Personalized Recommendations
Systems
Overview
Industry research indicates that successful online retailers,
such as Amazon, are generating as much as 35% of their business from
recommendations. This means that for many companies, sales are being lost
every day.
More and more, customers are expecting to be able to personalize
their shopping experience beyond wish lists. Netflix and Amazon are great
examples of how to companies can do this.
The race is on to personalize, because once someone invests the
time to train a web site to learn their preferences, they will prefer to
shop there versus somewhere else. Imagine how much harder it will be for
Blockbuster to steal Netflix customers, once they have given three months
worth of feedback.
Shoppers can tell when recommendations dont make any sense.
When this happens an opportunity is lost.
Recommendations that Drive Sales (Creating a truly personalized
shopping experience)
Other systems promise to provide recommendations that
- Are relevant
- Lead to discovery of new products
- Engage the customer
But, ours are more relevant because, we
- Use better data (feedback from your customers and transactions)
- Apply leading-edge methods (we statistically map your brands in
multi-dimensional space)
Shoppers will discover new products if they trust and use a great system
So, we engage the customer by
- Collecting product ratings
- Using scrollable recommendations
The better the data, the better the recommendations. Its
that simple.
Product Ratings Data: Since people arent always
happy with their purchase, customer feedback improves the quality of
recommendations. Collecting this data is also key to building a relationship.
Sales Data: After ratings data, purchase data is the most useful.
Removing gift purchases ensures that matches are based on the customers
true personal tastes.
Click Data: Widely available and used only as a last resort, click
data provides little guidance as to whether the shopper will like a product
enough to buy it - let alone be satisfied with it. Strategically, it does
nothing to build a relationship.
Philosophy: Collect the best data, derive a custom solution, then
check the results by hand to make sure they make sense. Once foundation is
set, profiles update in real time.
Leading-Edge Methods
The Also Bought method shows whats popular, not whats
most relevant. We can do better
Our approach is different from what you see on other web sites, because we
map brands in multi-dimensional space.
This provides extraordinary accuracy with recommendations, not only for new
shoppers, but especially for returning customers.
Also has implications for pricing strategy and product line opportunities.
Strategy
Strategy
New Shoppers
Product Pages Scrollable, community recommendations
Checkout Accessories, such as batteries
Search Displays everything the customer needs (not just list of products).
Sort by popularity, price, etc.
Repeat Buyers
Home and Product pages Scrollable, personalized recommendations.
Profile Building: Collect product ratings
Smart Search User preferences influence results
Clients Customize Recommendations by
Profit margin
Inventory availability
On-sale items and other criteria
space
Getting Started (Its easier than you imagine)
Provide data to Dart for analysis
We use it to build a demo
You evaluate it and when ready to implement
Insert the ready-to-use code provided by Dart
No software hassles
Four web-based patents
Other uses for customer profile data
My-Gift Store (Recommend Gifts)
Recommendations for when bill-to and ship-to dont match
Email gift reminders with recommendations
Encourage customers to add new gift recipients
New reasons to email
Post-sale requests for product ratings
Personalized promotions
Telemarketing Up-Selling
Direct Mail: Highly recommended products
Catalog Merchandising
More About Personalized Recommendations
Slideshow /
Brochure (pdf) /
Compare Results (pdf)
DVD Demo -
Judge for yourself. Select from thousands of DVDs!
Coffee Demo
- See how even new shoppers can influence
recommendations. Features unique brand map.
DART
Marketing, LLC |