Download - Epiphany Summer Conference 2016
COPY
RIGH
T JA
YWIN
G PL
C 20
15
COPY
RIGH
T JA
YWIN
G PL
C 20
15
The new face of search
Data, audience and the programmatic
marketer
Why We Need a Cross-Device Solution
ALESSANDRA ALARI, GOOGLE
Understanding Performance Used to be Easy in 2007
Moment Search Conversion Report
$
~10 Years Later
Example Forecast
Desktop
2012 2013 2014 2015 2016
Mobile
Sear
ch Q
uerie
s
Mobile Moment
Tablet
2017
Mobile Queries Overtook Desktop Queries Globally in 2015
Across millions of websites using Google Analytics today, we're seeing more than half of all web traffic now coming from smartphones and tablets
Google Analytics Data, US, Q1 2016
Quarter of Year
Conv
ersio
n Ra
teDesktop
Mobile
Tablet
Source: Monetate Ecommerce Quarterly - http://www.monetate.com/resources/research/
On Paper, Mobile Seems to be Underperforming
Nearly
80% of online users use multiple devices
Source: Google / Ipsos Connect, March 2016, GPS Omnibus, n=2,013 US online respondents 18+
Today’s customer journeys are much more fragmented
Moment Search Conversion Report
Understanding Performance in a Multi-Device World is Complicated
© Google Inc. 2016. All rights reserved.
User clicks on Retailer’s desktop ad
Later orders flowers on his
tablet
Web to Mobile Browser to Browser
Buys three pairs of running shoes
on Retailer’s Mobile site
User clicks on Mother’s
Day flowers ad on his laptop in the office
What’s a Cross Device Conversion?
3
How can we capture
multi-screen behaviour?
Moving Beyond Cookies
Industry Options Available
Algos based on non-PII public info (e.g., device
type, IP address) to approximate device links;
does not create persistent ID
Statistical Linkingor Probabilistic
Uses identifiers that a user cannot control or reset - e.g., fonts, screen resolutions - to create a persistent, unique ID for
that device
FingerprintingTrack users who log-in to
a service on multiple devices
User Log-In or Deterministic
Benefits of the Google Solution
Tracking that Works Across the Entire Google Ecosystem
SCALE METHODOLOGY BREADTH
7x 1B+ Users
iOS
Attribution 360
Measurement Today
Attribution
Cross Device
Last Click
No. In Reality, We Are Here:
6% Our current solution provides an adequate single view of our customers
47% We have a solution in place but it still has gaps in coverage of devices, touch-points and data
32% We have not yet implemented a solution to achieve a single view of our customers
Source: Signal, “Preparing for Cross-Channel Success: Solving the Identity Puzzle” , March 2015
19% We have implemented a solution but its shortcomings severely limit its practical effectiveness
Cross-Device is Not For
Everyone
Photobox proves the value of mobile search ad campaigns through geo experiment
5x Incremental ROI driven by mobile
“Now we have a view of the holistic performance of
mobile search advertising on our overall sales numbers, this
has enabled us to more effectively attribute the spend
to sales.” James Duggan, Head of Digital Marketing, PhotoBox
UK
Proprietary + Confidential
Made.com used Adwords cross-device reports to gain insight into the devices their customers use to search, browse and buy
+204% YoY Increase
in mobile conversions
+21% Increase in
overall conversions
+105%
Increase in mobile ROAS
The Bright Future
9:34 AM
contentbazaar.co/20160314/pi-way
9:34 AM
contentbazaar.co/20160314/pi-way
Slow Loading
ContentShifting
Non-ResponsiveContent
9:34 AM
contentbazaar.co/20160314/pi-way
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.
A Better Mobile Web
Attribution
PRECISION
SCAL
ABIL
ITY
Rules-based
Attribution
Data-driven Attribution
Controlled Experiments
Freq
uent
One
-off
Correlational Causal
In Summary
Use Insights to Value Mobile
Go from Last Click toCross Device
Accept the Complexity
Thank [email protected]
Personalising your customer journey
Chris Rowett, Director of Performance, EpiphanyMitch Vidler, Head of Digital Analysis, Jaywing
COPY
RIGH
T JA
YWIN
G PL
C 20
15
The path to personalisation
Single Customer
View
Audience Segmentation
Tracking & Rules
Personalisation
COPY
RIGH
T JA
YWIN
G PL
C 20
16
It’s Hard.
COPY
RIGH
T JA
YWIN
G PL
C 20
16
Single customer view
COPY
RIG
HT
JAYW
ING
PLC
2015
Customer interactionsCustomer interactions
COPY
RIGH
T JA
YWIN
G PL
C 20
16
Audiences – granular segmentation
Researching Their First Pet
First Time Pet Purchaser
Long Term Pet Owner
COPY
RIGH
T JA
YWIN
G PL
C 20
16
Feed information into rulesMitch & Chris can share the same audience, but still need individual creative
Mitch Chris
COPY
RIGH
T JA
YWIN
G PL
C 20
16
The challenge
Static Customer
DataName
AddressDOB
Transactional historyLoyalty data
Dynamic Behavioural
DataOnline data
Browsing behaviour Clicks & views
Web trafficeCRM analytics
Hidden away in products such as GA
Visible but not useableWhat to capture? What
to do with it?How to link with current
CRM activity?
Used to create a Single Customer View
COPY
RIGH
T JA
YWIN
G PL
C 20
16
The challenge
Static Customer
DataName
AddressDOB
Transactional historyLoyalty data
Dynamic Behavioural
DataOnline data
Browsing behaviour Clicks & views
Web trafficeCRM analytics
COPY
RIGH
T JA
YWIN
G PL
C 20
16
• Media Placement Found and Purchased
• Blank framework calls the rules engine
• Information from the Single Customer View is loaded
• Personalised creative is served
Level 1 Level 2
Integrate a personalised rule set
Media Partner
COPY
RIGH
T JA
YWIN
G PL
C 20
16
Display personalisation
COPY
RIGH
T JA
YWIN
G PL
C 20
16
Hyper-personalised email
COPY
RIGH
T JA
YWIN
G PL
C 20
16
Onsite content personalisationAs with media creative, your own website can be personalised to suit the user, increasing engagement and conversion.
A good Conversion Rate Optimisation program can feed into site personalisation
COPY
RIGH
T JA
YWIN
G PL
C 20
16
Better Product Recommendations
COPY
RIGH
T JA
YWIN
G PL
C 20
15
Existing approaches
Rules-based Attribute Modelling
Collaborative Filtering
Simple rules-based approach where data
volume and complexity is low
Incorporates customer demographic information and
product featuresLimited incorporation of
purchase history Does not scale well to large
product inventories Does not identify patterns in
purchase behaviour
Scales to large product inventories
Identifies patterns in purchase behaviour
Does not incorporate demographic information or
product features (suffers from “colds starts”)
COPY
RIGH
T JA
YWIN
G PL
C 20
16
Better predictions at all stages
Attribute ModellingCollaborative FilteringJaywing Approach
Customer / Product Lifecycle
Pred
ictive
Powe
r
COPY
RIGH
T JA
YWIN
G PL
C 20
15
AlmanacAlmanac can bring together disparate data sources and feed this into an enhanced Single Customer View
Mobile app (store visit)
Personalised display/
retargetingPersonalised
emailTargeted content
Enhanced Single Customer View
COPY
RIGH
T JA
YWIN
G PL
C 20
16
A virtuous circle of activity
Richer Data
Better Recommendations
Putting optimal product recommendations in front of customers drives more sales
An increased propensity to buy builds deeper relationships with customers, who move more of their buying to youDeeper customer relationships
generate more useful data on purchase patterns and behaviours
Our advanced algorithm uses the richer data to further improve recommendations
More Purchases
Deeper Relationships
COPY
RIGH
T JA
YWIN
G PL
C 20
16
Many new customers but also loyal existing customers
KWD First Visitor79%
KWD Return Visi-tor
21%
KWD Visitor Type
KWD First Visitor KWD Return Visitor
2 Visits21%
3-4 Vis-its
25%
5+ Visits54%
Repeat KWD Visitor Breakdown
2 Visits3-5 Visits5+ Visits
Personalisation also presents the option to down weight spend on existing loyal customers
COPY
RIGH
T JA
YWIN
G PL
C 20
16
Optimise campaigns for profit
SKU-level Product Profitability Data
Profit Optimised Campaigns
Granular SKU-level Product Recommendations
Product 1
Product 2
Product 3
Product 4
Product 1
Product 2
Product 3
Product 4
Product 1
Product 2
Product 3
Product 4
Confidential & ProprietaryConfidential & Proprietary
Search in 2016The Audience RevolutionOli Petas, UK Search Audience & Automation Lead
7x Increase in Click Through Rate
7x Increase in Conversions
Return of £6 for every £1 invested
Goals Approach ResultsGoals Approach Results
Using On-Site DataI Clarks
“Search retargeting gives the consumer a much more personalised experience.” John Ashton, Head of Multi-Channel Retail, Clarks
Increase the impact of web campaigns
Get more in-depth analysis of customer activity online
Improve the conversion rate
Placed Remarketing Lists for Search Ads (RLSA) tags on every page of the Clarks website
Analysed the behaviour of website visitors
Tailored keywords, creative and bids for individual users
Full Case Study | Video
13% Lower Cost per Acquisition
52% Higher Click Conversion Rate
12% More Revenue per Conversion
32% Higher Click Through Rate
Goals Approach ResultsGoals Approach Results
Connecting CRM Data through Customer Match I Pegasus Airlines
“Customer Match empowered us to tailor our messages to both our clients and prospective clients. The strategy has increased our revenues and decreased cost per sale.” Didem Namver, Digital Marketing Manager, Pegasus Airlines
Decrease cost per ticket sale
Increase traffic
Take advantage of new optimisation opportunities
Implemented Customer Match in Adwords
Automatically adjusted bids and messaging to connect with high-value customers
Full Case Study
Maximising Marketing Impact
Rohan Gifford, Research Manager, Google
Across 56 sales correlation studies in EMEA YouTube had
a Return on Investment (ROI) higher than TV in 43
(77% of total) studies
I Studies spanned countries and verticals
AutomotiveClassifieds & LocalConsumer Packaged GoodsMedia & EntertainmentRetailTechnology
I Three methodologies included
Media mix models
Consumer scan panel studies
Geo-tests
Across 17 studies where optimisation was undertaken the recommendation was to
increase YouTube spend in 100% of cases
In 14/17 optimisations (82%) the recommendation was to more than double
YouTube spend
YouTubeRe
venu
e
OptimalCurren
t
Spend
Making the Future Happen7 Evidence Based Tips
Rohan Gifford, Research Manager, Google
I
TrueView Skippable ads on YouTube
View Through Rate % of TrueView ads that are not skipped
Brand Lift brand lift experiments built into YouTube
Three terms to know
1.FrequencyFirst impressions count most, but brand lift builds with frequency up to 6 a week
Top of Mind Ad Recall
Control 1 2 3 4 5 6 7 8 9 10
0%
3%
Frequency
Source: Google/eye square research 2015/2016.
2.DecayReach people regularly across the campaign
3.DeviceMobile delivers even higher brand impact: plan mobile first
4.Contentget higher brand lift next to the most engaging content
Ad Recall Lift increases with increasing Preference Score
Source: Internal Google Data, based on 5,500 TrueView Brand Lift studies which ran between April and October 2015. Confidence interval represents 95% CI
5.Targetingaffinity targeting drives better results
6.CreativeData is creative friend
7.YouTube and TVAND not OR