Consumer Intelligence Analytics Workshop

> Beyond Analysis <
Turning data into actionable
insights that boost ROI

> Workshop overview
§ Turning data into insights
– Metrics framework
– Data governance
§ Turning insights into action
– Oranisational structure
– Combining data sources
– Optimising the funnel
– Testing & optimisation
– Measuring success
August 2013 © Datalicious Pty Ltd 2

ACME Corp
Metrics framework

Awareness Interest Desire Action Satisfaction
> AIDA and AIDAS formulas
Social media
New media
Old media

> Importance of social media
Search
WOM, blogs, reviews,
ratings, communities,
social networks, photo
sharing, video sharing
August 2013 © Datalicious Pty Ltd
Promotion
8
Company Consumer

Reach
(Awareness)
Engagement
(Interest & Desire)
Conversion
(Action)
+Buzz
(Delight)
> Simplified AIDAS funnel

People
reached
People
engaged
People
converted
People
delighted
> Marketing is about people
40% 10% 1%

People
reached
People
engaged
People
converted
People
delighted
> Standardised roll-up metrics
Unique browsers,
search impressions,
TV circulation, etc
Unique visitors,
site engagements,
video views, etc
Online sales,
online leads, store
locator searches, etc
Facebook
comments, Tweets,
ratings, support calls, etc
Response rate,
Search response rate,
TV response rate, etc
Conversion rate,
engagement rate,
checkout rate, etc
10%40% 1%
Review rate,
rating rate, comment
rate, NPS rate, etc

People
reached
People
engaged
People
converted
People
delighted
> Provide context with figures
40% 10% 1%
New prospects vs. existing customers
Brand vs. direct response campaign

Exercise: Providing context

> Exercise: Providing context
§ Brand vs. direct response campaign
§ New prospects vs. existing customers
§ Competitive activity, i.e. none, a lot, etc
§ Market share, i.e. small, medium, large, et
§ Segments, i.e. age, location, influence, etc
§ Channels, i.e. search, display, social, etc
§ Campaigns, i.e. this/last week, month, year, etc
§ Products and brands, i.e. iphone, htc, etc
§ Offers, i.e. free minutes, free handset, etc
§ Devices, i.e. home, office, mobile, tablet, etc

> Conversion funnel 1.0
August 2013
Conversion funnel
Product page, add to shopping cart, view shopping cart,
cart checkout, payment details, shipping information,
order confirmation, etc
Conversion event
Campaign responses
© Datalicious Pty Ltd 16

> Conversion funnel 2.0
August 2013
Campaign responses (inbound spokes)
Offline campaigns, banner ads, email marketing,
referrals, organic search, paid search,
internal promotions, etc
Landing page(hub)
Success events (outbound spokes)
Bounce rate, add to cart, cart checkout, confirmed order,
call back request, registration, product comparison,
product review, forward to friend, etc

> Additional success metrics
Click
Through
Add To
Cart
Click
Through
Page
Bounce
Click
Through $
Click
Through
Call back
request
Store
Search ? $
$
$Cart
Checkout
Page
Views
?
Product
Views
Use additional metrics closer to the campaign origin

Exercise: Statistical significance

How many survey responses do you need
if you have 10,000 customers?
How many email opens do you need to test 2 subject lines
if your subscriber base is 50,000?
How many orders do you need to test 6 banner executions
if you serve 1,000,000 banners
Google “nss sample size calculator”

How many survey responses do you need
if you have 10,000 customers?
369 for each question or 369 complete responses
How many email opens do you need to test 2 subject lines
if your subscriber base is 50,000? And email sends?
381 per subject line or 381 x 2 = 762 email opens
How many orders do you need to test 6 banner executions
if you serve 1,000,000 banners?
383 sales per banner execution or 383 x 6 = 2,298 sales

> NPS survey and page ratings
Page ratings

Sentiment
ReachInfluence
> Measuring social media

> Relative or calculated metrics
§ Bounce rate
§ Conversion rate
§ Cost per acquisition
§ Pages views per visit
§ Product views per visit
§ Cart abandonment rate
§ Average order value

> Align metrics across channels
§ Paid search response rate
= website visits / paid search impressions
§ Organic search response rate
= website visits / organic search impressions
§ Display response rate
= website visits / display ad impressions
§ Email response rate
= website visits / emails sent
§ Direct mail response rate
= (website visits + phone calls) / direct mail pieces sent
§ TV response rate
= (website visits + phone calls) / (TV ad reach x frequency)

Exercise: Metrics framework

Level Reach Engagement Conversion +Buzz
Level 1,
people
Level 2,
strategic
Level 3,
tactical
Funnel
breakdowns
> Exercise: Metrics framework

Level Reach Engagement Conversion +Buzz
Level 1,
people
People
reached
People
engaged
People
converted
People
delighted
Level 2,
strategic
Display
impressions ? ? ?
Level 3,
tactical
Interaction
rate, etc ? ? ?
Funnel
breakdowns
Existing customers vs. new prospects, products, etc
> Exercise: Metrics framework

> Establish baseline to measure lift
Switch all advertising off for a period of time (unlikely)
or establish a smaller control group that is
representative of the entire population
(i.e. search term, geography, etc)
and switch off selected
channels one at a
time to min.
impact

IR − MI
MI
= ROMI + BE
> ROI, ROMI, BE, etc
IR − MI
MI
= ROMI
R − I
I
= ROI
R Revenue
I Investment
ROI Return on
investment
IR Incremental
revenue
MI Marketing
investment
ROMI Return on
marketing
investment
BE Brand equity

> Importance of calendar events
Traffic spikes or other data anomalies without context are
very hard to interpret and can render data useless

> Potential calendar events
§ Press releases
§ Sponsored events
§ Campaign launches
§ Campaign changes
§ Creative changes
§ Price changes
§ Website changes
§ Technical difficulties

ACME Corp
Data governance

> Process is key to ongoing success
Source: Omniture Summit, Matt Belkin, 2007

> Importance of data governance
Initial
Setup
Initial
Setup
Initial
Setup
When corporate standards are not established, updated, or enforced
by an organization, bad data begins to build within the (web)
analytics platform, which erodes its value over time.
<6 Months 6-12 Months >12 Months
Failure to manage
analytics platform
Deterioration of
business intelligence
Misinformed
business decisions
Loss of revenue
and increased costs

ACME Corp
Organisational structure

CEO
CMO
Analyst
CTO
Developer
CIO
Analyst
> Common organisational structure

CEO
CMTO
(CTO) Developer
CIO
Analyst Analyst
> New organisational structure

> Scaling teams with required skills
Data visualisation/reporting
Data mining/analysis
Data modelling
Fast analytics
Data processing/enhancing
Big data
Data collection
The Datalicious team
§ Data scientists
§ Business analysts
§ Data engineers
§ Web engineers
§ Platform admins
§ Project managers
§ Data strategists
Datastrategy

ACME Corp
Combining data

> The consumer data journey
To retention messagesTo transactional data
From suspect to To customer
From behavioural data From awareness messages
TimeTime
prospect

Transactional data
> Combining data sources is key
3rd party data
+
Whole is greater
than sum of its parts
Behavioural data
Prospects
Customers
Repeat customers

> Maximise identification points
20%
40%
60%
80%
100%
120%
140%
160%
0 4 8 12 16 20 24 28 32 36 40 44 48
Weeks
Cam
paign
response
Em
ailsubscription
Online
purchase
Repeatpurchase
Confirm
ation
em
ail
Em
ailnew
sletter
W
ebsite
login
Online
billpaym
ent
−−− Probability of identification through Cookies
August 2013 45© Datalicious Pty Ltd
App
dow
nload/access

http://www.acme.com/email-landing-page.html?
CampaignID=12345&
CustomerID=12345&
Demographics=M|25&
CustomerSegment=A1&
CustomerValue=High&
ProductHistory=A6&
NextProduct=A7&
ChurnRisk=High&
[...]
> Email click-through identification

acme.com/christianbartens redirects to amp.com.au?
CampaignID=12345&
CustomerID=12345&
Demographics=M|25&
CustomerSegment=A1&
CustomerValue=High&
ProductHistory=A6&
NextProduct=A7&
ChurnRisk=High&
[...]
> Personalised URLs for direct mail
Catch on
acme.com
404 error page

Customer data exposed in page or URL on login and logout
CustomerID=12345&
Demographics=M|25&
CustomerSegment=A1&
CustomerValue=High&
ProductHistory=A6&
NextProduct=A7&
ChurnRisk=High&
[...]
> Registration and login pages

> Identify customers across devices
Mobile,
Phone
Home
PC
Work
PC
Tablet POS Etc

Exercise: Identification points

> Identification best practice
Maximise data integrity
Age vs. year of birth
Free text vs. options
Use auto-complete
wherever possible

> Social single-sign on services
http://vimeo.com/16469480
Gigya.com
Janrain.com

> Power of geo-segmentation
Geo-segmentation can
help identify and target
under/over-performing
customer segments in
defined geographic areas
down to a postcode level.

> Address based data enhancements

http://www.acme.com/?
CampaignID=FB:12345&
Location=Sydney&
Demographics=M|25&
Interests=Traveling

ACME Corp
Optimising the funnel

> ACME cross-channel targeting
“Optimising the funnel from the bottom up”
Brand new
Prospects
Existing / engaged
Prospects
Existing / intent
Prospects
Existing
Customers

> Targeting profitable customers
Awareness Engagement Conversion Loyalty
Prospects Customers

Brand new
Prospects
Existing / engaged
Prospects
Existing / intent
Prospects
Existing
Customers
Up/cross-sell
Process
re-initiation
Re-targeting
Audience
extension

Transactional data
> Combining data sources
3rd party data
+
Whole is greater
than sum of its parts
Behavioural data
Prospects
Customers
Repeat customers

> Transactions plus behaviours
+
one-off collection of demographical data
age, gender, address, etc
customer lifecycle metrics and key dates
profitability, expiration, etc
predictive models based on data mining
propensity to buy, churn, etc
historical data from previous transactions
average order value, points, etc
CRM Profile
Updated Occasionally
tracking of purchase funnel stage
browsing, checkout, etc
tracking of content preferences
products, brands, features, etc
tracking of external campaign responses
search terms, referrers, etc
tracking of internal promotion responses
emails, internal search, etc
Site Behaviour
Updated Continuously

> Customer profiling in action
Using website and email responses
to learn a little bite more about
subscribers at every
touch point to keep
refining profiles
and messages.

1,875% ROI

Brand new
Prospects
Existing / engaged
Prospects
Existing / intent
Prospects
Existing
Customers
Up/cross-sell
Process
re-initiation
Re-targeting
Audience
extension

1,333% ROI

Brand new
Prospects
Existing / engaged
Prospects
Existing / intent
Prospects
Existing
Customers
Up/cross-sell
Process
re-initiation
Re-targeting
Audience
extension

PREMIUM
OFFER
1300 PRIORITY
PREMIUM
EXPERIENCE

PREMIUM
EXPERIENCE

> Network wide re-targeting
Product A
Product B
prospect
Product A
prospect
Product A
customer
Product B Product C
Product C
prospect
Product B
prospect
Product B
customer
Product A
prospect
Product C
prospect
Product C
customer

> Network wide re-targeting
Product B
prospect
Product A
prospect
Product A
customer
Product C
prospect
Product B
prospect
Product B
customer
Product A
prospect
Product C
prospect
Product C
customer
Group wide campaign with approximate impression targets by product rather than hard budget limitations

Closer
Message 1
Message 1
Message 1
> Story telling or ad-sequencing
Influencer Influencer $
Message 2
Message 2 Message 3
Message 2 Message 3 Message 4
Message 3
Message 4
Message 4
Introducer
Product A
Product B
Product C

> Ad-sequencing in action
Marketing is about
telling stories and
stories are not static
but evolve over time
Ad-sequencing can help to
evolve stories over time the
more users engage with ads

Exercise: Re-targeting matrix

Purchase
Cycle
Segmentation based on: Search keywords,
display ad clicks and website behaviour Data
Points
Default,
awareness
Default
Research,
consideration
Product
view, etc
Purchase
intent
Checkout,
chat, etc
Existing
customer
Login, email
click, etc
> Exercise: Re-targeting matrix

Purchase
Cycle
Points
Default Product A Product B
Default,
awareness
Acquisition
message D1
Acquisition
message A1
Acquisition
message B1
Default
Research,
consideration
Acquisition
message D2
Acquisition
message A2
Acquisition
message B2
Product
view, etc
Purchase
intent
Acquisition
message D3
Acquisition
message A3
Acquisition
message B3
Checkout,
chat, etc
Existing
customer
Cross-sell
message D4
Cross-sell
message A4
Cross-sell
message B4
Login, email
click, etc
> Exercise: Re-targeting matrix

> Unique phone numbers
2 out of 3 callers
hang up as they
cannot get their
information fast
enough.
Unique phone
numbers can
help improve
call experience.

Purchase
Cycle
Points
Default Product A Product B
Default,
awareness
1300 000 001 1300 000 005 1300 000 009 Default
Research,
consideration
1300 000 002 1300 000 006 1300 000 010
Product
view, etc
Purchase
intent
1300 000 003 1300 000 007 1300 000 011
Checkout,
chat, etc
Existing
customer
1300 000 004 1300 000 008 1300 000 012
Login, email
click, etc
> Website call center integration

800% ROI

Brand new
Prospects
Existing / engaged
Prospects
Existing / intent
Prospects
Existing
Customers
Up/cross-sell
Process
re-initiation
Re-targeting
Audience
extension

> Audience extension (Cookies)

> Audience extension (Address)
August 2013
RDA Research
geoTribes
Roy Morgan
Asteroid
Offline media
behaviour
Online media
behaviour
Experian
Hitwise
Experian
Mosaic
Veda
DataExpress
Online media
planning
Offline media
planning
Customer
address
Geo-demographic
segmentation
Prospect,
customer segm.
Customer
value potential
Customer
targeting
Roy Morgan
Value Segments
Customer
transactions
Customer
segmentation
Current
customer value

> Targeting profitable customers
Awareness Engagement Conversion Loyalty
Prospects Customers
Geo-demographic segmentation

> Working back from existing clients
[…]
3rd party
media
insights
Geo-
segments
Customer
address
Historic
sales
Roy
Morgan
(offline)
Experian
Hitwise
(online)
Profitable
customers
Other
segments
Who are my profitable
customer and where do
I find more of the same?

Awareness Engagement Conversion
Audience
purchased
Geo-
segments
Audience
purchased
Audience
engaged
Geo-segments based
on historic sales
Audience 1 Segment 1 Segment 1
GAP Segment 2 GAP GAP Segment 2
Segment 3 GAP Segment 3
GAP Segment N GAP GAP Segment N
> Identifying gaps = opportunities
Audience 1 = Segment 1
Audience 2 = Segment 3

480% ROI

Brand new
Prospects
Existing / engaged
Prospects
Existing / intent
Prospects
Existing
Customers
Up/cross-sell
Process
re-initiation
Re-targeting
Audience
extension
1,875% ROI
1,333% ROI
800% ROI
480% ROI

Exercise: Optimisation ROI

ACME Corp
Testing & optimisation

Don’t reinvent the wheel

> Small things sometimes count

> Introducing hero vs. challengers
Hero #1
CTR = 1%
Challenger #1
CTR = 0.5%
Challenger #2
CTR = 1.5%
Challenger #3
CTR = 1%
Challenger #4
CTR = 1%
New hero #2
= Challenger #2

Rather than testing all combinations of alternative page content (i.e. A/B
testing), the Taguchi Method (i.e. multivariate MV testing) is a way of
reducing the number of different test scenarios (recipes) but still yield
useful test results. Essentially, the optimal page design is ‘predicted’
from the test results by analysing which page elements and element
combinations were most influential overall.
> A/B vs. MV (Taguchi) method
Test elements
(i.e. parts of page)
Test alternatives
(i.e. test content)
Full set of test
combinations (A/B)
Reduced Taguchi
test scenarios (MV)
3 2 8 4
7 2 128 8
4 3 81 9
5 4 1024 16

Offer
Issue
Offer
> Design and test experiences
Email
Live chat Phone call
Phone call Letter Email
Issue
All customers
Segment A, B, C
Segment D, E
Influencers
Lovers
Display
Postcard
Display
FAQs

Exercise: Statistical significance

How many click-throughs do you need to test 3
landing pages if you have 30,000 visitors?
How many conversions do you need to test 3
How many click-throughs do you need to test 3 landing pages
if you have 30,000 visitors but only expose 10% to the test?

How many click-throughs do you need to test 3
369 per test or 1,107 clicks in total
How many conversions do you need to test 3
369 per test or 1,107 conversions in total
How many click-throughs do you need to test 3 landing pages
if you have 30,000 visitors but only expose 10% to the test?
277 per test or 831 clicks in total

Exercise: Testing matrix

Test Segment Content Success Difficulty Potential
> Exercise: Testing matrix

Test Segment Content Success Difficulty Potential
Test 1 Product 1
Offer 1A
Clicks Low $100kOffer 1B
Offer 1C
Test 2 Product 2
Offer 2A
Clicks High $100kOffer 2B
Offer 2C
> Exercise: Testing matrix

Targeting before testing

> Garbage in, garbage out
Avinash Kaushik:
“The principle of garbage in, garbage out
applies here. [… what makes a behaviour
targeting platform tick, and produce results, is
not its intelligence, it is your ability to actually
feed it the right content which it can then target
[…. You feed your BT system crap and it will
quickly and efficiently target crap to your
customers. Faster then you could
ever have yourself.”

ACME Corp
Measuring success

Direct mail,
email, etc
Facebook
Twitter, etc
> Channels influence each other
POS kiosks,
loyalty cards, etc
CRM
program
Home pages,
portals, etc
YouTube,
blog, etc
Paid
search
Organic
search
Landing pages,
offers, etc
PR, WOM,
events, etc
TV, print,
radio, etc
= Paid media
= Viral elements
Website, call
center, retail
= Sales channels
Display ads,
affiliates, etc

> First and last click attribution
Chart shows
percentage of
channel touch
points that lead
to a conversion.
Neither first
nor last-click
measurement
would provide
true picture
Paid/Organic Search
Emails/Shopping Engines

> Media attribution approaches
Success
$100
Success
$100
Display Affiliate
Search
$100
Success
$100
Last channel
gets all credit
First channel
gets all credit
All channels get
equal credit
Success
$100
All channels get
custom credit
Display
$100
Affiliate Search
Display
$33
Affiliate
$33
Search
$33
Display
$15
Affiliate
$35
Search
$50

> Ad clicks inadequate measure
Only a small minority of people actually click on ads, the majority
merely processes them (if at all) like any other advertising without an
immediate response so advertisers cannot rely on clicks as the sole
success measure but should instead focus on impressions delivered

Closer
Paid
search
Display
ad views
TV/print
responses
> Full purchase path tracking
Influencer Influencer $
Display
ad clicks
Online
sales
Affiliate
clicks
Social
referrals
Offline
sales
Organic
search
Social
buzz
Retail
visits
Lifetime
profit
Organic
search
Emails,
direct mail
Direct
site visits
Introducer

> Combine paths across devices
Mobile Home Work
Tablet Media Etc

> Media attribution models
$100
Even/linear
attribution
Time decay
attribution
Custom
attribution
10% 15% 25% 50%
Display
impression
Display
impression
Display
click
Search
click
10% 10% 50% 30%
25% 25% 25% 25%

10% 30% 10% 50%
10% 50% 30%10%
> Custom (weighted) attribution
$100
Weighted
attribution
$100
Weighted
attribution
Display
impression
Display
impression
Display
click
Search
click
Display
impression
Search
click
Display
impression
Display
click

Touch
point 1
> Analytics to pick the best model
Touch
point 2
Touch
point 3
Touch
point N
CloserInfluencer Influencer $Introducer
Touch
point 1
Touch
point 2
Touch
point 3
Touch
point N
Touch
point 1
Touch
point 2
Touch
point 3
Touch
point N
✖
✔
✖

> Attribution models compared
COST PER CONVERSION
Last click
attribution
Custom (weighted)
attribution

> Insights to maximise media ROI
COST PER CONVERSION
Last click
attribution
Even/weighted
attribution
?
Email
?
Direct
mail
?
Internal
ads?
Website
content
?
TV/Print

> Redistributing media spend
ROI FULL PURCHASE PATH
TOTALCONVERSIONVALUE
Maintain
spend
Increase
spend
Reduce
spend
Publisher 1
Publisher 2
Publisher 3
[…]
Publisher N

Contact me
cbartens@datalicious.com
Learn more
blog.datalicious.com
Follow us
twitter.com/datalicious

Smart data driven marketing

> Conversion funnel design
Visits
Product Views
Cart Adds
Checkouts
Conversions
Visits
Non-Bounces*
Engagements**
Leads**
Conversions
* Non-bounce event
** Serialised events,
i.e. once per visit

> Success: ROMI + BE
§ Establish incremental revenue (IR)
– Requires baseline revenue to calculate additional
revenue as well as revenue from cost savings
§ Establish marketing investment (MI)
– Requires all costs across technology, content, data
and resources plus promotions and discounts
§ Establish brand equity contribution (BE)
– Requires additional soft metrics to evaluate subscriber
perceptions, experience, attitudes and word of mouth
IR − MI
MI
= ROMI + BE

> Combining data sources

> Combine data across devices
Mobile Home Work
Tablet Media Etc

> Importance of online experience
The consumer decision process is changing from linear to circular.
Consideration
set now grows
during online
research phase
which increases
importance of
user experience
during that phase
Online research

> Increase revenue by 10-20%

> Targeting: Quality vs. quantity
30% existing customers with extensive
profile including transactional history of
which maybe 50% can actually be
identified as individuals
30% new visitors with no
previous website history
aside from campaign or
referrer data of which
maybe 50% is useful
10% serious
prospects
with limited
profile data
30% repeat visitors with
referral data and some
website history allowing
50% to be segmented by
content affinity

> The holy trinity of testing
1. The headline
– Have a headline!
– Headline should be concrete
– Headline should be first thing visitors look at
2. Call to action
– Don’t have too many calls to action
– Have an actionable call to action
– Have a big, prominent, visible call to action
3. Social proof
– Logos, number of users, testimonials,
case studies, media coverage, etc

> Best practice testing roadmap
§ Phase 1: A/B test
– Test same landing page
content in different
layouts
§ Phase 2: MV test
– Test different content
element combinations
within winning layout
§ Phase 3: Repeat
– Hero vs. challengers
§ Phase 4: Re-targeting
Element #1: Prominent headline
Element #2:
Call to action
Supporting
content
Element #3: Social proof / trust
Terms and conditions

Consumer Intelligence Analytics Workshop

More Related Content

What's hot (20)

Similar to Consumer Intelligence Analytics Workshop (20)

More from Christian Bartens (20)

Recently uploaded (20)

Consumer Intelligence Analytics Workshop