Intuit relies on deep analysis to measure and optimize online campaign effectiveness, improve user experience, and increase customer retention. However, it was spending millions of dollars on multichannel marketing campaigns without knowing which elements of those campaigns were contributing to the bottom line. By analyzing cookie and website data, Intuit was able to conduct better ad targeting and price/volume analysis, understand customer behavior, customize offerings, increase customer satisfaction, and maximize revenue.
Brad Terrell, VP & General Manager, Digital Media at Netezza, and Arup Ray, Director, Data Warehousing & Business Intelligence Development and Architecture at Intuit, present on:
- Using advanced analytics to drive better decisions for multichannel marketing campaigns.
- Learning the benefits of more precise, targeted marketing.
- Discovering where customers are in the ad funnel and how to target them appropriately.
14. The Segmentation Problem
High
Age: 18-24
Income: High Age: 25-35
Recency: Low Income: High
Income
Recency: High
Recency of visits to
High
YourWebSite.com
Age: 50-65
Income: Low
Recency: Low
Age: 36-49
Income: Low
High Recency: High
17. The Bidding Problem
Price/Volume Forecast Market X: Price/Volume Curve
Curves Show the Optimal
1,200,000
IMPRESSIONS PER DAY
1,000,000
Bid in Each Market: 800,000
600,000
max((R − B) ⋅ N(B))
400,000
200,000
0
0.00 1.00 2.00 3.00 4.00 5.00
R = revenue derived from showing ad CPM Bid
B = ad’s bid
N(B) = impression volume for bid B as Market X: Profit Curve
predicted by P/V Curve 3,500
3,000
2,500
PROFIT
2,000
1,500
1,000
500
0
0.00 1.00 2.00 3.00 4.00 5.00
CPM Bid
18. Customer Insight Through Deeper Analysis
(The Fastest Climbers Win)
Optimization
Predictive
Analytics
BI Reporting and
Ad-Hoc Analysis
• What is the
best choice?
• What will happen?
• What will the impact be?
• What happened?
• When and where?
• How much?
20. Intuit Company Information
Who We Are …
A leading provider of business and
financial management solutions
Company Founded: 1983
Date Public: 1993
Stock Symbol: NASDAQ: INTU
GAAP Net Revenue (Fiscal 2010): $3.2 billion
Employees: ~ 8000
Locations: 60 locations in 40 cities around the globe
20
23. Life stage of customer acquisition
Pre- Post-
Awareness
Conversion Conversion
First impression
Authentication
Conversion
23
24. What is working?
• Which channels are yielding results? Which sub-channels?
• Are channels cannibalizing each other?
• What works at what stage of the acquisition funnel?
24
25. High Level Goals of Intuit Marketers
• Optimize spend across
channels
• Coordinate spend across
channels
• Optimize messaging in
ads
• Use early converter
behavior for mid-course
adjustments
26. Where is the data to help me?
• Doubleclick tracks all converters and non-converters across
web sites as well as on key pages of our web site
• Omniture tracks customer behavior on our site
• These data sets are warehoused in our Customer Experience
Data Warehouse running on Netezza
• Massive volumes of data – the impression table alone has
more than 40 Billion rows!
26
27. What does the data tell me?
• A lot …
– About the effectiveness of the acquisition channel as well as specific
campaigns
– About channel interferences
– About product issues
– About customer behavior with regards to acquisition campaigns
27
32. Portal Home Pages – Distinct Visitors
0131 MSN HP Money Module Txtlink
0201 MSN HP Money Module Txtlink
0131 MSN HP
0201 MSN HP
0125 MSN HP
0215 MSN HP
0214 MSN HP
0131 Yahoo HP RM
0207 Yahoo Marketplace - Logo + text
0124 MSN HP
0124 MSN HP Money Module Txtlink
0201 AOL RM - User Initiated
0131 AOL HP RM - Host
0201 AOL HP RM - Host
0131 AOL HP Textlink
0201 AOL HP Textlink
0124 AOL HP Textlink
0125 AOL HP
0221 AOL HP - Textlink
0125 AOL HP Textlink
0124 AOL HP
0221 AOL HP 300x250
0201 AOL HP NB
0131 AOL HP NB
AOL HP 01/10
0 1000000 2000000 3000000 4000000 5000000 6000000 7000000 8000000
32
33. Portal Home Pages - Impressions
0131 MSN HP Money Module Txtlink
0201 MSN HP Money Module Txtlink
0131 MSN HP
0201 MSN HP
0125 MSN HP
0215 MSN HP
0214 MSN HP
0131 Yahoo HP RM
0207 Yahoo Marketplace - Logo + text
0124 MSN HP
0124 MSN HP Money Module Txtlink
0201 AOL RM - User Initiated
0131 AOL HP RM - Host
0201 AOL HP RM - Host
0131 AOL HP Textlink
0201 AOL HP Textlink
0124 AOL HP Textlink
0125 AOL HP
0221 AOL HP - Textlink
0125 AOL HP Textlink
0124 AOL HP
0221 AOL HP 300x250
0201 AOL HP NB
0131 AOL HP NB
AOL HP 01/10
0 10000000 20000000 30000000 40000000 50000000
33
34. Portal Home Pages – Distinct Converters
0131 MSN HP Money Module Txtlink
0201 MSN HP Money Module Txtlink
0131 MSN HP
0201 MSN HP
0125 MSN HP
0215 MSN HP
0214 MSN HP
0131 Yahoo HP RM
0207 Yahoo Marketplace - Logo + text
0124 MSN HP
0124 MSN HP Money Module Txtlink
0201 AOL RM - User Initiated
0131 AOL HP RM - Host
0201 AOL HP RM - Host
0131 AOL HP Textlink
0201 AOL HP Textlink
0124 AOL HP Textlink
0125 AOL HP
0221 AOL HP - Textlink
0125 AOL HP Textlink
0124 AOL HP
0221 AOL HP 300x250
0201 AOL HP NB
0131 AOL HP NB
AOL HP 01/10
0 50000 100000 150000 200000 250000 300000 350000
34
35. Portal Home Pages – Distinct converters
per Impression
0131 MSN HP Money Module Txtlink
0201 MSN HP Money Module Txtlink
0131 MSN HP
0201 MSN HP
0125 MSN HP
0215 MSN HP
0214 MSN HP
0131 Yahoo HP RM
0207 Yahoo Marketplace - Logo + text
0124 MSN HP
0124 MSN HP Money Module Txtlink
0201 AOL RM - User Initiated
0131 AOL HP RM - Host
0201 AOL HP RM - Host
0131 AOL HP Textlink
0201 AOL HP Textlink
0124 AOL HP Textlink
0125 AOL HP
0221 AOL HP - Textlink
0125 AOL HP Textlink
0124 AOL HP
0221 AOL HP 300x250
0201 AOL HP NB
0131 AOL HP NB
AOL HP 01/10
0.00% 0.50% 1.00% 1.50% 2.00% 2.50% 3.00%
35
36. Portal Home Page – Impressions per
distinct converter
0131 MSN HP Money Module Txtlink
0201 MSN HP Money Module Txtlink
0131 MSN HP
0201 MSN HP
0125 MSN HP
0215 MSN HP
0214 MSN HP
0131 Yahoo HP RM
0207 Yahoo Marketplace - Logo + text
0124 MSN HP
0124 MSN HP Money Module Txtlink
0201 AOL RM - User Initiated
0131 AOL HP RM - Host
0201 AOL HP RM - Host Good value?
0131 AOL HP Textlink
0201 AOL HP Textlink
0124 AOL HP Textlink
0125 AOL HP
0221 AOL HP - Textlink
0125 AOL HP Textlink
0124 AOL HP
0221 AOL HP 300x250
0201 AOL HP NB
0131 AOL HP NB
AOL HP 01/10
0 20 40 60 80 100 120 140 160 180 200
36
38. Paid Search Clicks and Affiliate Clicks
Paid Search does not currently interact significantly with Affiliate
3%
No Affiliate click
before
At least one
Affiliate click before
97%
15%
No Paid Search click
before
85% At least one Paid
Search click before
38
39. Paid Search Clicks and Affiliate Clicks
Paid Search does not currently interact significantly with Affiliate
Number Paid Search click alone
of Affiliate click alone
cookies
Paid Search click followed by Affiliate click
with…
Affiliate click followed by Paid Search click
0 1000000 2000000 3000000 4000000 5000000 6000000
Paid Search click alone
Conversion
rate of Affiliate click alone
cookies Paid Search click followed by Affiliate click
with…
Affiliate click followed by Paid Search click
0% 2% 4% 6% 8% 10% 12% 14% 16% 18% 20%
41. Why do people not convert?
Which are the
problem pages?
Create alternate flows
and do A/B testing
Pre- Post-
Awareness
Conversion Conversion
Authentication
Conversion
impression
First
41
43. Converted on Display Click…
People who convert on display click tend to stick with display clicks…
Awareness Stage Navigation Stage
Email Click
Email Click
Email Impression 1.46 Email Impression 1.78
Affiliate Click 1.92 2.40
1.41 Affiliate Click
1.69
Paid Search Click 1.43
Paid Search Click 1.96
1.39 1.92
Display Click 52.0 Display Click 56.5
Display Impression Display Impression
0% 50% 100% 0% 20% 40% 60% 80% 100%
Average number of exposures,
Percentage of converters
for converters who had at least one exposure
who had at least one exposure
43
48. Messages that are funnel aware
• Based on where the user is in the funnel we can target them
with relevant messages
• This could be a better offer if they are in awareness stage or a
reinforcement of value proposition if they are in navigation stage
• Doubleclick provides reach beyond the confines of the Intuit
websites
• The impressions in post-conversion stage are used to cross-sell
other products and hence are not wasted
48
49. Keys to success
• Ability to load large data volumes daily
• Complex queries that process billions of rows in minutes and
not hours
• Ability to leverage intelligent messaging mechanisms within
Intuit website and Doubleclick for intelligent messages across
the web
Data Insight Action
49