|KDnuggets : News : 2009 : n10 : item3||PREVIOUS | NEXT|
Subject: Data Mining at 1-800-Flowers
Here is my interview with Aaron Cano, Vice President of Enterprise Customer Knowledge at 1-800-Flowers, the world's leading florist and gift shop, about data mining and flowers.
Many thanks to SAS for arranging this interview.
Gregory Piatetsky-Shapiro: Can you share with us some interesting examples of how you use data mining?
Aaron Cano: We have SAS Enterprise Miner which the life blood of what goes on.
In the customer knowledge group, it really helps us understand who the customers are, and more importantly - what are the drivers of customer behavior and what motivates their transactions with us.
We have a lot of history, we have 12 different brands, including the Popcorn Factory, Cheryl&Co. (cookies), Fannie May (chocolates), Harry London (chocolates), ...
We have transactions and data for all these brands, and we do overlays.
The important thing for us is how we understand the cross-channel, and cross-brand marketing opportunities, and the data mining tool allow us to really hone in on these opportunities.
Consider for example Cheryl&Co. We can go into our enterprise-wide database and try to find prospects who look like the Cheryl&Co. best customer.
GPS: What brand you have most success cross-selling with flowers?
AC: usually Cheryl&Co., The Popcorn factory, and Fannie May chocolates. Chocolates, popcorn, and cookies are where we had a lot of success with flower buyers buying from other categories.
One of the major objectives is to have gift buyer also buy other things that you can also send to your loved ones.
GPS: You probably do customer segmentation?
AC: We have gone down to the level of personas; that helps associates better understand who the customer is. Some of the more important work we are doing is educating the business units on who the customer is.
What their likes and dislikes are; put them in groups.
From a marketing perspective we do a lot of segmentation Traditional RFM does not work that well.
But we can differentiate who the best customers are, who the good customers are, and we differentiate their likes and dislikes by products, and by occasion, among other categories.
GPS: What would you recommend to such a customer?
AC: A large part of our business is same day or next day delivery.
So if a customer calls in the morning and says it is my wife's birthday, we can get her a gift by the afternoon.
We also have a reminder program for this purpose - to send emails reminding people of the birthdays coming up.
GPS: I saw on the SAS page on success with 1-800-Flowers that you were talking about real-time targeting. Tell us more about it.
AC: We are in development to have a real-time "manager".
The idea is to leverage customer data in real time, so when a customer comes in by whatever channel, and by knowing who they are and their past experience, we can differentiate their current experience based on this knowledge.
To give you an example, if a person only buys roses for his wife, and he does it 5 times a year. When this person is on the web site, only show him a banner for roses. It would be best not to distract him with other ads.
GPS: What about local information? Flowers are local - are flowers available in California different from ones available in Vermont?
AC: We do a lot of analytics around geography. Some of our customers will come to make a purchase, and send it around the block.
We do a lot of local connections, but also a lot of long-distance connections.
I may send a gift to Cincinnati one day, and next day to my neighbor.
You mentioned merchandising differences. Most of our products are available throughout the country. We source products all over the world, from Holland, from California, from Florida, from South America. It does not matter where the recipient is, we can send what the customer wants.
GPS: How do you integrate across multiple channels (web, phone, ...)
AC: We provide the enterprise a full customer view.
We also do analytics at the channel level as well - email vs. portal vs. direct mail vs. TV/radio. More importantly, we look at the value of the customer. Regardless of the channel they came in, we provide them the means to purchase what ever they are looking for, and we offer so many channels for the customer convenience.
GPS: I assume that you are quite satisfied with SAS currently. Looking forward, what additional improvements you want from SAS or other analytics tools to keep improving your business?
AC: We talked about real time targeting - this would be a great thing for analytics. Further down the road, we want to know how to integrate the web logs and the customer data.
How do we combine click-stream data and the customer data? It is something we hope to use in the future.
GPS: Can you give an example of a real time targeting you are looking for?
AC: Every customer has a different need every day. Say, last week I bought something for my wife's birthday, and this week I am looking for a thank you gift for my neighbor.
We need to change our engagement style based on what the customer is looking for. Because I bought a birthday gift last week, don't assume that I am looking for a birthday today.
By taking click-stream data and marrying it with customer data, I can get much better data for my data mining tools, and I can much better understand who my customer is and what they want.
GPS: What do you like to do when away from a computer?
I would like to say working on the house, but it is more of a necessity; it is not really what I like to do.
But I do enjoy A quick trip to the beach, going to a baseball game, spending time with my family.
GPS: What are the most popular things, flower arrangements people buy for mother's day?
AC: Roses, and rose bouquets, tulip baskets. The interesting thing is that people come in and purchase for multiple mothers. The celebrate their mother, their wife, their grandmother - all the different moms in their life.
|KDnuggets : News : 2009 : n10 : item3||PREVIOUS | NEXT|
Copyright © 2009 KDnuggets. Subscribe to KDnuggets News!