CHAPTER
1 Defining Business Requirements Additional Web Content
Building the foundation
Business Requirements Example: Adventure Wors C!"les The sample business intelligence databases beginning with SQL Server 2005 are based on a fictitious company called Adventure or!s "ycles #A"$% a multinational manufacturer and seller of bicycles and accessories&
1
According to the descriptions in 'oo!s (nline% the company is based in 'othell% ashington% )SA and has regional sales offices in several countries& *n this section we provide a summary of a subset of business re+uirements for a company li!e Adventure or!s&
This document describes what the business re+uirements gathering process might loo! li!e at a company such as Adventure or!s "ycles&
Assuming the ,-'* system already has good business sponsorship% the first ma.or step step in the Lifecycle Lifecycle is re+uirements definition% and the first tas! in re+uirements definition is preparation&
#ntervie$ Preparation at Adventure Wors C!"les Typically% you/d carefully review all the information about Adventure or!s or!s "ycles that you could find% reading through strategy documents% annual reports% mar!eting plans% competitive analyses% analyses% and presentations from from senior management/s management/s annual offsite planning meeting& 'ecause Adventure or!s "ycles is a fictitious company% you can/t really do the !ind of research you should& SQL Server 'oo!s (nline #'(L$ provides some bac!ground information about Adventure or!s "ycles in the section Adventure or!s "ycles 'usiness Scenarios& ou may want to review the materials in '(L to get a general sense for Adventure or!s "ycles and the Adventureor!s
2
transaction system database& *n this section% we have enhanced the 'oo!s (nline content by doing some analyses of data from the database itself& e also provide you with business re+uirements information from our imaginations& THE H#%T&R' &( AD)E*T+RE W&R,% T-e Adventure Wors database -as been evolving at .i"rosoft for several !ears/ #t0s been used in various forms as a demo database for .i"rosoft0s CR. solution1 as a training database for /*ET ar"-ite"ture "ourses1 as an appli"ation to demonstrate .i"rosoft Transa"tion %erver1 and as an example to demonstrate online s-opping using Commer"e %erver/ #n t-ese last t$o in"arnations1 Adventure Wors $as a "amping gear retailer/
Adventure Wors C!"les Basi" Business #nformation There is an incredible amount of information about your organi3ation buried in your transaction systems& 'uried is the !ey word here4not many people can get at it% which is why you are building a ,-'* system in the first place& owever% as a competent systems professional% you should be able to do the data profiling needed to get at this data& A few +ueries against the transaction database can reveal much about the dynamics of the business and the nature of the data at the same time& The Adventure or!s "ycles sample databases can be downloaded from "ode6le7&com& The relational transaction database for SQL Server 2008 92 is called Adventureor!s (LT6 200892 and it re+uires the SQL Server :ull;Te7t Search install option& e/ll cheat a bit and use the relational AdventureWorksDW2008R2 database instead% because it typically installs without any problems&
3
The data warehouse version of the Adventure or!s database ma!es our +ueries easier because it has collapsed many of the transaction tables into dimensions and facts& *t is still not as easy to use as it could be because it has two separate fact tables for the company/s two main distribution channels< internet and reseller& This data e7ploration tas! is a perfect opportunity to ta!e advantage of 6ower6ivot for =7cel& e loaded both fact tables and associated dimensions into a 6ower6ivot database and added the relationships it didn/t automatically identify& e then created a computed column that added the total reseller sales amount and the total internet sales amount to get a combined TotalSales amount& e also added a new dimension called ,im"hannel using a lin!ed table from =7cel& This table has two columns< "hannel>ey and "hannel,escr& *t only has two rows% one for the internet channel and one for resellers& :inally% we added a computed column to each fact table called "hannel>ey with the value of 1 in the :act*nternetSales table and 2 in the :act9esellerSales table& e finally added relationships between the two fact tables and the ,im"hannel table based on the "hannel>ey columns& Adventureor!s,200892 holds data for three fiscal years from ?uly% 2005 through ?une @0% 2008& A few +ueries on this data reveal that Adventure or!s "ycles is doing well% at least in terms of orders and growth& The rand Total line in :igure 1&1 shows that orders have been increasing rapidly% with a BB percent increase from 200C to 200D and a 5B percent from 200D to 2008& :igure 1&1 also shows that Adventure or!s "ycles sells products in four ma.or product categories< bi!es% components% clothing% and accessories& (ver the three years in the data% bicycles
4
account for more than 8C percent of orders% with clothing and accessories ma!ing up about @ percent& (igure 2/2: Adventure Wors C!"les Total %ales Amount b! (is"al 'ear and Produ"t Categor!
Adventure or!s "ycles sells a lot of bi!es& Additional database +ueries tell us where their bi!e orders come from& :igure 1&2 reveals that A" has sales in si7 countries% with about 5C percent of orders coming from the )nited States over the three years& The percentage of orders coming from outside the )&S& has increased from about @@
percent in 200C to close to 51 percent in 2008& (igure 2/3: Adventure Wors C!"les Produ"t &rders b! Countr!4Region
'ecause we added a sales channel dimension% we can see how the two ma.or parts of our business compare% even though the data comes from two different fact tables& This drill;across function is handled automatically in 6ower6ivot% once the relationships are defined& :igure 1&@ shows that the bul! of orders come from bicycle stores and
5
distributors% also !nown as 9esellers& This channel accounts for about D@ percent of orders and is handled by a direct sales force of 18 people& Adventure or!s "ycles broadened its business during the time period covered in the database by opening up a direct sales channel to consumers on the *nternet& (ddly% *nternet orders have dropped in 200D& hile we don/t !now why this is so% we predict the E6 of Sales will ta!e credit for being able to grow the direct sales channel as fast as the *nternet has grown& 'y the way% we would chec! the underlying data to ma!e sure our 6ower6ivot database isn/t defined incorrectly before we made any ma.or public announcements based on this report&
(igure 2/5: Adventure Wors C!"les Produ"t &rders b! %ales C-annel
Fow you have a sense for what Adventure or!s "ycles sells% where they sell it% and how they sell it& The ne7t +uestion is to whom do they sellG :igure 1&B gives a snapshot of customers bro!en down by sales channel& This split is vital to understanding A"/s customers because the two channels are very different&
6
(igure 2/6 Adventure Wors C!"les "ustomer and order d!nami"s
The 9eseller channel had a total of C@5 customers who ordered between 200C and 2208% BH@ of whom placed an order in 2008& The *nternet channel is much larger% with a total of 1D%H18 active customers% 1D%B12 of whom placed an order in 2008& The 9eseller channel .ustifies a direct sales force in part because the average order is around I21%000 compared with an average of I1%0H8 for the *nternet& (n the other hand% the *nternet business should be significantly more profitable because the price would be closer to retail rather than wholesale% and the cost of selling would be much lower without a direct sales force& ou may hear more about this during the business re+uirements interviews& As the ,-'* system manager% you should continue from here% researching top reseller customers and their historical buying patterns #seasonality% product lifecycle% and so on$& *nvestigate the *nternet customers as well because you have demographic information on them that tells you who are they% what they buy% and where they come from&
7
These few reports demonstrate the power of analysis& They give you a general sense for the si3e and shape of Adventure or!s "ycles& owever% they do not give you enough information to build a ,-'* system& There is much more to learn in terms of the business% its strategies and plans% competitive environment% and !ey players& This is where you depart from the facts and venture into the fu33y;edged land of organi3ation and politics&
#ntervie$ Planning The Adventure or!s "ycles documentation doesn/t include an org chart% but the ,im=mployee table in the Adventureor!s,200892 database lists a total of 2H0 active employees& The =mployee table is self;referencing with a 6arent=mployee>ey field that allows us to generate the company/s org chart& The resulting org chart% shown in :igure 1&5% list seven direct reports to >en Sanche3% the "=(&
8
Ken Sánchez CEO (290)
Terry Duffy VP Engineering And Tool Deign !"D (#$)
%&ur& 'or&n CO in&nce " *! E+ec ,en " Adin (29)
D&-id %iu in&nce (.)
5&e *&ilon VP Producion Producion 1&nuf&curing (.)
Peer Kre8 Producion Conrol 1gr Producion Conrol 1&nuf&curing (20#)
,&ry Al&n &ciliie 1gr6 &ciliie " 1&inen&nce (7)
P&ul& /&rreo de 1&o *u&n !eource ()
/ri&n 3el
1uli:le 1&n&ger; 1fg6 Producion (#7.)
Ay Al8er Euro:e&n S&le 1gr ($)
Pil&r Ac
Sy&d A88& P&cific S&le 1gr6 (2)
3endy K&hn Purch&ing (#4) A6 Sco 3righ 1&er Scheduler 1fg6 Producion Conrol (>)
5e&n Tren&ry =S 1&n&ger =nfo Ser-ice E+ec ,en " Adin (#0)
Se:hen 5i&ng@ 'orh Aeric&n S&le 1gr6 (##)
*&+e A8olrou ?A 1&n&ger ?u&liy Aur&nce (##)
(igure 2/7 T-e Adventure Wors C!"les organi8ation "-art
The org chart should raise a few +uestions for you% such as why ?ames amilton% the E6 of 6roduction% has only one direct report% and why :acilities and Jaintenance #with seven people$ is the only group reporting to him& Jeanwhile% why does 6eter >rebs% the 6roduction "ontrol Janager #not a E6$ have all the rest of the 200;person manufacturing organi3ation reporting to himG hy is ,avid 'radley% the Jar!eting Janager% only a manager when
9
D&-id /r&dley 1&r
'rian el!er is the Sales E6G *f this were your organi3ation% these are some of the important political and power;related +uestions to get answered& This is why e7amining the org chart early on is generally a good idea& iven the number of senior e7ecutives and the si3e of the organi3ation% you would probably plan on more than 10 but fewer than 20 interviews at Adventure or!s "ycles& ,epending on availability% e7pect this to ta!e a minimum of a wee!% more li!ely one and a half to two wee!s& The rest of this write up highlights !ey parts of the re+uirements definition process in the conte7t of the Adventure or!s "ycles case study& (nce you/ve completed the pro.ect re+uirements definition step% you will be ready to begin designing your dimensional model% as covered in "hapter 2 of the boo!&
Adventure Wors C!"les Enterprise Business Requirements 'egin with a series of enterprise re+uirements interviews to get a broad sense of the important business processes and their business value% and to identify potential sponsors& Summari3e each interview by grouping the individual re+uirements into common analytic themes& This brings a useful structure to what is often a fairly scattered interaction& hile the same analytic re+uirement may come up several times in the conversation% you need to summari3e it only once& The typical summary for an hour;long interview ta!es about three to four pages& =ach summary should include the business narrative along with e7ample
10
analyses and potential data problems& The narrative should capture an estimate of the value of each business process& e/ve included an e7ample interview summary for 'rian el!er% the E6 of Sales& The following summary gives you much of the information you need to understand the business processes and the dimensional modeling decisions made in the ne7t chapter& There are additional abbreviated interview summaries in a separate document for the following people<
>en Sanche3% "=( 6eter >rebs% 6roduction "ontrol Janager ,avid 'radley% Jar!eting Janager Jary ibson% *nternet "hannel Analyst ,avid Liu% :inance Janager
Adventure or!s "ycles Eice 6resident of Sales% 'rian el!er% is a big believer in the power of information& e and his group have ta!en over one of the *T organi3ation/s analysts full time .ust to generate reports and analyses& The team decided to interview 'rian early on to get some positive reinforcement& After you/ve discussed roles and responsibilities% you might start the interview by as!ing 'rian< ow do you tell when you/re doing a great .obG *f he/s as smart as you thin!% his answer might be< hen my sales planning is accurate% when my sales grow% when * can leverage special offers% and when * have good customer satisfaction& That would be +uite an answer% but it/s a gold mine to drill down into his information needs& our .ob is to react to his answer point by point% drawing out more detail& The following summary captures the results of that drill;down&
11
Example #ntervie$ %ummar! Interviewee: Brian
Welker, VP of Sales
Date: 7/25/2009 Interviewer: Joy
Mn!y Scribe: Warren "#orn$#%ai$e Additional attendees: S$ar$ &'er, (arolyn (#a, Joy Byr!, )a*e Won+
Roles and Responsibilities
'rian el!er is head of the sales organi3ation& e/s responsible for sales to 9esellers% which was I@C&2 million in 2008% or about C8 percent of total sales& e has 1D people who report to him% including @ regional sales managers& 'rian is e7cited about his team and eager for them to be successful& They are all bi!e frea!s who love to ride bi!es and love to tal! about them4perfect bi!e sales people& 'rian is measured on achievement of the total 9eseller sales target for the year& #nformation Requirements
'rian is particularly frustrated with how difficult it is to get information out of the company/s systems& hen he as!s for a report% it can ta!e days or wee!s to get the information& (ften he/s told *t can/t be done& The ma.or analytic areas that 'rian wor!s with are as follows<
Sales planning: 6lanning for the year begins in
the fall of the previous year with the Sales planning process& Sales territories are based on geography& All new customers are assigned to a sales territory when they place their first order based on where they are located& Sales planning includes loo!ing at the following<
12
Growth analysis: (verall mar!et% new
products% new geographies% new sales people& Customer analysis: ho are the top customers% how have they changed over the last yearG Territory analysis: here are top customers located% what are the current sales territories% and how balanced are theyG ow does this map to sales regionsG
'rian and 9amesh Jeyyappan #the analysts who wor! in *T$ also loo! at sales by sales regions% which are groupings of customers based on the state where they/re located& 9egions overlap sales territories and are based on seasonal buying patterns and regional preferences& 'eing able to group historical sales li!e this helps the sales team do a better .ob of forecasting monthly sales& They usually loo! at regional sales a lot during the sales planning cycle and then compare actuals to the forecast during the year& =very year% they change the regions a bit to line up with changes in buying patterns& 9amesh does all of this data analysis for the Sales forecasting and +uota assignment process in a spreadsheet& The spreadsheet includes territory growth factors% allocations% and manual ad.ustments& The planning process is totally manual and ta!es a couple of months in the fall% and maybe a wee! per month to do the reporting during the year&
13
,uring the annual planning process% 'rian wants to be able to see reseller customer orders by year by customer territory% regardless of the sales rep assigned to the territory& *n previous .obs% 'rian would ad.ust the si3e of the territories by moving customers from one territory to another with the goal of ma!ing the territories more even& e has not done this at A" yet% so all customers are still assigned to their original territory& Sales reps can be reassigned to different territories% usually when a sales rep leaves& Sales performance: (nce the planning process is
done% 'rian wants to see sales according to the new territory assignments% all the way bac! through history so he can compare with actuals as they come in& At any time% Sales must be able to re;create historical sales and commission reports based on what happened at the time of the order% not which territory gets credit today& 'rian also wants to loo! at orders from a sales rep perspective& The first thing he wants to see at the start of the wee! is how his sales reps are doing year to date& *f 'rian sees a problem in the higher level data% he wants to be able to drill down to detailed orders for individual reps& (f course% 'rian has other reports he would li!e to see< for e7ample% top 20 customers and orders by 9eseller versus online& Basic sales reporting: 'rian wants to ta!e better advantage of the customer information buried in the orders transaction system& The sales reps would really appreciate it if they could get a list of the customers in their territory ran!ed by
14
orders& 'ecause most sales go to a small percentage of 9esellers% the sales reps would concentrate on ma!ing sure those important customers are happy& 'eyond this% 'rian !nows that 1D percent of 2002 customers did not reorder in 200@& And to date in 200B% he still has not heard from an additional 1D percent or so& is sales people could use this information to bring the best of these customers bac! to the business& Price lists: The fact that the sales reps are out in the field most of the time ma!es it difficult for them to !eep their price lists current& The price list changes fairly often% but only a few things on the list change& *t would be great to get a report to the sales reps that flagged changes and special offers% and maybe even highlighted the relevant customers& Special offers: The special offers could be a great sales tool& A" .ust finished an inventory clearance sale on the silver Jountain 500s& The color didn/t sell as well as others% resulting in too many in stoc! at model changeover& Jary ibson% the mar!eting assistant for mountain bi!es and ,avid Liu in :inance put their heads together to come up with ideas to stimulate demand and came up with a B0 percent off offer& This is something the sales people can wor! with% but it/s a random process& They/d li!e a report that shows which of their customers bought a lot of the product that is on special offer to see if they/re interested in more at a great price&
15
'rian would li!e his sales people to start with the biggest potential customers first and !eep selling down the list until they run out& Actually% 'rian thin!s the business would be better served if they contacted the more profitable customers about special offers first& Some of the biggest customers are big because they scoop up specials% which don/t ma!e a lot of money for A"& That/s another thing< The sales reps need to !now when out of stoc!s occur on special offers& Customer satisfaction: 'rian would li!e to create some measures of customer satisfaction and has been trying to get more information out of the customer care system lately% with limited success& e would love to be able to trac! calls by complaint type% product% sales region% and customer to get a sense for customer satisfaction and product +uality& e also thin!s comparing order date and ship date in the sales data to identify late orders% and determining the percentage of returned items might be indicators of customer satisfaction& This would ma!e a great start at a customer satisfaction scorecard& International support: The company has been growing internationally% but the transaction systems haven/t !ept up& The systems do ta!e orders in multiple currencies% but none of the descriptions has been translated from =nglish& This is a problem for the sales people% who have split up the product list and done the translations themselves& This doesn/t wor! in the long run because no one !nows if they/ve translated the information correctly& All materials must be
16
bilingual to comply with "anadian law& The product tags and documents are already bilingual% but the sales materials are not& Additional Issues
'rian e7pressed a frustration on the part of his sales force about the difficulty they have using e7isting reports& *t is our sense that 'rian would li!e an analytic system that provides his sales reps most of the information they need in a standard format with .ust a few !eystro!es& The time 3one differences ma!e it hard for some of them to get live support from head+uarters& *f they need to get custom information% he would li!e it to be easy for them to get it themselves& Success Criteria
'rian would li!e the system to provide him and his team with<
=asy access to basic sales data for the whole field
organi3ation :le7ible reporting and analysis tools All the data in one place #especially sales and
forecast data$
Requirements %ummaries: W-at0s .issing9 The short answer to what/s missing is a lot& This particular summary doesn/t include much indication of business value& And one summary will never provide the range of business re+uirements you get from a full set of interviews& *n addition% you would need *T and business
17
analyst representation to get a sense for the level of effort or even availability of some of the data re+uested& owever% there is enough information here to help you ma!e sense of the ne7t few steps&
Anal!ti" T-emes and Business Pro"esses The interview summaries have already grouped similar re+uirements into common analytic themes& e/ve found that it/s easiest to identify the business processes that support each analytic theme as you write the summary& :or e7ample% Sales 6lanning is an analytic theme that is supported by information from the orders business process& :igure 1&C is an e7ample of how you might distill Adventure or!s "ycles/ business processes from the analytic themes based on 'rian el!er/s interview summary& our final list will include many more themes and business processes& (igure 2/ Anal!ti" t-emes and supporting business pro"esses from t-e intervie$ summar!
18
The Supporting 'usiness 6rocess column in :igure 1&C allows you to group themes that rely on data from the same business process and to identify themes that re+uire data from multiple business processes& 9ecall that themes that rely on data from multiple business processes% or consolidated themes% are more difficult to implement& Loo! for the highly leveraged opportunities where several themes can be delivered with data from a single business process& *n most organi3ations% this opportunity is the sales business process% which Adventure or!s "ycles calls orders&
19
THE *A.E ;A.E #* ACT#&* T-e business pro"esses listed in (igure 2/ do not ne"essaril! mat"- up $it- t-e terms $e -eard from Brian Weler in t-e intervie$/ He0s t-e )P of %ales1 and tends to use t-e $ord -e is measured on orders/ T-e pro"ess of determining business pro"ess1 table1 and attribute names starts -ere/ 'ou need to be as pre"ise as possible in !our use of terms/ #n fa"t1 as soon as !ou understand t-e distin"tion bet$een sales and orders1 !ou s-ould re$rite t-e intervie$ summar! to refle"t it/ We left t-e ambiguit! in our summar! to mae t-e point/
(nce you fill in the complete list of themes% re;sorting it by business process will reveal that many of the business re+uirements rely on data from a few business processes& :or Adventure or!s "ycles% data from the orders business process is all that is needed to enable a wide range of analytic re+uirements% many of which are not on 'rian/s list in :igure 1&D& As the full set of re+uirements would reveal% Sales is not the only department interested in orders data& This is important because it adds to the overall business value of implementing the orders business process& All or most of the following re+uirements could be met with data from the orders business process<
Sales planning input 'asic sales reporting Special offers 6roduction forecast input 6roduct planning and monitoring *nternet customer demographics #with a well; designed customer dimension$ "ustomer profiling "ustomer loyalty program Standardi3ed currency reporting
20
'rian also discussed a few consolidated analytic themes that re+uire data from more than one business process #from :igure 1&C$<
Sales performance: (rders and forecast Customer satisfaction: "all trac!ing% orders% and
returns The prioriti3ation process focuses on business processes because these are the coherent units of wor! for the ,-'* system& *n describing each business process% you need to tie it bac! to the analytic themes it supports% so senior management can assess its business value& The business processes from 'rian/s interview are shown in Table 1&1& 'rian/s supported analytic themes are underlinedK the rest came from other interviews& Table 2/2 Business Pro"esses Derived from Brian Weler0s #ntervie$ LETTE
BUSINESS
R
PROCESS
SUPPORTED ANALYTIC THEMES
A
Order
Order re:or0ing &nd &n&lyi order forec&0ing &d-er0iing effec0i-ene cu0o(er &0if&c0ion :roduc0ion forec&0ing :roduc0 :rofi0&8ili0y cu0o(er :rofi0&8ili0y
/
Order forec&0
S&le :erfor(&nce 8uine :l&nning :roduc0ion forec&0
C
C&ll 0r&c
C&ll cen0er :erfor(&nce cu0o(er &0if&c0ion :roduc0 Bu&li0y cu0o(er :rofi0&8ili0y :roduc0 :rofi0&8ili0y
D
!e0urn
Cu0o(er &0if&c0ion :roduc0 Bu&li0y cu0o(er :rofi0&8ili0y :roduc0 :rofi0&8ili0y ne0 &le
21
'y the time you/re finished with the interviews and write;ups% you should have a list of business processes that represents the ma.or activities of your organi3ation< 25 to 50 is a reasonable count at this point& ou/ll leave some items off the list% even though they will be discussed in the document and with senior management& *n 'rian/s case% price lists and international support are important issues to his organi3ation% but they are transaction system problems because they involve enhancing transaction systems or building new *T infrastructure& ou can help with better reporting% but you shouldn/t be dealing with connectivity and data capture issues if you can avoid it&
Adventure Wors C!"les Bus .atrix 'usiness processes are the units of wor! in building the information layer of the '* system& The bus matri7 shows the business processes and the dimensions needed to support them& The more you !now about how the business and its supporting systems wor! the easier it is to define the bus matri7& our full version of Table 1&5 would be the starting point for an initial version of your bus matri7& :igure 1&D shows a high;level version of the Adventure or!s "ycles bus matri7 for the business processes from Table 1&1&
22
(igure 2/? T-e Adventure Wors C!"les bus matrix
The bus matri7 gets another level of detail during the dimensional modeling process when you add in the different types of fact tables& ou/ll get more on this in "hapter 2&
T-e Adventure Wors C!"les Prioriti8ation Pro"ess As we described in "hapter 1 of the boo!% the prioriti3ation process is a facilitated session where the ,-'* team presents the preliminary pro.ect findings and facilitates discussion with !ey business representatives& The primary goals of the prioriti3ation process are to ensure a common understanding of the re+uirements findings and business processes% and prioriti3e the business processes based on business value and feasibility& The top priority business process becomes the focus of the initial ,-'* pro.ect&
23
:igure 1&8 shows the resulting grid from the Adventure or!s "ycles prioriti3ation session& *t includes a few more business processes than we listed in Table 1&1% but it/s still not a full set&
(igure 2/@ T-e Adventure Wors C!"les prioriti8ation grid
Fote that there are two items on the grid that are not actually business processes< "ustomer and product profitability are consolidated themes that senior management has e7pressed significant interest in analy3ing& These have been included on the grid to show their importance% but they are far over to the left to indicate the difficulty involved in building all the needed business processes& iven the number of analyses supported by data from the orders business process% it should come as no
24
surprise that orders is the top priority theme& The team should get to wor! on this right away
Business Requirements for t-e &rders Proe"t etting to wor! on the orders business process re+uires holding an additional set of interviews to drill down on orders;related analyses& The team needs to understand several issues that were raised in the enterprise re+uirements process% such as what the different !inds of regions that people described are& The team would get more specific about the !inds of new reports and analyses people want to see as input to the '* Application trac!& All of this information becomes the grist for the Adventure or!s "ycles business dimensional modeling process case study in "hapter 2&
%ummar! This write;up went through an abbreviated e7ample of the business re+uirements gathering process at Adventure or!s "ycles& e started with some data profiling to get a sense for the nature of the Adventure or!s business& e then went through an interview summary for the E6 of Sales& e then showed how the analytic themes from that interview tied to the business processes that fed into the bus matri7 and the prioriti3ation process&
25