Will this slight
discrepancy
dramatically
shift budgets or
efforts? NO!
Slide 11
Slide 11 text
Transactional
data must be
100% correct
but analytical
data doesn’t.
DATA IS
FUZZY
Slide 12
Slide 12 text
You need to be
pragmatic
02/
Slide 13
Slide 13 text
Gain actionable insights,
even if the data isn’t perfect.
Slide 14
Slide 14 text
Pizza Meter
In the 90s, Domino’s pizza orders
would increase during heightened
Pentagon activity (political crisis or
military operation).
Slide 15
Slide 15 text
Your sample
size must be 70
times larger
than the
number of
variables
SARAH
CROOKE
QUOTE
Slide 16
Slide 16 text
Data
Pragmatism
Actionable insights over
theoretical perfection
Slide 17
Slide 17 text
Data
Pragmatism
Tailoring data analysis and
models to meet specific
business needs
Slide 18
Slide 18 text
Data
Pragmatism
Data is presented in a way that
triggers action in stakeholders
Slide 19
Slide 19 text
Spot Trends
No need for perfect
data
Directional
reporting informs
decision-making.
Slide 20
Slide 20 text
You Rely on Trend Analysis Already...
Glimpse
Real-time data from
multiple sources.
Exploding Topics
AI scraping + trend
analysis.
Google Trends
sample of web
searches.
Slide 21
Slide 21 text
YoY in Google Trends
Slide 22
Slide 22 text
Semrush
Enterprise has a
“What Has
Happened”
Trend Analysis
Solution
Slide 23
Slide 23 text
AlsoAsked.com
monitors intent
shift trends in
PAAs
Slide 24
Slide 24 text
Trend Analysis
Crash Course
03/
Slide 25
Slide 25 text
3 main types of trend analysis
Comparative analysis
trends across
categories
Regression analysis
relationship
between variables
Time-series
data points over
time
01/ 02/ 03/
Slide 26
Slide 26 text
Emerging keywords Search intent changes Keyword popularity by
region
Seasonal keywords Search volume correlations Intent proximity
Keyword Trendspotting
Slide 27
Slide 27 text
“With Revolut Bank,
trending queries show
that folks inquire about
its trustworthiness and
legitimacy.”
MARK
WILLIAMS-
COOK
Slide 28
Slide 28 text
Create Your
Own Metrics
03/
Slide 29
Slide 29 text
Spotify has 6,000
music genres.
01/
For optimal
recommendations
02/
Using meta data,
raw audio data &
listening patterns.
03/
Slide 30
Slide 30 text
To the Spotify algorithm, an artist's genre is
more of a "cluster of collective listening
patterns" than a traditional label.
Slide 31
Slide 31 text
Spotify’s algorithm uses those
labels
to map out a network to figure
out other genres
you are most likely to engage with.
Slide 32
Slide 32 text
Focus on metrics that align
with your company's unique
value and goals.
Slide 33
Slide 33 text
Occupation
Rate
Tracking rented units in GA4 as
a metric to optimize Google Ads
& to check vs indexation
Slide 34
Slide 34 text
Core Web Vitals metrics are
weighed by Google.
Slide 35
Slide 35 text
CLS is the one important
metric for Google Ads
Slide 36
Slide 36 text
Create a modified weighed system
to reflect a combined SEO & PPC baseline
Slide 37
Slide 37 text
How to get
started
04/
Slide 38
Slide 38 text
Change the
conversation,
not the data.
Slide 39
Slide 39 text
Blueprint
Articulate the challenge
1.
Define Custom Metrics
2.
Assemble the data sources
3.
Identify patterns
4.
Communicate trends &
drive action
5.
Slide 40
Slide 40 text
#1 Articulate the challenge
How frustrated are our
customers?
Ecommerce example
Slide 41
Slide 41 text
#2 Define Custom Metrics
Slide 42
Slide 42 text
Digital Happiness Index
Page load speed
(fast)
01/
High value orders
placed
02/
Return purchases
03/
Slide 43
Slide 43 text
Digital Happiness Index
Add to wishlist
04/
Five star
reviews
05/
Slide 44
Slide 44 text
#3 Assemble the data sources
to mitigate their inherent limitations
Slide 45
Slide 45 text
Identify patterns
Slide 46
Slide 46 text
#5 Communicate Trends & Drive Action
Improving page load
speed is crucial.
It impacts the
happiness index.
It's linked to CX &
sales.
Slide 47
Slide 47 text
Google Ads
GSC data
CRUX data
Google Merchant
Center data
My current
hyperfocus is
trend analysis in
BigQuery.