Business objects SET ANALYSIS 6.5.1 User Manual

Set Analysis User’s Guide

Set Analysis 6.5.1

2 Set Analysis User Guide

Copyright

Trademarks

Use restrictions

Patents

Copyright © 2004 Business Objects. All rights reserved.

Business Objects, the Business Objects logo, Crystal Reports, and Crystal Enterprise are
trademarks or registered trademarks of Business Objects S.A. or its affiliated companies in the
United States and other countries. All other names mentioned herein may be trademarks of their
respective owners.

This software and documentation is commercial computer software under Federal Acquisition
regulations, and is provided only under the Restricted Rights of the Federal Acquisition
Regulations applicable to commercial computer software provided at private expense. The use,
duplication, or disclosure by the U.S. Government is subject to restrictions set forth in
subdivision (c)(1)(ii) of the Rights in Technical Data and Computer Software clause at 252.227-

7013.

Business Objects owns the following U.S. patents, which may cover products that are offered
and sold by Business Objects: 5,555,403, 6,247,008 B1, 6,578,027 B2, 6,490,593, and
6,289,352.

Set Analysis User Guide 3

Contents

Preface Maximizing Your Information Resources 5

It’s in the documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Reader’s comments form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

Chapter 1 Sets 9

What are sets? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Sets view . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Set building steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Visual data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Visual sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Freehand SQL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Event analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

Import a file . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

Set membership . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

Set metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

Chapter 2 Export List 45

Export lists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

Define a list . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

Chapter 3 Set Analysis 53

Interactive analytics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

Membership analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

Set metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

Migration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

Contents

4 Set Analysis User Guide

Chapter 4 Individual Analytics 71

Chapter 5 Appendix 87

Chapter 6 Index 91

Individual portrait . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

Membership . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

Metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

New Sets Rebuild Engine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

Contents

Maximizing Your Information Resources

preface

6 Set Analysis User Guide

It’s in the documentation

Online guides

Application Foundation documentation strives to deliver product information that
is rich, convenient, and easy-to-use.

Whether you’re a novice or experienced user, Application Foundation
documentation is the place to go for discovering our products, exploring their
features, or locating precise information.

Product information has been substantially expanded to encompass not only
facts about product features but also tips, samples, and troubleshooting
instructions.

For your convenience, our documentation is available from all products in
Acrobat PDF, and print media.

Documentation has been designed first and foremost with speed and ease of
navigation in mind. All the information you require is readily available just a few
mouse clicks away.

The next sections highlight new and key features of our documentation.

!

User’s guides

All Application Foundation user’s guides are available as Acrobat Portable
Document Format (PDF) files. Designed for online reading, PDF files enable you
to view, navigate through, or print any of their contents.

From a PDF file, you can search for specific occurrences of a word using the Find
command, or navigate to the exact location of a topic by clicking a cross­reference or an entry in the Index or Table of Contents.

To open a document, you can select it from the Help menu provided that you
have installed the Adobe Acrobat Reader, version 4.0 or higher on your machine.
You can download it for free from Adobe Corporation’s web site at:

http://www.adobe.com/products/acrobat/readstep.html

What to do fo r m o re in formation

If you cannot find the information you are looking for, then we encourage you to
let us know as soon as you can. Feel free to send us any requests, tips,
suggestions, or comments you may have regarding this or other Application
Foundation documentation using the Reader’s Comment Form at the back of this
guide.

Maximizing Your Information Resources

Audience

Set Analysis User Guide 7

To search for anything specific use the index.
Below is a list of other Application Foundation guides:

• Dashboard Manager User Guide

• Predictive Ananlysis User Guide

• Process Analysis User Guide

• Application Foundation Installation Guide

• Application Foundation Configuration Guide

This guide is intended for users that want to create sets that allow you to define
complex selections and custom groupings of customers, products or other
individual items in your database.

It’s in the documentation

8 Set Analysis User Guide

Reader’s comments form

Set Analysis User Guide
Version 6.5.1
Part Number: 3C1-50-300-01

Your company’s name: _______________________________________
Company address: __________________________________________
Telephone/fax number: _____________ E-mail address: ____________
What is your job position? _____________________________________
How long have you been using Application Foundation? ________________

Business Objects welcomes your comments and suggestions on the qual ity and
usefulness of this publication. Your input is an important part of the information
used for revision.

If you find any errors or have any suggestions, please indi cate the topic, chapter
and page number below:

_______________________________________________________________
_______________________________________________________________
_______________________________________________________________
_______________________________________________________________
_______________________________________________________________
_______________________________________________________________

E-mail your comments to:[e-mail]
Mail it to:
Technical Publications
Business Objects Americas
3030 Orchard Parkway,
San Jose, CA 95134

Maximizing Your Information Resources

Sets

1

chapter

10 Set Analysis User Guide

What are sets?

Sets allow you to define complex selections and custom groupings of customers,
products or other individual items in your database. They are represented as lists
of unique identifiers and are stored in table alongside your source data. For
example, you may have a list of all your clients in a main database and source
data that indicates other characteristics and recent purchasing behavior. Sets
allow you to identify and target groups of these clients based on their meeting or
not meeting criteria that source data allow you to specify. For instance, you may
want to select all your clients living in a particular geographic area who own their
own homes, are below a certain age, have children and have never purchased a
children’s products from the company.

Each set:

• Contains items based on a single "subject" which determines the primary key

• Is uniquely identified by its name, folder and author

• May or may not have a description

• Has a last modified date

• Is allocated to a folder

• Contains a count of its members, only if the set has been built. Otherwise it

just has a definition.

• Contains members that can record an accumulated score based on user

defined weightings.

• Can be processed from many steps, each step based on a single table or

view.

The ’subject’ defines the central theme for the sets which are based on it. A set
can contain data for only one subject; for example, customers, accounts,
products, etc. In the context of a ’customer’ subject, you can define sets that
include conditions based on the various tables and views that describe the
customer entity.

Sets can be scheduled. Refer to Scheduler in the Dashboard Manager User
Guide for more details.

Sets

Sets view

Set Analysis User Guide 11

You can perform the following functions from Sets view:

• View a list of existing sets

• Delete a set

• Name and describe sets

• Define the current set

• Create a new set

• Create a new Folder

• Define set scope

• Access and create Groups

• Create taxonomic groups and rebuild (processing) groups

Sets view

12 Set Analysis User Guide

Set building steps

Create a new set

Every time a set is created it goes through certain actions before it is processed.
These actions are called steps that are crucial for building a set. You can build a
set in five ways - Visual Data, Visual Sets, Event Analysis, Import, and Freehand
SQL. The next few pages explain the process for each.

!

Select a subject

To choose a subject:

1. Click Subject drop-down. Select a Subject. As you select a subject, all existing

folders, and sets are displayed.

The subject becomes the current subject for all successive set operations.
The current subject is the central theme of the set. A s et contains data for only
one subject; for example, customers, accounts, or products.

To create a new Set within the selected subject:

1. Click Add. Select the New Set option.

Sets

2. As you select the option, the next page displays fields that allow you to enter

all the details for the new set.

3. You have to enter the Name of the new set. Each set must have a name that

identifies it for later operations. The name you choose should be unique to
you within the host folder. If you want to use duplicate names for sets, you
must allocate the sets to different folders.

Set Analysis User Guide 13

4. Select the Folder the new set is going to belong to. Each set belongs to a

folder, which enables sets to be shared among user groups (by giving them
Read Only or Read/Write access) or kept private to the user who created it.

5. Select the Scope of the set which can either be public or private. Choose the

scope for the set. The scope may be Private or Public, which allows you either
to share the set with other users or keep it private. When you create a s et and
allocate it to a public folder, its scope is defined to be public. When you create
a set and allocate it to a private folder, its scope is defined to be private.

6. Add any descriptive notes about the set.

7. Frozen – This specifies that the set cannot be changed or reprocessed, either

by the set owner or by anybody else. The only way to change or reprocess
the set is to uncheck this field. This flag is useful when creating a snapshot
set that must not be reprocessed, for example, a specific mailing campaign.

8. The External check box specifies that this set has been created by an external

action. This is used if set membership is being calculated and managed by
some other application. The other application can create a set and store it
within the repository. An external set can only be created and reprocessed
from the other application, but can be used as a starting point for any other
set.

9. The Terminal check box specifies that the set cannot be used as the s tarting

point for any other set. So it is not possible to use this set within the Visual
sets view.

Set building steps

14 Set Analysis User Guide

If your installation has been set up to use Calendars, you can choose the Type
of set you want to create – static or dynamic. If you choose Dynamic, the
Dynamic Profile tab is displayed. Dynamic sets permit the membership to change
based on changes in the source data. The history of dynamic set membership
can also be preserved and used for time-based analysis.

10. If you are creating a Dynamic set, enter the following details:

- At what time interval the set has to be rebuilt. The drop-down allows you to
select the time interval.

- The type of dynamic set, for example, Snapshots, Temporal or Limited.
Click any one of the option buttons

- Enter the rebuild information. Dynamic sets are typically rebuilt according to
the periods associated with refresh time interval or calendar selected for the
set. The default is to prevent multiple processing in the current period as you
may not normally want to rebuild the set twice in the same period.

- For new sets, the period associated with the first build is called the base
period. For subsequent builds, the most recently ended period is associated
with the set. However, if you check Allow Auto-fill, the associated period
comes after the period for the most recent build. Thereafter, each build
increments the active period for that set. If the s et is dynamic and Allow A uto­fill is checked, the AutoFill button in Sets view is enabled.

Selecting Stop at current period causes normal behavior to be adopted once
the auto increment rebuild period is equal to the current period.

11. Once you have entered all details click Apply Changes to save the new set

defined.

Sets

Visual data

A new set created is empty. An empty set needs to have members added to it
before it can be used. You fill the set in the V isualData view. You can perform the
following functions from the VisualData view:

• Navigate through the database using the predefined object trees to locate the

categories of data required to fill the current set.

• Create your own object trees for data navigation.

• View the contents of the database categories in the detail windows on the

right of the view. These detail windows provide information about the content
of the categories and how they relate to other categories.

• View the components of a category by moving it to the top detail window.

When you move a category to the top window, its sub-categories appear in
the lower window. You move a category to the upper window by double­clicking it in the lower window.

• View the relationship between the upper and lower window categories as

indicated in the Object Tree.

• View the number of members currently in the set.

You can get into this option by clicking the drop-down list to Add a metric and
select the Visual Data option.

Set Analysis User Guide 15

!

Starting to build a set

Once you have created and saved a set, you can add members to it in the
VisualData
listing showing segments for members in the database that you can drill down on.

The Application Foundation administrator creates these default object trees in
Application Foundation Architect.

Create visual set

To create a visual data set:

1. Click Add tab under set building steps and select Visual Data option.

view. The VisualData view Object Tree contains a directory tree-like

Visual data

16 Set Analysis User Guide

Once you have created and saved a set, you can add the data field to it. The
VisualData view Object Tree contains a directory tree-like listing showing
dimensions and measures in the database that you can drill down on. The
Application Foundation Administrator creates these default object trees in Set
Analyzer Architect.

2. As you select Visual Data, the Content page appears.

3. From the list select and add a dimension or measure to visualize your

selection using the arrow button. You can also select by double clicking the
dimension.

REMINDER

Please make sure that you sele ct a count fo r e ac h dim en si on. It i s only then that y ou
can perform an operation on your set.

Sets

4. To remove the dimension or measure, click the left facing arrow key.

Set Analysis User Guide 17

When you select a measure Field Condition window automatically appears.
But if you select a dimension, you have to first check the Yes/No Condition
and click OK to allow this window to appear.

The Field Condition window allows you to define the filter and specific
condition for the dimension. For example, you want to create a set of people
with Age (dimension) that is greater than 45 (filter, condition).

You also have the option of not defining any condition. To do so check the
Yes/No Condition radio box. This deactivates all filters and you can then view
all values for the selected dimension.

If the dimension is Date type, then the field condition window allows you to
enter calendar details.

On the window you can enter whether the date is Relative or Floating.
Floating is when you can specify the floating period.
Relative is when you have to enter the specific date.
If you selected dates with a condition Between then you have to end the

ending period which again can be either relative or floating.
The Description field specifies exactly what you select as the calendar dates.

5. Click the drop-down and select a filter option. The description is automatically

displayed.

6. Click OK and the main page has the filter condition displayed on the se lected

dimension.

Visual data

18 Set Analysis User Guide

7. As you click the dimension, you are prompted with a window asking if you

prefer exact count.

8. Dimensions that have calculative or numeric value can be filtered. To do so,

click the dimension once you have moved it to the set building area of the
page. As you select the dimension, he Fx button appears.

Filter function tab

9. Click Fx tab and a window appears.

Once you have entered all details for the set content, you need to enter the
Attributes.

10.To enter the attributes, click Next. The attributes page appears.

Sets

11.Enter a description of the step to be processed, for reference purposes.

You can allocate a number of points to each step or process operation. For
example, select Tennis Players and award 1 point. Select Contacts in New
York, and award 2 points, and select Contacts spending more than $200 and
award 4 points. Each member can exist only once in a set, but if a member is
selected for more than one reason, then the points allocated are ac cumulated
on a member by member basis. In this example, someone who matches on

Set Analysis User Guide 19

all three conditions accumulates 7 points.

NOTE

The points score must be an integer and can be positive or negative. Limits for
number size vary by d atab as e a nd installation. Typi cally, the points val ue is no t ve ry
large.

12.Allocate points for score carding by using the spinners or typing directly into

the Points box.
Sampling allows you to further refine your selection by defining the exact

characteristics of the output. Normally selection focuses on defining the
criteria that defines your output. Normal selection, however, may produce too
many people for your purposes. For example, what happens if you want only
the top 100 people from the 129 who match on other specific criteria?

If you select from a database of 3 million customers and your selection is
based on certain age, gender, and income groupings, you may produce a
target of 270,320 customers. If you want only 10,000 people, how do you
select these people? If you then want to split them into five cells of 2,000
people in order to test different treatments in order to determine the most
effective treatment to use on the remaining 260,230, then what do you do?

The sampling module is a single interface that provides a mix between
random sampling and the ability to choose data based on ranked attributes.
In addition, it is possible to stratify each selection into groups. The benefit of
being able to stratify or group your data allows you to produce skewed
samples. For example, your data may be based on a 60/40 male/female split.
However, you may want to target 25% male/75% female, which is easy to do
using the sampling module.

Application Foundation provides four specific types of sampling scenario:

- Random sample creates a different result set each time it is created. It
selects 100 people randomly from the sample frame. If you select Random,
you have to check either Sample Size - Shows the numbers that is used to
process the sample - or Sample % of Total - Count as a percentage of the
sample total - check box.

- Ranked selects the top 100 people from the sample frame. Sort on the
column(s) matching the field to be used for the ranking, then select OK. You
can sort the column in either ascending or descending order to specify
whether you choose the top n or bottom n. Descending order produces the
top n.

Visual data

20 Set Analysis User Guide

Creating a sample

- Stratified Random sampling selects 50 males and 50 females randomly
from the sample frame.

- Stratified Ranked sampling selects the top 10 people in each town in the
sample frame.

13.Choose whether you want a random or ranked sample.

In addition to selecting between rank or random, you may also define a
grouping or strata by clicking the In Groups. Selecting a group organizes the
data by the groups you select.

14.If you select Random, click Group tab.

Sets

15.Select the columns you want to group by.

16.Click OK.

A grid listing available strata appears.

Set Analysis User Guide 21

17.Enter the following:

- Source Count Total: Total size of the original sample frame.

- Source Count: The breakdown of the original source, grouped by the
selection you made in the Choices dialog box using In Groups.

- Source %:Grouped source count as a percentage of the total.

- Sample Count shows the numbers used to process the sample.

- Sample % of Source: The sample count as a percentage of the source
count.

- Sample % of Total: Count as a percentage of the sample total.

NOTE

You can change the sample counts. For exam ple, if Unknow n gender is not re levant,
you can enter a value of 0 in the Samp le Count column. You can als o apply a constant
number to the Sample Counts colum n, by entering the number in the Constant no. box
and clicking the > button.

18. Within the grid, type in the counts required, then click OK.

19.If you select Rank, select the ranking.

20.Sort on the column(s) matching the field to be used for the ranking, then select

OK. You can sort the column in either ascending or descending order to
specify whether you choose the top n or bottom n. Descending order
produces the top n.

21.As you complete all the details about the set, click OK. The set built is

displayed.

Visual data estimates

Visual Data estimates allow Set Analysis to use estimates rather than getting
counts from source tables. This is not as accurate as gettin g counts directly from
the source tables, but can be much quicker.

Visual Data estimates use aggregated objects to calculate subsequent counts for
objects at a lower level of an ad-hoc object tree.

Estimates on these aggregated objects differ from standard object trees because
they do not have to be defined using the “Object Tree” creator in Architect. The
counts for the tree are predefined but does not calculate if you add an object to
the predefined tree.

The example shows an ad-hoc object tree of Country, Bank and Gender all of
which have had “Fast Counts” calculated on an individual basis, they are not pa rt
of defined “Object Tree”

Visual data

22 Set Analysis User Guide

If “England” and “Abbey National” are selected for the first 2 levels of the tree,
when “Gender” is expanded, the system recognizes that there is a fast count for
“Gender” and the objects above it and it uses a smart calculation to produce an
estimate for ”Gender”.

In this case, if the source table is queried, counts are returned thus:
Female= 4,456,448; Male = 4,128,768; Unknown = 65,536. These are exact

counts from the source table.

Sets

sIf “Estimates are now used, the counts are returned thus:Female= 4,356,808;
Male = 4,218,497; Unknown = 69,156.

NOTE

The accuracy on calculations for estimates improves if the object being
calculated follows a predefined “Object Tree” e.g If a geography hierarchy with
fast counts has been defined as Country, State, City and the Gender object is
added to the bottom of this hierarchy, the estimate is more accurate than for an
ad-hoc assortment of objects.

Set Analysis User Guide 23

Options can be set by the user that affect the way in which estimates work. These
options can only be changed by using the “Options” icon located in the top right
hand corner of the screen.

• Always use Source if required – always use estimates without prompting the

user.

• Always use Estimates if available – always query the source table without

prompting the user

• Default Yes/No condition to Yes gives the abi lity to selec t a defaul t condition.

This allows the administrator to default yes/no conditi ons from architect, or for
the user to set the default per session from the set applet.

Visual data

24 Set Analysis User Guide

The final analytic looks as below:.

You can perform the dashboard actions to the analytic. The analytic allows you
to:

• navigate to any of the VisualData analytic UI portlets

• select edit to create the SVG framework

• create the VisualData Query as you would from the sets applet

• save as an SVG

• refresh to update

• add to my dashboard

Refer to the Dashboard Manager user guide for more details.

Sets

Visual sets

Once you have created a set, you can add or remove members to or from it using
the database navigation facilities of the VisualData view. You can also add or
remove members to and from a current set from a previously created set. You
can perform the following functions from the Visual sets:

• Construct sets that contain members that several sets have in common.

• Perform set mathematics on any combination of sets.

• Manage complex exclusions based on existing sets.

• Visual sets counts.

Set Analysis User Guide 25

The Collections list shows you the sets that have been selected for inclusion in
the Visual sets operation. You can combine sets for inclusion in the universe by
dropping them in the same collection. By dropping sets in separate collections,
they are displayed individually in the graphical universe area. The color of each
collection matches the color of the set in the universe.

1. You add a set to a collection by selecting the set and clicking the arrow keys

to move it to a particular collection.

2. You can rename the collections by clicking the collection a window appears

to edit the collection information.

Visual sets

26 Set Analysis User Guide

3. To select the collections you can either click the diagram or click Select. The

drop-down lists all possible selections that you can make.

Sets

4. The Count calculates the number of members (and percentage) currently

selected from the available sets in the universe.

5. Click Next to select the operation you want to perform on the sets.

6. Enter the Description to the Visual Set.

7. Change the step type using the drop-down list that includes the Start button.

You can select Add, Subtract, Keep, or Points.
With SQL tools and selection tools in general, al l matching items are returned

as equal. This is not always the case in marketing, for example, so Appl ication
Foundation provides the ability to score each selection. The scores for each

Set Analysis User Guide 27

selection are then accumulated.

8. Select the Set Operation to be performed. The operations are:

- Add - Adds the selected members in the sets in the universe to the current
set.

- Subtract - Subtracts the selected members in the sets in the universe from
the current set.

- Keep - Keeps the intersection of the current set and the selected member in
the sets in the universe.

- Score - Updates the points for the selected members in the sets in the
universe.

NOTE

The first time you create a set, the only default choice in set operation is added and
therefore you do not see any options as shown in the screen above

9. Click OK to execute the operation. On the next page the set added to th e li st

and a graphical diagram of the operation performed is displayed.

Visual sets

28 Set Analysis User Guide

Freehand SQL

When you need to go beyond the standard Application Foundation user module
functions to create your own set, Freehand SQL allows you to write your own
SQL to process sets. For example, this may be necessary to link in with a fact
table or if some information necessary to create the set is not held in the
dimension table.

1. In Sets view, select an existing set, or create a new set.

2. Click Freehand SQL icon. The Freehand SQL view is displayed.

Sets

3. Type in the SQL in order to generate a list of keys to form the set.

4. Click Next. On the next page, enter the Description of the SQL process and

define any Additional Points.

5. Click OK to process the SQL. The new or modified SQL set defined in the Set

Building Set list can be seen.

NOTE

The normal set operations (Start, Add, Sub tract, et c.) are us ed to dete rmine h ow the
results of the free hand SQL are applied to the current set. The SQL statement can
return as many columns as necessary, but one column must be named ID. This
column is used as the list of ID s with which t o create the set . It is not poss ible to crea te
a set containing IDs th at are not listed in the dimension ta ble. When you proc ess a set
from freehand SQL, any IDs that are not in the dimension table are discarded.

Set Analysis User Guide 29

The final page that you see after adding or defining operations for a set looks at
illustrated below.

!

Autofill

Used to process a set from a selected base date contained in predefined
calendar to a specific date or for n periods from the base date. The AutoFill is
enabled only if the set is dynamic and if the AutoFill opti on is enabled in the set’s
properties. AutoFill can back fill the dynamic membership of a set.

For new sets, the period associated with the first build is called the base period.
For subsequent builds, the most recently ended period is associated with the set.
However, if you check Allow Auto-fill, the associated period comes after the
period for the most recent build. Thereafter, each build increments the active
period for that set. If the set is dynamic and Allow Auto-fill is checked, the AutoFill
in Sets view is enabled

REMINDER

You can do Autofill for Dynamic Sets.

To proceed with Autofill, follow the steps below. The steps explain how to create
a dynamic Set and then define autofill.

Freehand SQL

30 Set Analysis User Guide

1. Select new set from the ‘Add’ drop-down then name the set and select a

folder.

2. Select ‘Dynamic’ from the ‘Set Type’ drop-down and choose a calendar for an

analysis period

3. Define parameters for the set history:

- Snapshot: Contains the current membership, no history

- Temporal: The set contains all the membership history

- Limited Temporal: Contains the last n specified membership History

Once the dynamic set has been defined it can be AutoFilled from either the
content or membership screens using the AutoFill Icon on the build menu bar.

4. Click AutoFill icon to bring up the AutoFill Properties screen and define the

parameters for AutoFill.

Sets

Set Analysis User Guide 31

5. Check on either

- Complete: This processes the set for the all of the periods defined in the
calendar

- Partial: This processes the periods defined in the ‘ AutoFill to’ or ‘ Periods to
AutoFill’ available in the ‘AutoFill Requirements’ section

Both of these parameters adjust depending on which one is being changed

TIP

What dates you can cho ose to ‘ AutoFil l to’ are det ermined by the se lected base d ate
and whether or not the 'Stop at Current Period' option is checked in the set’s
properties. AutoFill Summary This provides summary data such as start and end
periods, t otal number of periods, etc.

6. Once the AutoFill parameters have been defined, the set can be processed.

Freehand SQL

32 Set Analysis User Guide

Rollback

Rollback allows the set to contain the hi story that has been specifi ed. It gives the
ability to rollback and refresh a selected number of periods in a temporal set.

7. Specify the Rollback to so that the action performed is from the specifi ed date.

8. Click on the calendar to select a period.

Sets

9. Click OK.

10.Specify Refresh To. It is not necessary to refresh the action.

After the operation has been completed, the set only contains the history that has
been specified.

!

Build now

This button allows you to build the sets that you defined. It calculates based on
the steps defined. As you build the steps, the sets displaying -1 can be seen. The

-1 signifies that the step has been combined with other steps to form a single SQL
statement, this gives a significant improvement in set processing.

To learn more about the New Rebuild Engine, refer to New Sets Rebuild Engine

on page 88

Set Analysis User Guide 33

NOTE

Please note that only 1SQL block can be used per set, all VisualData and Visual
Set steps can be combined into a block, all other steps cause individual
statements to be run, sampling within a step also causes the block to be broken

!

Finish

Individual steps can still be built using “Finish”. This builds the last step defined
whilst keeping the counts from previous steps intact. This method do es not take
advantage of the new “Block wise” method but gives a slight improvement in
processing time over the full client product.

!

Checking for dependencies

You can check whether the set has any dependent sets, by clicking
Dependencies in the tool bar. The Set window is displayed, showing a list of all
sets that are dependent on the current set. This shows you whether changes to
the current set have any effect on any other sets.

!

Edit a set

1. Select a Set from a folder.

2. Click Edit. The next page displays the process that was defined for the

selected set. You can modify the definition and operations for the set.

!

Remove a set

1. Select a Set from a folder.

2. Click Delete. A confirmation window pops-up letting you to confirm the

deletion of the selected set.

3. Click OK. The selected set is deleted from the folder.

!

Empty a set

A click on the button empties all the defined sets to be built again.

Freehand SQL

34 Set Analysis User Guide

Event analysis

Event Analysis performs analysis on the membership of a dynamic set (joiners,
leavers, stayers and visitors), specifies analysis periods and uses the activity
within the set for analysis or select defined parameters to specify the event type
to analyze.

To define an event analysis follow the steps below:

1. Click the drop-down list in the Set Building Steps. As you click the event

analysis page appears.

Sets

2. On this page you can select the folder and the set under it. Click the arrow key

to select the folder.

3. Specify if this is an Activity or a Summary for the set.

- An Activity button creates an action based on activity within the analysis
period.

- A Summary button creates an action based on the summary of activity
within the analysis period.

4. You now specify the analysis period. You can simply specify start and end

dates, or you can use the calendar associated with the set to specify the
analysis period.

Set Analysis User Guide 35

NOTE

You cannot proceed until you define the Analysis Period.

The event activity depends upon whether you have selected the Activity or
Summary button.

5. When you click Activity, you can choose an Activity type to analyze.

- Joiner: Analyzes the selected set for people who joined the set in the
analysis period.

- Leavers: Analyzes the selected set for people who left the set in the analysis
period.

- Stayers: Analyzes the selected set for people who were always present
within the set in the analysis period.

- Visitors: Analyzes the selected set for people who were at one time or
another in the set in the analysis period.

Event analysis

36 Set Analysis User Guide

6. When you click Summary, you can construct a summary action for the

analysis.

- Cycle allows you to select members based on their type of presence with
respect to the calendar periods.

- Condition allows you to set a condition for inclusion.

- Number allows you to set the number of periods over which to perform the
selection.

- Time Interval allows you to select the time periods to use.

7. Click Next.

8. Define the Description and Set Operation for the event.

Sets

9. Click OK and an event analysis is created for the set.

Import a file

The Import options allows you to import lists of keys from a .c sv file. The file must
contain a column based on the key. The keys must exist in the primary source
table for the current subject in order to be imported.

• In Sets view, select the set into which you want to import keys. The Import

view is displayed.

• Click Select Import File and from the Open dialog box, select the file from

which you want to import keys.

The import process imports keys that exist in the dimensi on table for the c urrent
subject. If a customer file contains 100 items and only 99 of the customer primary
keys exist in the customer table, then 99 items are inserted in the set.

TIP

If you are not sure of the type of file format required to import into Application
Foundation, then c reate a set w ith a small nu mber of items . Then choos e Export from
the File menu (do not change the default options for delimiters, etc.). Then create a
new set and import fro m the same fi le. Con firm th at yo u hav e importe d the items into
your set. Now open the comma separated file (.csv). You see how the first row
contains a column name. This is required for the import to work successfully.

Set Analysis User Guide 37

Import a file

38 Set Analysis User Guide

Group

How to create a group

Groups are a way of grouping sets. These groups can be used for classification,
analysis and scheduling purposes. Application Foundation supports two types of
groups - Rebuild and Taxonomic. Both are explained further in the chapter.

You can create a group by clicking the option button at the bottom of the page.

Click to add a Group

Sets

Click to view the Groups

1. Click Show Groups. The groups view is displayed.

2. Click the Add and select New Group.

Set Analysis User Guide 39

3. As you select the New Group option, the page displays information that can

be defined for the new group.

4. Enter the group Name.

5. Select the Folder name in which the new group is to be located.

6. Select the Type of group. There are two kinds of groups-

- Distinct - Distinct Groups allow you to group together a number of sets and
ensure that there is no overlap between them. Set Analyzer creates a new list
of members that removes any duplication of members. This can be very
useful, for example, if you have created several mailing campaigns and you
want to ensure that no contact is mailed more than once. The distinct model
allows you to define a priority for each set and ensure that each member is
allocated to only one set.

Refer to section Distinct groups on page 40 for more details.

- Rebuild - This type is a group of sets with or without similar content that can
be refreshed at a defined time interval. At the time of refresh, the
dependencies are taken into account. A rebuild group can be scheduled and
all sets under it are taken into account. For example, if set A is dependent on
Set B before it is refreshed, rebuild type group takes care of that. Refer to
section Rebuild groups on page 41 for more details.

- Taxonomic - This is also a group of set for classification and analysis
purpose. Taxonomic groups cannot be scheduled but can be only used as a
way of classifying sets in all places where sets can be selected in a tree list.
These groups are also used in group benchmark analytic. For example, if you
want to club all sets related to high income spenders then you can define the
group as taxonomic and include the relevant sets into that.

7. Enter the Description to the group.

8. Click Apply to save the new defined group.

Once the group name is saved, you can now enter the group content. Group
content is nothing but adding sets.

Group

40 Set Analysis User Guide

Distinct groups

When you select a distinct group follow the steps below:

1. As you select a distinct group the page below appears. After applying the

definition, use the Edit icon to enter the add set screen.

The next screen displays the available sets to be selected.

Build Order

Display Order

Sets

2. Sets can be moved to the group in 1 of 2 ways:

- The > < buttons add and remove individual sets

- The >> and << buttons move/remove all the sets from one pane to another.

3. Once the required sets have been added to the group, the user can reorder

according to:

- Build Order – Changes the order that the sets are built in The order in which
sets are listed as component items for a group determines the priority order
for the calculation of distinct items. The group is made up of all the items from
the first set plus only those items from subsequent sets that have not already

Once the sets have been added to the group it is ready for processing. The group
is processed on creation, The group is reprocessed after:

• A set is added to or subtracted from the group

• The build order is changed

• The display order is changed

• One of the sets within the group is refreshed. In this case, the group is

Rebuild groups

Rebuild groups allow you to group a number of sets together so that they can be
processed and reprocessed together. For example, if you have processed a final
set which has been processed from several other sets, in order to reprocess the
final set you need to reprocess all the other sets first, and ensure that each
dependent set has finished processing before the set starts. This can be a very
lengthy and manual task, and so building a processing group allows you to group
all these sets together and let Set Analyzer work out which sets must be
processed in which order.

Follow the steps below.

1. As you select a distinct group the page below appears. After applying the

Set Analysis User Guide 41

been allocated to the group. Thus, the first set has the highest priority

- Display Order – Changes the order that the sets are displayed in e.g Within
a BusinessObjects report

refreshed automatically. It is important to remember that refreshing a set that
is apart of a distinct model can be a lengthy process as the group is refreshed
as a part of the refresh process. The system displays a warning dialogue box
before the group is reprocessed

definition, use the Edit icon to enter the add set screen.

Group

42 Set Analysis User Guide

2. From the “Add Set” page, add and remove sets using the < > icons, when you

are finished, press the “OK”.

NOTE

Please note that the set highlighted in red on the screen shot is a parent of the
set “V set”. This automatically adds to the group when the child set is selected.
This set is also placed above the child set in the processing order.

3. Now the Group can be processed by clicking the “Process Group” Icon.

Sets

4. A progress bar indicates the current set being processed.

NOTE

Please note that if any of the sets within the group are rebuilt manually, outside
of the context of the group, the system reprocesses the whole group.

Set membership

This option displays members for the selected set. It shows the status of stayers,
joiners and leavers. Refer to Set Based Analytics for more details.

Set Analysis User Guide 43

Set membership

44 Set Analysis User Guide

Set metrics

This option displays the metrics for the selected set. Refer to Set Analysis on

page 53 for more details.

Sets

Export List

2

chapter

46 Set Analysis User Guide

Overview

Application actions is a feature that allows you to take actions on the insight you
gained through analysis. There are three major actions that you can take once
you have gone through your analysis.

• Rules - automate some of the analysis or to track exception when data

changes.

• Lists - generate lists of people according to criteria.

• Schedule - schedule refresh of metrics and rules

Export List

Export lists

The List feature enables the user to get a subset of individuals and products from
a given set. List is a powerful way to get a clear picture of your analysis. For
example, list of people in high profile subset or list of joiners in the Collectors set
etc.

A list is defined by the administrator through the System Configuration section but
once you access the list, you can customize it to your needs. For example, a list
of joiners may have their customer id, age marital status and income, which is
termed as the content of the list.

There are different ways of sending a list, for example, emails, URL etc., to allow
better integration with operational systems like Campaign Management.

A list can be based on single or two sets. For example, you can get a list of
leavers in High Spender set to Low Spender set.

A List can be linked to Individual Profiler. The Profiler Output enables you to
profile an individual of the list. Therefore based on your List template, the
Individual analytics generates the names corresponding to the content of the li st
which further allows you to get into deeper analysis for each individual.

Set Analysis User Guide 47

Export lists

48 Set Analysis User Guide

Define a list

To define a List do the following:

1. Click Export List from Analytic Services.

Select List Output

Select Subject Area

Select List Compositions

Next

2. On the List Generation page select an output from the List. Different outputs

can be defined by the administrator through System Configuration. Outputs
describe how the format of the list.

3. If you select exporting to a URL, you have to enter the URL string.

Export List

NOTE

This subject can have already been sele cted by the ad minis trator. The ad minis trat or
can set a default name, which the user can not change. Only the administrator sets
the email addresses and can also define a default selected objects list.

4. Select the List Composition. This indicates if you want to see a list of one set

or a migration which is a movement from one set to another or measures for
two sets.

Set Analysis User Guide 49

5. Click Next to define the Set for the List.

6. Based on the List Composition, the set lists are displayed. The set list is

based on the combination of subject area and list composition selected in the
previous section. You can view the set list either by group or set.

If you selected single set as your list composition then you have one set list
displayed.

If you selected migration as your list composition, two set lists are displayed
which are source and destination sets. It indicates the movement from the
source set to the destination set.

If you select any other list composition, two set lists are displayed to make a
comparison.

Define a list

50 Set Analysis User Guide

7. Click Next to define the Content of the list.

Content List

Selected Content

Arrows to sort the order

Arrows to move

8. A list of available information is displayed. Sel ect as many and move it to the

List content column by clicking the right arrow key. You have to move one
content head at a time.

Once you have moved the content heads, you can move them in the order
you want them to appear in the list.

If you need to remove any content heading, click the left arrow key.

9. Click Next to define the filters.

Export List

Filter List

Number of rows in the list

Sort order

Selected Filters

Set Analysis User Guide 51

10.Select at least one filter which is a subset for the list. Use the arrow keys to

move the filter name to the Conditions applied column.

11.You can choose optional size and ranking of the list. You may have only the

first 20 rows, which give you information about only 20 customers, when you
generate your list every time. You can also sort the list by the content defined.
Select the Order By list and if you want in ascending or descending order.

12.Click Generate to generate the list.

NOTE

The generation of List may take a while as the system reads the entire database
based on your list specifications.

As the list is generated it is displayed on the screen. This list can also be stored
on the server as a WebIntelligence document, downloaded and saved on your
local hard drive, linked or published as a URL or sent through an email.

You can also generate a metadata XML which has to be defined at the time of
setting up Application Foundation.

Define a list

52 Set Analysis User Guide

Export List

Set Analysis

3

chapter

54 Set Analysis User Guide

Overview

If a business deals with customers, each individual customer belongs to a group
with something in common. For example, age group or state they live in or marital
status. These groups are known as sets and every customer can be classified,
analyzed and treated in different ways using set techniques. You can create very
broad or very specific sets to categorize your customers in ways that are relevant
to your business. Based on your business, you can create very specific sets and
categorize your individual customers.

The diagram below illustrates the link between subject area, sets and subsets
with an example.

LEVELS

EXAMPLE

Subject Area

Set Group

Set Group

Set

Set

SubsetSubset

Subset

The Set List contains two types of entries:

• Individual Sets

• Set Groups

Individual Sets and Set Groups reside in a Set Analyzer set repository that your
Application Foundation administrator makes available for analysis. The Set List
is the most visible component of the Set Explorer.

Individual Sets may or may not belong to Set Groups, but Set Groups generally
contain more than one individual set. Grouping sets can be a very useful way of
classifying and organizing sets into meaningful categories to facilitate analysis.

For more information about the creation of subject areas, sets, and Set Groups,
consult your Application Foundation administrator.

Member Joiners Leavers

Customer

Single Set

Single Set

Collector

Collector

Set Analysis

Interactive analytics

Membership

The Membership analytics allows you to view for the selected set for the specified
period.

Set Members are those entities that currently reside in the set as of the most
recent set refresh. For example, suppose your
customers who purchased a case of premium wine in the last month. It is now
July 1st, and your
of every month. If a customer purchased a case of premium wine on the 15th of
April, that customer is included in the Member count of the
when you log into Application Foundation to check the Overview analytic.

Set Stayers are those entities that currently reside in the set as of the most recent
refresh and also resided in the set as of the refresh prior to that. In the original
example, if the same customer had purchased a case of premium win e during the
months of March and April, that customer is included in the Stayer and Member
counts of the

Wine Collector set gets refreshed at 11:59 p.m. on the last day

Wine Collector set.

Set Analysis User Guide 55

Wine Collector set is defined as

Wine Collector set

The top part of the page displays a bar diagram with the statistical information
which are indicators based on set history.

As you click the bar, the subset trend graphs changes accordingly.
You can change the month period by clicking the arrow key near the month

name.

Interactive analytics

56 Set Analysis User Guide

Migration

Migration is movement of members from one set to another. This can be viewed
in a form of a graph using the migration analytic.

1. Click Migration analytic.

You see the graph for the selected set. You can selec t any set from the set list.
The graph clearly displays the number of joiners and number of leavers for

the selected set.

2. Select the flow type from the drop-down list.

You can view either the In-Out flow or the Net Flow.
Net Flow = In Migration - Out Migration

Same time joiners and leavers

Set Joiners are those entities that currently reside in the set as of the most recent
refresh but did not reside in the set as of the refresh prior to that. In the previous
example, if the same customer had not purchased a case of premium wine during
the month of March, that customer is included in the Joiner and Member counts
of the

Wine Collector set.

Set Analysis

This analytic shows relevant events that happen at the same time. You can
change the parameters on the graphic to observe new results.

1. Click Joiners and Leavers analytic.

You get a graph showing same time leavers and joiners.

2. Select the flow type from the drop-down list.

You can view either the In-Out flow or the Net Flow.
Net Flow = In Migration - Out Migration

Visual data counts

Refer to Visual data on page 15 for more information.

Set Analysis User Guide 57

Interactive analytics

58 Set Analysis User Guide

Membership analysis

Membership breakdown

Set Analysis

Membership breakdown shows how the number of Members is decomposed into
number of Joiners and Stayers (positive bars) over time. They can be compared
to the number of Leavers represented as negative bars.

A “stayer” is an individual who was present in a set in a given period as well as
the prior period.

A “leaver” is an individual who was not present in a set in a given period but was
present in the prior period.

A “joiner” is an individual who was present in a set in a given period but was not
present in the prior period.

The period of analysis here corresponds to the set refresh period. No metrics are
required for this analysis to run.

Membership change

Set Analysis User Guide 59

Membership change shows the source of change from the membership at the
beginning of the period to the membership at the end.

The amount of change is represented in orange with joiners (positive change)
and leavers (negative change).

No metrics are required for this analysis to run.

Membership analysis

60 Set Analysis User Guide

Membership status

Membership status is useful to see how a Membership has evolved within a
period that is wider than the set refresh interval (for example, six months if the set
is refreshed monthly).

A “leaver” is an individual who was not present i n a set at the end of the period of
analysis but was present just before this period.

A “continuous stayer” is an individual who was present in a set at the end of the
period of analysis and was present just before this period and who didn’t left
within the period of analysis.

A “broken stayer” is an individual who was present in a set at the end of the period
of analysis and was present just before this period but who left within the period
of analysis.

A “joiner” is an individual who was present in a set at the end of the period of
analysis but was not present just before this period.

“Turnover” is the number of leavers as a percentage of the members at the end
of the prior period.

“Retention” is the number of stayers as a percentage of the members at the end
of the prior period.

No metrics are required for this analysis to run.

Set Analysis

Set metrics

For Set Metrics, you have to make sure that you select a Set and a Metric. Follow
the steps below. They are the same for all set metrics and migration analytics.

1. As you click any set metric from the catalog, the page below appears.

Select the subject area

Select the set

Set Analysis User Guide 61

Select a Metric

2. Once you select a metric, specify a Time Period for which you want to view

3. You can view the information in various ways. Select one from the drop-down.

Multi-set scatter

This compares members, joiners, leavers etc. with respect to the current values
of the metrics simultaneously, for example, cost and revenue. This is a scatter
plot comparing two different metrics.

Multi-Subset Scatter-plot shows a snapshot of two or more subsets of a set. Two
metrics define the scatter-plot X-Y axes. Use this analysis to simultaneously
compare current values for differing subsets with respect to two measures. For
example, if you want to know whether there is an emerging trend in revenue/cost
ratios among Joiners and Leavers, you can plot Revenue vs. Cost for Joiners and
Leavers and compare their relative positions.

Analysis-Specific Tip

• You can easily page through different periods.

the analytic. Pick one from the drop-down list.

Set metrics

62 Set Analysis User Guide

Percent of group total

Set Analysis

Percent of group total shows how much a set contributes to the total of a group.
It is appropriate to mutually exclusive sets.

The percent of group total enables the comparison of two metrics that are
measured in different units.

Applied to a set of mutually exclusive tiers that covers all the individuals, with a
Count metric (for example, Number of individuals) and a Revenue metric, this
analysis is a 80:20 rule chart over time.

Percent of enterprise

Set Analysis User Guide 63

Percent of enterprise shows how much a set contributes to the enterprise total.
The formula for Percent of Enterprise is: Set Metric / Enterprise Metric

Set metrics

64 Set Analysis User Guide

Set vs. enterprise

Set Analysis

Set vs. enterprise is useful to benchmark a Set metric against the enterprise.
Percent of origin makes the two lines start at the same value (100%) in order to
facilitate the comparison.

Sets vs. group average

Set Analysis User Guide 65

Set vs. Group average is useful to benchmark a set raw value against the
average value of a group. Group average is calculated by averaging the values
of the sets within the group. Therefore if the metric is “Average Order Size”, the
group average computes an average of the “Average Order size”.

Set metrics

66 Set Analysis User Guide

Sets vs. group total growth

Set Analysis

Set vs. Group total growth is useful to benchmark a Set raw value against the
group total growth.

Group Total Growth solves the scale issue of Set raw value versus group total. It
makes the group total trend relative to the starting value of the Set trend.
Therefore the group total growth always starts with the same value as the Set
trend.

Sets vs. group turnover

Set Analysis User Guide 67

Set vs. Group turnover is useful to benchmark the turnover of a set against the
turnover of a group.

“Turnover” is the number of leavers as a percentage of the prior period’s
members. It reveals the amount of change, or churn, that has occurred in a
membership.

The period of analysis here corresponds to the set refresh period. No metrics a re
required for this analysis to run.

Set metrics

68 Set Analysis User Guide

Set compare

Set compare is useful to compare sets at a point in time.
It distinguishes which sets are above the average (blue flat line) from those that

are below.

Set Analysis

Migration

Migrants compare

Set Analysis User Guide 69

In the last 7 months we
managed to re-activate
only customers that
are the most profitable.

Migrants compare is useful for comparing metrics based on migrants.
The standard net flow enables the comparison of metrics that are measured in

different units. For this analysis to run, migrants metrics must be defined.

Migration

70 Set Analysis User Guide

Migration by tier

Migration by tier is especially useful for seeing at a glance whether there is a
general migration trend toward high-tier sets or toward low-tier sets.

It is a tabular analysis which, in addition to showing the number of migrants
between sets representing different tiers, as well as the amount of overall up­migration and down-migration.

For this analysis to run, the names of the sets must have an alpha-numeric
ordering that represents their relative positions (for example, Tier 1, Tier 2, Tier
3…). and Migrants metrics must be defined.

Set Analysis

Individual Analytics

4

chapter

72 Set Analysis User Guide

Overview

Individual analysis delivers detailed historical information describing the
evolution of an individual’s relationship to the business. It can be used to profile
customers, products, partners, etc. Individual analysis supports discovery of
pattern that occurs over a period of time, which can lead to new ideas for
segmentation.

It delivers not only individual-level attribute data, but also the information needed
to understand relationship history or "customer journey." As an embedded or
adjunct facility in an operational environment, it provides rapid access to current
set membership information, as well as access, as needed, to greater detail on
set membership over time, standard metrics, and attribute data.

For example, high value customer profiling can uncover patterns that foretell high
value behavior within a set.Or an individual behavior can be a benchmark to a
wider population within the set.

The individual analysis has three main tabs that help you understand the
behavior of the selected individual in relation with one or more sets.

• Individual Portrait

• Membership

- Current Membership

- Membership Summary

- Membership History

• Metrics

- Metrics Summary

- Metrics trends

- Metric Sets vs. Individuals

Individual Analytics

Individual portrait

The portrait tab is the entry point of Individual analysis. It provides a snap shot of
descriptive data about the individual. Reports and documents can be l inked that
provide additional details, such as, transaction history, if they have been set up
during configuration.

The Portrait Tab is intended to give users a quick and convenient overview of
information on the selected individual. It provides an interface to two kinds of
information:

• Standard Dimension or Attribute Information

• Detail information, such as transaction histories or other contact or click

Information for the individual

On the Portrait page, you have to perform a search on a particular individual.

Set Analysis User Guide 73

1. Enter either the Customer Id, Name or City. You can use wild card "*" to

perform a search. You can either enter one or all three fields. The search
engine looks for a combination entered.

2. Click Search.

As a customer is found in the database, de tails get displayed in the respective
fields. You can view the basic portrait of the selected individual. The Header
takes up a single line, and is always displayed on top within this analytics.

Profile Header

REMINDER

To search for another

individual

Remember the fields displayed on the page is dependent on what your system
administrator defined through the System Configuration option.

Individual portrait

74 Set Analysis User Guide

Membership

Current membership

The membership tab shows the individual’s membership to various sets. You can
see membership as of current date or a history or just a summary. This tab
provides information for individual as well as helps benchmark them against
predefined metrics. Membership tab is a powerful analysis workflow where the
analyst can navigate from high level membership information to detailed
comparison. You can get membership information by a single date or by sliding
a time window.

The Membership Tab provides three sub-tabs for examining the history of set
membership for the individual being profiled:

• Current

• Summary

• History

All views share the same Set Tree List, including the as sociated Display Options
and Filter Options

This view provides you with information on the individual's relationship to the
selected sets at a point in time. The point in time defaults to Today, or to the End
Date for the selected time window if the Point Summary is visited after you visit
a Period view (Summary or History). For the sets that have been selected, you
see the same columns as in the Overview sub-tab, with all calculations being
performed as of the selected date (information on membership subsequent to the
as of date is excluded).

Apart from its usefulness by itself, this view helps you better interpret the
information in the Summary View (for example, telling whether the Period
information is consistent with historical behavior or indicative of a change).

Individual Analytics

Set Analysis User Guide 75

To view the Current Membership do the following:

1. Click Current Membership tab.

Move arrow to
change time period

Zoom in/out

Indicates Previous
Membership

Ways to view data

Indicates Current

Membership

2. You see all the sets and the following information -

• Member- indicates if the individual is a member of any set. Blue indicates a

current members; grey indicates the individual was a previous member but is
not current member.

• Time since Joined - number of days since the individual joined the sets until

the current date. This is for current membership only.

• Time since Last Left - number of days since the individual left the set until

current date. This is for prior membership which is no longer current.

• First Joined - the date when the individual first joined the set.

• Presence - total number of days present in the set sine the first day of joining.

• Presence % - percentage value of the present days in the set.

3. Click the option buttons at the bottom of the page which gives you the option

of viewing data in different ways. You can view by sets, by group or view only
the selected sets or sets which have data for it.

4. You can zoom in or out of the time period by clicking the zoom in/out button.

5. You can click the set name to see more details.

Through this tab you saw basic information about the selected individual and
associated sets. You can also gather information as to when the individual left a
set, joined a set and remained in a set.

Membership

76 Set Analysis User Guide

Membership summary

The membership summary tab displays summary statistics on the individuals
membership in various sets. This is similar to the Current Membership, except
that the time ruler is used to select an interval instead of a single date. The
columns are also similarly defined but are restricted to the selected time window
on the ruler. You can click the set to see membership episodes dates within the
selected time window.

Search for

Select time period grain

another individual

Membership history

Individual Analytics

Ways to view data

You see the following information:

• Presence - total number of days present in the set since the first day of joining.

• # Occurrence - total number of times the individual left and came back into the

set for the selected time period.

• Avg Duration - the average duration for which the individual was part of the

set.

• Presence % - percentage value of the present days in the set

The Membership History tab displays individual periods of membership in the
selected sets as a line graph. Each row displays all epi sodes of mem bership for
the selected individual within the selected time interval. Eac h line set rep resents
a single episode of membership. The minimum length of any episode displayed
for a given set is determined by its refresh frequency.

Set Analysis User Guide 77

The time window ruler enables selection of a time window within which the hi story
is displayed. The grain of the ruler determines whether the time ruler cursors
work in days, weeks, months, quarters or years. Zoom widens/reduces the date
range displayed in the ruler. Varying the width of the time window can change the
episode display - wider intervals widen the grain of the display.

If the interval between adjacent episodes is too short to display adequately, given
the scale of the window width, then the character used to display the line is
changed to signify the missing detail.

Sets are selectable via the standard tree list operation.

Search for another

Select time period grain

individual

Ways to view data

Shows a movement
pattern within sets

This page gives you information about individual membership history over a
period of time. You may see broken lines which indicates that the individual has
moved within a set.

Membership

78 Set Analysis User Guide

Metrics

Metrics summary

The Metrics tab provides the user with a powerful facility for observing the
evolution of an individual with respect to one or more metrics over time. To
facilitate meaningful interpretation of individual data, the Summary and Trend
sub tabs support comparison of individual data with one or more metrics.

The sub tabs are:

• Summary

• Trend

• Set vs. Individual

The time window over which the metrics are calculated are selectable based on
the Time Ruler, as is the grain (aggregation level or calculation interval) for
calculations in the Multi-Trend view.

Grain selection in the Metric tab is based on which metric is selected first.

This view presents measure values for the individual based on a selected period
as well as "closest-in-time" metric values.

In Metric Summary view, the calculation interval for the indi vidual is the same as
the Observation Time Interval.

The most recent refresh date for the metric value is displayed and then the metric
value. The format of the date reveals the metric grain. To ensure that the
individual's values are directly comparable to the metric values, you can choose
a grain and double click Time Ruler to select a single period of that grain. Values
of selected metric of that grain have their values for the selected period displayed
in normal font. Otherwise, the values are shown in gray and the following rules
are used to select the "best matching" metric value.

Choose the metric value to display as that one whose calculation was based
upon:

• the most recent, entirely contained sub-interval of the selected period,

• the larger period (with data) that includes the greatest portion of the selected

period, or

Individual Analytics

Set Analysis User Guide 79

• that smaller period among the 1 or more smaller periods touching the selected

interval that is the most recent (with data).

Move arrow to

Select single period

grain

Zoom in/out

Ways to view data

change time period

A metric value is displayed in regular font if the selected time interval corresponds
to the period for the metric and a value exists for it. Metric periods are single day,
week, month, quarter or year.

To view metric summary do the following:

1. Select a single period, choose the grain (day, week etc.) and double click at

some point on the ruler.

2. Click Change Metrics. A window appears which allows you to select metrics.

Check the metric name and click OK. The metrics selected are displayed on
the top panel.

3. Select sets for which you want to see the metric summary.

4. Click Update. The numbers for the metrics are displayed for the selected sets

for the individual. As you update, the button is removed from the page. Only
if you change the metrics that it appears again.

When an update is executed all data is calculated for the customer for the
selected metrics based on the time grain. As you change the time period, the
metric information is re-calculated.

The calculation is also performed for each selected set based on the time period
selected.

Metrics

80 Set Analysis User Guide

Metric trend s

Through this tab you get a picture of the individual’s metrics for a peri od of ti me.
It also compares against a set. A set metric value al lows you to see the individual
contribution to the set for a specified period of time.

This tab displays trends for BusinessObjects measures calculated for a selected
individual over the selected time interval. These trends are for the selected
individual metrics. You can also select sets and see the trend of the individual
with one or more sets.

In metric trend, you can interactively choose which metrics to display by selecting
the set from the set tree list as well as which metric/transformation combinations
from those that are available.

You can first choose which metrics to display and then, in a subsequent step,
cause the calculation and display of the individual's values for the selected
metric.

Select time perio d

grain

Move arrow to
change time period

Individual Analytics

Select metrics by clicking the box

Ways to display the

chart

To view the individual metric trends do the following:

1. Select a single period, choose the grain (day, week etc.).

2. Select the Metrics. You cannot choose more than eight metrics.

3. Click Get Chart. This generates a chart based on the metrics selected for the

specified individual. The trend lines shows values of the metrics/transforms
calculated based on the selected grain over the selected time window.

Metric/transforms can be displayed as actual value, logs of the actual value
or percentage change of origin.

NOTE

Metrics/Multi-Trend makes no reference to sets. It can therefore be used along with
Portrait and Metrics/Summary to analyze an individual in the absence of set history
information.

Sets vs. individuals

Set vs. Individual permits comparison of a metric trend line for an individual to
that of one or more sets

This tab allows comparison of a metric trend line for an individual to that of one
or more sets. Values can be compared directly or trends can be compared using
percentage change from origin. It helps you benchmark an individual to set
averages over time, or alternatively, as may be useful in a B2B scenario, how
much of set totals are accounted for by the selected individual.

Use % change from origin when comparing trends lines for sums or counts, when
an ’average by individual’ metrics are not available. Alternatively, view the raw
values to analyze trends with respect to proportion of total metric value
accounted for by the selected individual.

Set Analysis User Guide 81

Metrics

82 Set Analysis User Guide

Select time period
grain

Zoom in/out

Move arrow to
change time period

Click to select
a metric

Individual Analytics

Ways to view data

Select a dimension/slice

To do so:

1. Select a single period, choose the grain (day, week etc.).

2. Click Change Metric and select the metric for which you want to make a

comparison. A window appears allowing you to select the metric.

3. As you click OK for the selected metric, the values are generated and a graph

is displayed.

4. You can make a comparison between the individual and a set for the selected

metric. Or even change the metric.

5. To do so, click one or more sets. The graph immediately changes to represent

the comparison. Each set is shown in a unique color.
You can view the graph in number of ways. By default it shows a line graph.

You have the option of a bar chart, % of origin, % change from origin.

Glossary of Terms

appendix

84 Set Analysis User Guide

Change

Current vs. Prior

Enterprise Analysis

Goal

Individual Analysis

Joiners

Leavers

Change is a subtraction between 2 values of a single metric.
Generic formula is: Metric value Current Period - Metric value
Prior period

Current Versus Prior is a division between 2 values of a single
metric.
Generic formula is: Metric value Current Period / Metric
value Prior period

An enterprise metric is a metric that is not Set based. It
provides an overall picture of a company or enterprise.

A goal is a variable representing an individual-level outcome
of interest.

Individual analysis delivers detailed h istorical information
describing the evolution of an individual’s relationship to the
business.

Joiners are those entities that currently reside in the set as of
the most recent refresh but did not reside in the set as of the
refresh prior to that.

Leavers are those entities that resided in the set as of the
refresh prior to the most recent refresh but do not reside in the
set as of the most recent set refr e sh.

Members

Metric

Migration

Percent Change

Glossary of Terms

Members are those entities that currently reside in the set as of
the most recent set refresh.

A Metric is a mechanism that allows you to track a measure
over time for a particular subset of a set.

Migration is movement of individual customers from one set
to another.

Percent change is the relative change between 2 values of a
single metric.
Generic formula is: (Metric value Current Period - Metric
value Prior period) / Metric value Prior period

Set Analysis User Guide 85

Percent Change from
Origin

Percent from Origin

Metric

Schedule
Stayers

Subset

Time Series Analysis

"% Change from Origin" is the relative change between a
Current value and an Origin value of a single metric.
Generic formula is: (Metric value Current Period - Metric
value Origin period) / Metric value Origi n period

% from Origin is a division between 2 values of a single metric.
Generic formula is: Metric value Current Period / Metric
value Origin Period

Metric is the mechanism that allow s you to create report
analyses to track customer set and subset behavior.

Scheduling is executing an action at a particular given time.
Stayers are those entities that currently reside in the set as of

the most recent refresh and also resided in the set as of the
refresh prior to that.

Subsets track the movement into and out of each set. Subset
types are: Joiners, Leavers, Members.

Time-series analysis refers to the unique ability of Application
Foundation to track the com plex details of millions of
customer behavior and their transactio ns in real-t ime, thu s
allowing you to respond rapidly to trends and take
preventative action.

Time Stamp
Variable

Time Stamp is the Date-time when the sampling needs to start.
A "variable" is simply a measured characteristic or attribut e.

86 Set Analysis User Guide

Glossary of Terms

Appendix

chapter

88 Application Foundation User’s Guide

New Sets Rebuild Engine

Individual steps can still be built us ing the “Finish” button. This builds the last step
defined whilst keeping the counts from previous steps intact. This method does
not take advantage of the new “Block wise” method but still gives a slight
improvement in processing time over the full client product.

Please note that only 1SQL block can be used per set, all Visual data and Visual
set steps can be combined into a block, all other steps cause individual
statements to be run, sampling within a step also causes the bloc k to be broken.

The illustration below explains the old engine and the new rebuild engine. In the
diagram, note that the earlier engine, calculated one step at a time. Where as in
the 6.1 engine, all the sets defined form one block and are calculated as one
block. Therefore you get a cumulative number but at the same time can attain
individual set calculation also.

Step 1 - Start

Step 2 - Add

h Application Foundation 2.5 Method

Step 2 - Add

Step 1 - Start

h Application Foundation 6.0 Method

Steps 1, 2 and 3

Step 3 - Subtract

Step 3 - Subtract

Appendix

Application Foundation User’s Guide 89

Below is an example taken to explain further the difference between the earlier
engine and the new rebuild engine.

h A simple set definition

i Country = France - 1 Point
i Gender = Male - 1 Point
i Home Ow ner = Y - 1 Point

n-1

2

2x2x2-1 = 7 Combinations

h Bits used to mark th e

Home Owner

Male

SQL SQL

France

SQL SQL SQL

SQL

France

SQL

Home Owner

Male

combinations

h Makes the selection

faster

h 1 table scan per query

where possible

h Using bits to mark membership

Gender = Ma leHome Owner = YDecimal Value

Gender = Ma leHome Owner = YDecimal Value

01

01
02

02
03

03
14

14
15

15
16

16
17

17

0

0
1

1
1

1
0

0
1

1
0

0
1

1

Country = France

Country = France

1

1
0

0
1

1
0

0
0

0
1

1
1

1

New Sets Rebuild Engine

90 Application Foundation User’s Guide

The new set processing allows for both Step wise and Block wise building
methods

• The new Application Foundation6.1 set processing method uses blocks of

SQL to build sets. The old method runs individual SQL statements to build up
the steps required for a set. This method allows for intermediate counts to be
displayed as follows:

- Step 1 VisualData Add - Country = “USA” 21000

- Step 2 VisualData Keep - Gender = “Female” 13000

- Step 3 VisualData Keep - State = “New York” 2100

• The new method appears as follows:

- Step 1 VisualData Add - Country = “USA” -1

- Step 2 VisualData Keep - Gender = “Female” -1

- Step 3 VisualData Keep - State = “New York” 2100

The -1 signifies that the step has been combined with other steps to form a single
SQL statement, this gives a significant improvement in set processing.

Appendix

Index

Set Analysis User Guide 91

A

Appendix 87
Application Foundation documentation

Guides in PDF 6
obtaining more inform ati on about 6

C

Cross-Set Analytics

overlapping memb ers trends 62

E

Event analysis 34
Export lists 47

F

Free hand SQL 28

G

Group 38

create 38
distinct 40
rebuild 41
type

distinct 39
rebuild 39
taxonomic 39

Group Benchmark

percent of group total 62

Individual Analytics

membership 74

current membership 74
history 76
summary 76

metrics 78

sets vs individual 81
summary 78
trends 80

portrait 73

Interactive Analysis

visual data 57

L

Lists

define 48

M

Migration

migration by tier 70
more information 6
Multi-Subset Scatterplot 61

O

online guides 6

P

PDF guides 6

I

Import a file 37
In the documentation 6

S

Sample 19

create 20

random 19

ranked 19

stratified 20

ranked 20

Index

92 Set Analysis User Guide

Set

autofill 29
build 32
building steps 12
definiti on 10
dependencies check 33
dynamic 14
edit 33
empty 33
finish 33
frozen 13
membership 43
metrics 44
new 12
remove 33
rollback 32
view 11

Set Analytics

interactive

membership 55
migration 56

Interactive Analysis

same time joiners & leavers 56
visual data counts 57

membership analysis 58

membershi p breakdown 58
membership change 59
membership status 60

Migration 69

migration compare 70
migration trend 70

Set Metr ics

multi subset scatter 62
multi-set scatter 61
percent of enterprise 63
percent of group total 62
set comparison 68
set vs enterprise 64
set vs group average 65
set vs group total growth 66
set vs group turnover 67

set metrics 61

Set metrics

individual & set groups 54

Single Set

membership analys i s 58

Subset Analysis

multi-subset scatter 69
subset trend comparison 69

V

Visual data 15

estimates 21
new 15

Visual sets 25

Index