dafiti and jetlore present at fashion digital, ny, 2015
TRANSCRIPT
The Latin King of Product Content: How Dafiti Grew into a Top Retailer by Providing Fresh, Engaging Content-Driven Experiences
Romulo Zarelli CostacurtaHead of CRM, Dafiti
Eldar Sadikov
CEO, Jetlore
About Dafiti
‣ Latin America’s largest fashion online retailer
‣ >$200M in online sales annually
‣ One of the fastest growing e-commerce
companies in Latin America
‣ Operates in 5 different countries with different
currencies and languages
‣ >100,000 SKUs
‣ Includes now 3 other ventures: Dafiti Sports, Tricae, and Kanui
‣ Accessible by means of all device types
Dafiti CRM program
‣ Email newsletters (along with triggers) are
critical for customer retention and activation
‣ High customer engagement in email has a
dramatic impact on customer LTV
‣ User engagement/activation in email is usually
accomplished through category-specific
promotions and discounts
‣ Heavily relied on segmentation to ensure
featured promotions were relevant to users
‣ But it's not always easy for the customers to
discover products via promotions and price is
not the decision driver for many customers
How can we make email content exciting and fresh everyday?
‣ We realized that product catalog is our best
content!
‣ But hand-picking product is painful and
suboptimal: different users want to see different
products
‣ Putting a carousel of last purchase or cart
abandonment recommendations may not reflect
the current fashion trends or customer’s tastes as
time goes by
‣ How can we create an engaging non-repetitive,
yet personalized product discovery experience?
We added a “product discovery” section to our email
‣ The product collection is algorithmically merchandised at
email open time
‣ Takes into account:
- historical purchases
- browsing behavior
- passive observations (what products were displayed before)
- catalog product performance
- customer context
‣ The collections adaptively iterate as Jetlore learns more about
the customer without the need for constant manual updates
to product selection strategy
Jetlore’s logic ensures a fresh, diverse and relevant collection to explore every day
Felipe has been browsing shorts and summer wear leading
to shorts being featured top left
Felipe has been browsing shorts and summer wear leading
to shorts being featured top left
Jetlore’s logic ensures a fresh, diverse and relevant collection to explore every day
Felipe didn’t click on Thursday’s Email, leading to exploration of casual shoes, a prior purchase
category. Hooded sweatshirts were also
trending.
Jetlore’s logic ensures a fresh, diverse and relevant collection to explore every day
Felipe didn’t click on Thursday’s Email, leading to exploration of casual shoes, a prior purchase
category. Hooded sweatshirts were also
trending.
Felipe browsed sweaters and backpacks Saturday morning before opening the email, leading to their
inclusion that day.
Results:Revenue Per Opened Email
14.6% 51.3%
All Days (including dayswith discounts on all products)
Low/No promotions (one third of all days)
vs
Click to Open Rate (CTOR) Higher and Steadier (11% )
With the Jetlore solution, Dafiti experienced an 11% increase in their click to open rate for promotional newsletters. This increase held steady throughout the measurement period and overall the CTOR experienced much less volatility than the control group.
This steady engagement not only improves short-term sales results, it also increases the long-term health of the email channel and overall customer retention.
“We like to call it ‘remarketing’ because it’s so successful in
activating newsletter subscribers” – Dafiti CRM Team
Click to Open Rate: Higher and Steadier with Jetlore
ControlJetlore
Results: Click-to-open Rate is Higher and Steadier
11% CTOR
How were these numbers and consistency achieved?
1. Jetlore ensures diversity of featured items spanning multiple product categories
vs
This exposes users to a
variety of items and enables
exploration (studying user
behavior to see if they react
to the items)
2. Jetlore keeps product content fresh and constantly evolving
‣ Standard recommendation systems pivot
off the last shopping cart item or last
purchase and repeat the same products.
‣ Jetlore’s engine dynamically adjusts to the
user behavior:
- if the user does not take action, it will iterate to offer new categories and items
- if the user takes an action, it quickly reacts by featuring more from that class of items
3. Jetlore immediately incorporates new items and trends
‣ Typical recommendation systems require a
bootstrapping time to start featuring each
item and, as a result, tend to over-focus on
the top/older items
‣ Jetlore immediately incorporates new
items without any bootstrapping time
‣ Jetlore automatically incorporates fashion
and seasonal trends via real-time
monitoring of product performance
Felipe’s profile
Shorts Summer wear
Backpack
Shoe
Ralph Lauren
Lacoste Casual
Navy BlueStripes
Ecco
Nike
Nike BackpackBlue
New Arrival
What’s next?
Discovery is not limited to products. It applies to promotions as well!
‣ In any given 2-week timeframe, we have 50-100
promotions created by Dafiti’s marketing team
‣ Each promotion is associated with a sub-catalog of
items (50-1000 items in each sub-catalog)
‣ Each user has a unique affinity to buy from a specific
sub-catalog
‣ Jetlore will select for each user the most relevant
promotions given the user profile at that point in time
- ensuring that promotions are not repetitive
Personalized discovery of promotions and products across all channels
Web: Homepage Mobile App: Home ScreenEmail: Newsletters
Thank you! Questions?
Content in the form of promotions and products is your best weapon. Use it to create personalized engaging discovery experiences.