In this deck I discuss why, and how data makes me tick. I've been doing work with data for decades but it's more exciting now than ever. Why do I get nerdy excitement from data? Because of my personal data, because data fuels change, because data is used for good. I'm immersed in data that has purpose on a daily basis.
Can you distinguish between data that has purpose and data that fuels vanity? Can you tell when data is for activation and when it's for reporting? Which data gets you promoted and which data gets you bogged down? A click bait teaser - you already know the answer...
2. “If you don’t have data,
you’re just another person
with an opinion.”
Media.Monks
3. “If you do have data, you’re
still just another person with
an opinion.”
Media.Monks
4. Unless your data is active in
decision making, it’s sole use
is “for reporting”.
TRUTH
Media.Monks
5. Media.Monks Proprietary & Confidential 5
Media.Monks Proprietary & Confidential 5
ACTUAL good stuff
Downloads, transactions, soft and
hard conversions. Channel data,
impressions, costs, clicks. Meta
data, cohorts, propensity scores,
consented data, modelled data
Bounce rate, scroll tracking, screen
resolution, flash version, device category,
browser brand and version
The rest is all vanity and confetti or ignored
What’s
tracked
Typical Strategic Alignment
What’s
useful
6. Media.Monks Proprietary & Confidential 6
Media.Monks Proprietary & Confidential 6
ACTUAL good stuff
Downloads, transactions, soft and
hard conversions. Channel data,
impressions, costs, clicks. Meta
data, cohorts, propensity scores,
consented data, modelled data
Bounce rate, scroll tracking, screen
resolution, flash version, device category,
browser brand and version
The rest is all vanity and confetti, or ignored
Typical Strategic Alignment Data without
purpose
Business/data
strategy aligned
Data with
purpose
Data for
reporting
26. User A - Booking probability 0.32
User A has already made a booking in the past 3 days - 2nd biggest
negative factor for this user’s low predicted score.
In today’s visit, user A didn’t choose “windscreen” as the damage type. Since a fix is
not urgent, this user is probably not in a hurry to make an appointment in the next 7
days. This factor has the biggest impact on why the model gives user A a low score.
Model explainer
- Why did user A get a low score?
29. Media.Monks Proprietary & Confidential 29
Media.Monks Proprietary & Confidential 29
Do you really need 6
Floodlights for 6 agencies to
cover the same pageview?
“
30. Media.Monks Proprietary & Confidential 30
Media.Monks Proprietary & Confidential 30
Yes, this is a reality
“
31. Media.Monks Proprietary & Confidential 31
Media.Monks Proprietary & Confidential 31
No, you don’t know it’s
happening but it is and you
need to know this and fix this.
“
32. Request
“I need a [insert tag name here] to fire on
[insert conditions here]”
33. Request
“I need a [insert tag name here] to fire on
[insert conditions here]”
Export
Upload GTM container export into BQ.
34. Request
“I need a [insert tag name here] to fire on
[insert conditions here]”
Export
Upload GTM container export into BQ.
Check
Visualise assets by type and function. Query
based on logic criteria. Prevent duplication.
35. Media.Monks Proprietary & Confidential 35
Media.Monks Proprietary & Confidential 35
Roll your own. Maintain your
data collection as part of your
data dictionary. Don’t add
more instrumentation until
you’ve checked to see if it
already exists.
“
40. Media.Monks Proprietary & Confidential 40
Media.Monks Proprietary & Confidential 40
Make sure your
arbitrary definitions of
what metrics mean are
well understood. Share
the definition.
“
41. Media.Monks Proprietary & Confidential 41
Media.Monks Proprietary & Confidential 41
Avoid arbitrary
definitions if possible.
Be normal. Use
sensible & standard
metrics please.
“
50. Media.Monks Proprietary & Confidential 50
Media.Monks Proprietary & Confidential 50
Production has grown from 20
million tonnes to 350 million
tonnes in 50 years
“
54. Media.Monks Proprietary & Confidential 54
Media.Monks Proprietary & Confidential 54
50% of greenhouse gas
emissions are due to livestock
production.
1 burger = 3000L of water = 2
months of showers
“
Agriculture consumes about 70% of fresh water worldwide;
~1000 L of water are required to produce 1 kg of cereal grain
~43,000 L to produce 1 kg of beef.
https://academic.oup.com/bioscience/article/54/10/909/230205?login=false
55. Media.Monks Proprietary & Confidential 55
Media.Monks Proprietary & Confidential 55
You want to reduce the
carbon footprint of your
food? Focus on what you eat,
not whether your food is local
“
68. ● KPIs that show change - not absolute values
● Trends - directionality
● Segment or die
● Channel-to-site-to-outcome - end-to-end
● Explanations
○ Why > what Context is king
● Actions
○ What? So what? What next?
○ Avoid “needs more data/analysis”
69. Don’t overload it
You can’t fit all metrics
in. You don’t need to
and readers don’t want
you to.
● 4-6 KPIs
● Only what’s
changed relative
to a target.
Dashboards in summary
Analyse for the reader
Explain changes in data
relative to targets, trends and
business context.
Recommend actions.
This is not Google Analytics
real time reporting - there’s a
tool for that. Don’t reinvent
the wheel.
Drive change
Answer questions,
provide direction.
Deliver dashboards inline
with the frequency and
cadence of data
collection.
70. Don’t just take my word for it
Storytelling with data The BIG BOOK of
Dashboards
Let’s practice
77. What?
US visitors to UK site but no conversion as
there’s no US delivery from the UK
78. What?
US visitors to UK site but no conversion as
there’s no US delivery from the UK
So What?
This is unexpected. It shows good SEO signals
for the UK content.
79. What?
US visitors to UK site but no conversion as
there’s no US delivery from the UK
So What?
This is unexpected. It shows good SEO signals
for the UK content.
What now?
Update and upgrade recipe content in the US
81. Half of US consumers accept
all cookies despite concerns
about how their data is
shared.
“
The Drum
Nov ’21
https://www.thedrum.com/news/2021/11/17/half-us-consumers-accept-all-cookies-despite-concerns-about-how-their-data-shared
82. Media.Monks Proprietary & Confidential 82
Media.Monks Proprietary & Confidential 82
So, what are they afraid of?
They aren’t exactly sure.
“
The Drum
Nov ’21
https://www.thedrum.com/news/2021/11/17/half-us-consumers-accept-all-cookies-despite-concerns-about-how-their-data-shared
83. Media.Monks Proprietary & Confidential 83
Media.Monks Proprietary & Confidential 83
76% of consumers say they
mistrust companies to
protect their personal data
and privacy online
“
84. Media.Monks Proprietary & Confidential 84
Media.Monks Proprietary & Confidential 84
76% of consumers say they
mistrust companies to
protect their personal data
and privacy online
75% of people say they only
want to see ads that are
relevant and useful to them
(BCG & Google)
“
85. Brand trust is the second-most
important purchasing factor
for brands across most
geographies, age groups,
gender, and income levels.
Privacy & 1P Data
Data Consultancy
Data driven Privacy
Centric
Measurement
Brand Trust
Data-driven paid media ads
typically unlock 20% efficiency
gains; 20-30% effective gains for
new acquisitions.
Maximise Observable Data
GA4, Consent Mode, Enhanced
Conversions, Server Side GTM,
Server to Server, Consent
Management Platform
Integration, Cookie Banner
Optimisation
86. Media.Monks Proprietary & Confidential 86
Google research has shown that providing
a positive privacy experience increases
share of brand preference by 43%
A negative privacy experience is almost as
bad as a data breach…
First Class
First Party
Privacy First
87. Media.Monks Proprietary & Confidential 87
Media.Monks Proprietary & Confidential 87
Authenticity
“What hinders me is
technical jargon, lengthy
explanations, and
companies possibly not
being upfront about what
they intend to do with my
info.”
UK Customer
Trust
The impact of a poor
privacy experience is
nearly as bad as a data
breach. 39% said they
would switch brand
loyalties in response to a
negative privacy
experience with their
preferred brand.
Spam
Unless I was absolutely
interested in buying an
item I would probably skip
a monetary discount offer.
…concerned about adding
to the many emails I
receive daily.
Poor experience impact
Quantify the
impact of privacy
success and
failure.
Convince the
skeptics with hard
facts, numbers,
reality rather than
a hunch.
“
88. Media.Monks Proprietary & Confidential 88
71%
People prefer to buy from brands that are HONEST about what data they collect and why.
Research participants who were more aware of how data sharing works were
26% more likely to agree that data sharing in return for more relevant ads
represents a fair value exchange.
https://www.thinkwithgoogle.com/intl/en-gb/future-of-marketing/privacy-and-trust/research-customer-privacy-practices/
89. Media.Monks Proprietary & Confidential 89
Media.Monks Proprietary & Confidential 89
The good news for marketers
is that nearly half (44%) of
consumers say they’d be
somewhat or very
comfortable sharing their
personal information in
exchange for a discount or
special offering from a brand
they liked.
“
The Drum
Nov ’21
https://www.thedrum.com/news/2021/11/17/half-us-consumers-accept-all-cookies-despite-concerns-about-how-their-data-shared
90. Media.Monks Proprietary & Confidential 90
Media.Monks Proprietary & Confidential 90
Consumers
reward
brands who
get this right
Personalise
More likely to
consider
purchasing
Consumers
more likely to
recommend to
friends and
family
More likely to
consider
repurchasing
1
3 2
https://www.mckinsey.com/business-functions/marketing-and-sales/our-insights/the-value-of-getting-personalization-right-or-wrong-is-multiplying
78%
76%
78%
91. Any privacy journey needs to
start with a deep understanding
of what people want.
The future of marketing will be
shaped to a large extent by
these win-win exchanges.
“
93. Media.Monks Proprietary & Confidential 93
Media.Monks Proprietary & Confidential 93
Verify
- Be certain you understand what the data means
Calibrate
- Is the data representative of reality?
Prioritise
- Which data points REALLY matter?
Apply governance
- Maintain what is collected and why, for what purpose
Dedupe - DCCO - Data Collection, Cleaning AND Organisation
Frame questions to provoke promote action
Visualise to facilitate action
Don’t just report. Analyse to guide action
- what, so what, what now
“
94. Media.Monks Proprietary & Confidential 94
Media.Monks Proprietary & Confidential 94
“ Failure is expensive, and inefficient
Success is highly rewarding - for us all
We’re not anti-business, we’re pro-consumer
95. Media.Monks Proprietary & Confidential 95
Questions?
Ask your favourite monk
doug.hall@mediamonks.com
Sources:
https://ourworldindata.org/
https://www.cowspiracy.com/facts
BCG & Google, 2022;
EY Global Consumer Privacy Survey, 2020
Google research privacy practices
https://www2.deloitte.com/uk/en/insights/focus/future-of-
mobility/electric-vehicle-trends-2030.html
https://www.climate.gov/media/12886
https://www.linkedin.com/posts/david-smallbone-
7396831_infratech-analytics-dashboards-activity-
6996457350861594624-za0Q
https://ourworldindata.org/food-choice-vs-eating-local
https://ksdigital.co/the-ga4-migration-monitoring-dashboard/
https://nutritionfacts.org/