@adrian_kingwell @mezzolabs
Making Data Personal
or How to sell Tom Cruise car insurance
Adrian Kingwell
Mezzo Labs
@adrian_kingwell @mezzolabs
What is personalised marketing?
The ability to deliver the right message
to the right person
through the right medium
at the right time
@adrian_kingwell @mezzolabs
What is personalised marketing?
It’s all about relevance and speed
@adrian_kingwell @mezzolabs
How do we sell Tom Cruise some
car insurance?
@adrian_kingwell @mezzolabs
What do we know about Tom?
• Who is he? Is he in our target audience?
• Does he need car insurance?
• Is he ready to buy?
• What would make him buy our insurance?
• How do we get our offer in front of him at the right
time?
@adrian_kingwell @mezzolabs
Tom’s demographic cohorts
• Actor – actors pay high premiums
• High net worth individual - worth $470 million
• Divorced, currently single
• Owns more than one car
• 53 years old – low risk
• Scientologist – doesn’t drink, smoke or take drugs
@adrian_kingwell @mezzolabs
Tom’s behavioural cohorts
• Ready to buy - got a quote for car insurance two
days ago
• Fussy buyer - spent 15 minutes ping-ponging
between the postcode field and the quote page
• Frequent OLB user - logs into online banking via the
app 3-4 times per week
• Lapsed customer - bought our car insurance a few
years ago
• Tech savvy - has multiple devices
@adrian_kingwell @mezzolabs
We need personal marketing
Let’s find Tom
Let’s reach him through the medium of his choice
Let’s send him a promotion that is relevant to him
Let’s do it at the right time
@adrian_kingwell @mezzolabs
Tom is in the neighborhood
• Tom is in Westfield shopping mall just off the A40
• He’s not looking at his iPhone but he is looking at
in-store display advertising
@adrian_kingwell @mezzolabs
@adrian_kingwell @mezzolabs
How did it go?
Did we find Tom?
Did we target him through the right medium?
Did we send him the right promotion?
Was it the right time?
Something went wrong in our personalised marketing
@adrian_kingwell @mezzolabs
How good is Minority Report marketing?
Targeted, but not exactly personalised
• More like programmatic marketing
Wasteful
• Tom was never going to be interested in Guinness
Expensive
• We failed to outbid luxury brands
It is still mass marketing
• Generic ads in locations that have the right demographics
@adrian_kingwell @mezzolabs
Let’s go back to basics
@adrian_kingwell @mezzolabs
First we need some data…
Website
Mobile App
Paid media
3rd party
CRM
Credit to Signal for this model
Blend the
data
@adrian_kingwell @mezzolabs
Next we create a single customer view…
Blend the
data
Match
unique
identifiers
Website
Mobile App
Paid media
3rd party
CRM
Credit to Signal for this model
@adrian_kingwell @mezzolabs
Then we create some segments…
Blend the
data
Match
unique
identifiers
Create
segments
Website
Mobile App
Paid media
3rd party
CRM
Credit to Signal for this model
@adrian_kingwell @mezzolabs
Finally we can do something with it…
Automation
Analysis
Blend the
data
Match
unique
identifiers
Create
segments
Website
Mobile App
Paid media
3rd party
CRM
Credit to Signal for this model
@adrian_kingwell @mezzolabs
Automation
Analysis
Email
Technology maps onto this…
Collect and
blend the
data
Match
unique
identifiers
Create
segments
Website
Mobile App
Paid media
3rd party
CRM
DMP
CRO
DSP
@adrian_kingwell @mezzolabs
CohortHypothesis
Website
Mobile App
Paid media
3rd party
CRM
Analysis
Customer ID
Content ID
Demographics
Behaviours
Website
Mobile App
Paid media
Email
Offline display
Blend the
data
Match
unique
identifiers
Create
segments
Relevance
Speed
@adrian_kingwell @mezzolabs
Challenges
Website
Mobile App
Paid media
3rd party
CRM
Website
Mobile App
Paid media
Email
In store
Blend the
data
Match
unique
identifiers
Create
segments
@adrian_kingwell @mezzolabs
Challenges
Website
Mobile App
Paid Media
3rd party
CRM
Web analytics
Mobile SDK
Adserving, video,
aggregators
Offline systems
CRM
User ID
How do we match
different IDs to
universal ID?
How do we match
the Universal ID
with 3rd party data?
Who are our users?
Event ID
What did they do?
When?
How much?
Where are they in
the buying cycle?
Identify the critical
few events and
filter out all the
rest
Campaign ID
What creative did
they see?
Where?
When?
At what cost?
What do they
respond to?
@adrian_kingwell @mezzolabs
Challenges
Website
Mobile App
Web analytics
Mobile SDK
Language
Do all data sources
define data the
same way?
Do we have a data
dictionary?
Do we have a data
layer?
Time
Are all data sources
on the same time
base?
How often can we
get access to the
data?
How quickly can we
process the data?
Governance
How do we ensure
that nobody breaks
anything?
Paid Media
3rd party
CRM
Adserving, video,
aggregators
Offline systems
CRM
@adrian_kingwell @mezzolabs
Challenges
Website
Mobile App
Web analytics
Mobile SDK
Analytics
Customer journey
is no longer linear
The funnel is no
longer funnel-
shaped
Customer goes
offsite > onsite >
app store > app >
branch
Management
Who owns this?
The ROI is difficult
Organisations are
not set up to do
this
Privacy, security
and compliance
What will Risk and
Compliance say?
Are your suppliers
ready?
Will your customers
think it is creepy?
Paid Media
3rd party
CRM
Adserving, video,
aggregators
Offline systems
CRM
@adrian_kingwell @mezzolabs
Back to Tom
• It’s all about speed and relevance
• If we can be relevant, our marketing won’t be wasted
• If we can execute quickly, we won’t get beaten
• We can’t read Tom’s mind
• But we can calculate his propensity to buy
• We can reduce our spend when he’s got a low chance of buying
• We can increase our spend when he’s got a high chance
• Are we ready to do this?
• Do we have the right skills
• Do we have the technology?
• Is the data ready?

Adrian kingwell, Mezzo Labs - Making Data Personal

  • 1.
    @adrian_kingwell @mezzolabs Making DataPersonal or How to sell Tom Cruise car insurance Adrian Kingwell Mezzo Labs
  • 2.
    @adrian_kingwell @mezzolabs What ispersonalised marketing? The ability to deliver the right message to the right person through the right medium at the right time
  • 3.
    @adrian_kingwell @mezzolabs What ispersonalised marketing? It’s all about relevance and speed
  • 4.
    @adrian_kingwell @mezzolabs How dowe sell Tom Cruise some car insurance?
  • 5.
    @adrian_kingwell @mezzolabs What dowe know about Tom? • Who is he? Is he in our target audience? • Does he need car insurance? • Is he ready to buy? • What would make him buy our insurance? • How do we get our offer in front of him at the right time?
  • 6.
    @adrian_kingwell @mezzolabs Tom’s demographiccohorts • Actor – actors pay high premiums • High net worth individual - worth $470 million • Divorced, currently single • Owns more than one car • 53 years old – low risk • Scientologist – doesn’t drink, smoke or take drugs
  • 7.
    @adrian_kingwell @mezzolabs Tom’s behaviouralcohorts • Ready to buy - got a quote for car insurance two days ago • Fussy buyer - spent 15 minutes ping-ponging between the postcode field and the quote page • Frequent OLB user - logs into online banking via the app 3-4 times per week • Lapsed customer - bought our car insurance a few years ago • Tech savvy - has multiple devices
  • 8.
    @adrian_kingwell @mezzolabs We needpersonal marketing Let’s find Tom Let’s reach him through the medium of his choice Let’s send him a promotion that is relevant to him Let’s do it at the right time
  • 9.
    @adrian_kingwell @mezzolabs Tom isin the neighborhood • Tom is in Westfield shopping mall just off the A40 • He’s not looking at his iPhone but he is looking at in-store display advertising
  • 10.
  • 11.
    @adrian_kingwell @mezzolabs How didit go? Did we find Tom? Did we target him through the right medium? Did we send him the right promotion? Was it the right time? Something went wrong in our personalised marketing
  • 12.
    @adrian_kingwell @mezzolabs How goodis Minority Report marketing? Targeted, but not exactly personalised • More like programmatic marketing Wasteful • Tom was never going to be interested in Guinness Expensive • We failed to outbid luxury brands It is still mass marketing • Generic ads in locations that have the right demographics
  • 13.
  • 14.
    @adrian_kingwell @mezzolabs First weneed some data… Website Mobile App Paid media 3rd party CRM Credit to Signal for this model Blend the data
  • 15.
    @adrian_kingwell @mezzolabs Next wecreate a single customer view… Blend the data Match unique identifiers Website Mobile App Paid media 3rd party CRM Credit to Signal for this model
  • 16.
    @adrian_kingwell @mezzolabs Then wecreate some segments… Blend the data Match unique identifiers Create segments Website Mobile App Paid media 3rd party CRM Credit to Signal for this model
  • 17.
    @adrian_kingwell @mezzolabs Finally wecan do something with it… Automation Analysis Blend the data Match unique identifiers Create segments Website Mobile App Paid media 3rd party CRM Credit to Signal for this model
  • 18.
    @adrian_kingwell @mezzolabs Automation Analysis Email Technology mapsonto this… Collect and blend the data Match unique identifiers Create segments Website Mobile App Paid media 3rd party CRM DMP CRO DSP
  • 19.
    @adrian_kingwell @mezzolabs CohortHypothesis Website Mobile App Paidmedia 3rd party CRM Analysis Customer ID Content ID Demographics Behaviours Website Mobile App Paid media Email Offline display Blend the data Match unique identifiers Create segments Relevance Speed
  • 20.
    @adrian_kingwell @mezzolabs Challenges Website Mobile App Paidmedia 3rd party CRM Website Mobile App Paid media Email In store Blend the data Match unique identifiers Create segments
  • 21.
    @adrian_kingwell @mezzolabs Challenges Website Mobile App PaidMedia 3rd party CRM Web analytics Mobile SDK Adserving, video, aggregators Offline systems CRM User ID How do we match different IDs to universal ID? How do we match the Universal ID with 3rd party data? Who are our users? Event ID What did they do? When? How much? Where are they in the buying cycle? Identify the critical few events and filter out all the rest Campaign ID What creative did they see? Where? When? At what cost? What do they respond to?
  • 22.
    @adrian_kingwell @mezzolabs Challenges Website Mobile App Webanalytics Mobile SDK Language Do all data sources define data the same way? Do we have a data dictionary? Do we have a data layer? Time Are all data sources on the same time base? How often can we get access to the data? How quickly can we process the data? Governance How do we ensure that nobody breaks anything? Paid Media 3rd party CRM Adserving, video, aggregators Offline systems CRM
  • 23.
    @adrian_kingwell @mezzolabs Challenges Website Mobile App Webanalytics Mobile SDK Analytics Customer journey is no longer linear The funnel is no longer funnel- shaped Customer goes offsite > onsite > app store > app > branch Management Who owns this? The ROI is difficult Organisations are not set up to do this Privacy, security and compliance What will Risk and Compliance say? Are your suppliers ready? Will your customers think it is creepy? Paid Media 3rd party CRM Adserving, video, aggregators Offline systems CRM
  • 24.
    @adrian_kingwell @mezzolabs Back toTom • It’s all about speed and relevance • If we can be relevant, our marketing won’t be wasted • If we can execute quickly, we won’t get beaten • We can’t read Tom’s mind • But we can calculate his propensity to buy • We can reduce our spend when he’s got a low chance of buying • We can increase our spend when he’s got a high chance • Are we ready to do this? • Do we have the right skills • Do we have the technology? • Is the data ready?