4. Let’s talk clustering!
• Your customers are different à Clustering reveals these different
personae
• Understanding clusters à personalization
• Personalized campaigns à better marketing ROI
5. What is customer clustering?
Athletes
Chronic Returners
Doting grandmothers
5
Creating homogenous groups of customers and identifying
personae to drive marketing actions
6. Then: Broad strokes
Basic retro-active
segmentation
Now: Clear picture
Clustering drives relevancy &
personalization
•
• What products do they
need?
•
12 month file versus 12+
month file
•
What brands do they
prefer?
•
Buys best selling brands
•
What offers will compel
them to buy?
•
6
1x buyer versus 2x+
buyers
Has loyalty
membership?
•
What’s the right channel
& frequency of contact?
7. 3 clustering models available
• Behavioral: Spending, channel, order interval, discounts, returns,…
• Product based
• Brand preference
• Clustering done mathematically via unsupervised learning
•
•
•
•
•
Let data drive what clusters you have in your customer base
No guesswork or leaning on intuition
Multi-variable, not based just on revenue, or a single product
Optimized for stability
Data refreshed daily
8. Example: behavioral cluster DNA
Long term, frequent buyers,
medium sized orders
High value, fewer orders,
big spend on 1st order
$99 average order
$2,261 total revenue
24 days between orders
24 total orders
57 total items
$76 first order revenue
1.7 products in first order
6% of orders on clearance
+10 more
$124 average order
$595 total revenue
67 days between orders
5 total orders
14 total items
$164 first order revenue
3.3 products in first order
3% of orders on clearance
+10 more
9. Use the DNA to personalize
Long term, frequent buyers
medium sized orders
High value, fewer orders,
big spend on 1st order
Goal: increase order size
(AOV)
Goal: decrease time to rebuy
Campaign: Give double
points if they spend 50%
more than usual
Campaign: Give double
points that expire sooner
than usual with purchase
Result: 125% increase in
AOV
Result: Time to next
purchase reduced to 2
months from 4 months
10. Example: product clusters
Product Clusters Developed
Cluster #1: Sweaters
• Based on what products customers
buy
Cluster #2: Stylish Men’s Wear
Cluster #3: Active Wear
Cluster #4: Gift Certificates/Family
Cluster #5: Elegant Ladies
Cluster #6: Kids Wear
Cluster #7: Underwear
Cluster #8: Accessories
10
4 March 2014
• Once clusters are created, this
reveals the natural groupings of
products
12. Example: Brand clusters DNA
Cluster 1 Brand Scale
Least Interest
Pleasure Doing Business
Wow Couture
Desigual
6126
L*Space
Preferred Brands
Tahari Arthur S. Levine
Calvin Klein
Eliza J
Adrienne Vittadini
Nine West
Preferred Brands
Cluster 3 Brand Scale
Least Interest
Collective Concepts
Wow Couture
Max and Cleo
Rene Rofe
Steve
Desigual
Dzhavael Couture
Custo Barcelona
Smash Wear
Salvage Rock’N’Rebel
13. DNA for each cluster drives relevancy in
campaigns
Cluster 1: Sweaters
Cluster 3: Active Wear
Goal: Drive customer engagement
Goal: Increase customer value
Campaign: Target cluster with all main
product related campaigns (E.g. New
sweater arrivals)
Campaign: Offer bundles of related
products from same cluster at discount
Result: 100% increase in AOV
Result: Improved email open rate by
25% and click rate by 66%
IN THIS video, I will talk about customer segmentation, what it is, when and how it is used, what are different types of segmentation and what to watch out for when segmenting customers.