a presentation by W H Inmon

1
BUSINESS ADVANTAGE IN THE EBUSINESS ENVIRONMENT
a presentation by W H Inmon
2
web site
web manager
web site data
data marts
cif
granularity manager
exploration warehouse
ODS
edw
analytical processing
near line/ secondary storage
3
web site
web manager
web site data
marketing to the mass market
a macro view
a micro view
exploration warehouse/ data mining
marketing to the individual
ODS
edw
4
continuous update
profile records for an individual customer
detailed historical records about activity
analysis done on an as needed basis - hourly
- daily - weekly
5
whats in the profile -
profile record
- cookie - voluntary information - when interest
shown - what bought - what browsed but not
bought - amount of purchase - style of purchase -
how much bought
6
an old style profile record -
profile record
- name - age - gender - occupation - own/rent -
salary - education - marital status - address
7
comparing an old profile record to a new profile
record -
old
new
- name - age - gender - occupation - own/rent -
salary - education - marital status - address
- cookie - voluntary information - when interest
shown - what bought - what browsed but not
bought - amount of purchase - style of purchase -
how much bought
8
profile record -
example - who do we sell mens clothing to?
- men or - cookies that have shown interest in
mens clothes
9
web site
web manager
web site data
using the micro profile record
1 - when the cookie has reentered the web
site 2 - for promotions give me all cookies
who have shown an interest in
Porsches
10
the profile record is created as a result
of sequences of transaction records that
enter the data warehouse
11
what can you use the profile record for?
cross selling - if they like Porsches they might
like Ferraris up selling - if they buy real
estate, they may need a home loan promotions -
give me all stamp collectors who like US
stamps buying habits - some people buy on week
ends, some people buy
large/small items selection - this item is
browsed but not selected this
item is not being browsed at all
12
web site
web manager
web site data
exploration warehouse/ data mining
the other perspective is from the market place
13
web site
web manager
web site data
the results of exploration and data mining go to
management and are used for many kinds of
decisions
14
exploration processing -
- what price ranges are selling/not selling this
Christmas season? - are more sales being made
this year than last? in what categories? by how
much? - what types of items are popular this
year? - is there a regionality of sales?
15
exploration processing -
looking across -
- time - classes of items - sizes of sales -
timing of sale - effect of promotions on sales
16
exploration processing -
price elasticity -
3,250 units sold at 1.25 3,497 units sold at
1.59 3,335 units sold at 1.75 3,402 units sold
at 2.00
175 units at 225 105 units at 279 76 units at
349 28 units at 425
17
price elasticity -
detailed, historical data
based on past sales, on an item by item basis
- when the price was raised, how many units were
sold when the price was lowered how many units
were sold
the result is the calculation of the price
elasticity ratio for any given item
18
exploration processing -
yield management -
Los Angeles to New York - Jan 25

Jan 15 - 980 Jan 17 - 852 Jan 19 -
783 Jan 21 - 1050 Jan 23 - 1156
19
yield management -
historical, detailed data
the history of flight sales is kept
based on the history of flight records, the
average number of seats sold as of some moment in
time is calculated
flight 223 - July 20 - Dallas to Denver avg seats
sold - July 10 - 72 avg seats sold - July 11 -
78 avg seats sold - July 12 - 81 avg seats sold -
July 13 - 84 ...
20
yield management -
flight 223 - July 20 - Dallas to Denver avg seats
sold - July 10 - 72 avg seats sold - July 11 -
78 avg seats sold - July 12 - 81 avg seats sold -
July 13 - 84 ...
based on flight and actual seats sold versus
average seats for the date, the price of the
ticket is raised or lowered
21
exploration processing -
customer loyalty -

loyal customer
disloyal customer
22
customer loyalty -
detailed, historical data
1 - who has left the company? who has
stayed? 2 - what characteristics does each
group have? 3 - based on characteristics,
create a profile

loyal customer
disloyal customer
23
customer loyalty -

cust id name address status classification .
4 - the profiles are used to go against the
existing customer records. Each
customer is classified
24
customer loyalty -
once classified, the customers are
treated differently, according to classification
- promotions - special treatment - discounts
25
detailed, historical, integrated

the intelligence to maximize business opportunity
once connected
the ability to connect
web data warehouse unparalleled
business opportunity