Preference Loyalty0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.29
0.05
0.09
0.17
How to read? If awareness increases by 1%, volumes
sales increase on average
by .29%
Copyright 2011 AiMarkCopyright© 2012 AiMark
Clicks
Branded search
Generic se
arch
#Web visits
#Pageviews per v
isitor
#Pos web co
nversations
#Neg web conversa
tions0.000.100.200.300.400.500.600.700.800.901.00
0.17 0.17 0.160.25 0.23
0.67
0.19
0.93
0.66
0.470.38
0.00
0.260.16
TV Online
How to read? If TV expenditures increase by 1%, # website visits
increase on average by .25%
Online, but also TV ads drive new funnel metrics
Copyright 2011 AiMarkCopyright© 2012 AiMark
How do our findings help managers?• Should I replace classic funnel with new funnel tracking? 1) Cost-benefit trade-off: while classic funnel metrics are ~ 5x as
expensive, adding them to new funnel improves out-of-sample prediction with 27% for low-involvement products
2) Cross-funnel effects imply overlap, but also help you track how campaigns move between digital to physical settings: note the high impact of TV on online engagement
• For my specific brand in my specific category?New funnel key for high-involvement categories, but best combination
‘hybrid’ funnel differs by brand and categoryA high-value tracker for small-budget brands could combine classic
awareness with new funnel metrics
• How can I leverage the real-time value of online behavior?
Copyright 2011 AiMarkCopyright© 2012 AiMark
The multi armed bandit principle
Copyright 2011 AiMarkCopyright© 2012 AiMark
“Cheap offer now”-mentality
“Branding” = “Indirect conversion”
Real time feedback loopfor long(er) term objectives
Copyright 2011 AiMarkCopyright© 2012 AiMark
Future of marketing: Feel The Force
1. Validate predictive valueof new & classic funnel
2. Define the best funnel for your brand
3. Connect with The Force: real-time consumer response
4. Launch & Iterate
Copyright 2011 AiMark
THANK YOU!!
QUESTIONS ?