I have a problem with my homework. Can somebody help me with ideas and a way how to solve the problem?
I'm thinking through solving it with clusters and a buffer that sounds reasonable. Are there also good ways with the network, business or geostatistical analyst?
Thx for any help
Here the problem:
The client is a beverage service distributor undergoing a sales force transformation effort. They
have 60+ divisions with approximately 50,000 customer sites. Each site is classified as either a
"street" account (e.g. mom-and-pop restaurant) or a "national" account (e.g. hospitals).
Typically, commissioned salespeople called Territory Managers (or TM's) handle the street
accounts, and salaried salespeople called National Account Managers (or NAM's) handle the
national accounts. Salespeople periodically visit their accounts in person.
However, based in part on geography, some national accounts are handled by TM's. Key West,
Florida, is a prototypical example – Key West's cluster of accounts is remote relative to the
client's other accounts, and contains both street accounts and a few national accounts. Rather
than dispatch a NAM's all the way to Key West to visit national accounts, the client has
assigned all the street and most national accounts there to TM's, so fewer salesperson are
inconvenienced by the journey to the remote location.
The client recognizes however that commission driven TM's are more expensive, harder to
manage and harder to staff than salaried NAM's, so the out team is investigating the
potential savings associated with moving some of the national accounts that are currently
managed by TM's over to NAM's. Our team wants to make an intelligent recommendation that
leaves situations like Key West as they are. Initially, the client wants GeoAnalytics to help
quantify the percent of national accounts currently managed by TM's that should remain the
responsibility of TM's rather than being shifted to NAM's.
The client has requested the following initial deliverables:
1) A percentage of national accounts to be excluded from the shift to NAM's;
2) A map of the excluded accounts;
3) An explanation of the approach to establishing the exclusions.
It's acceptable to limit the initial analysis to 5 or 6 divisions to speed up processing time if
When discussing analytical approaches with the client, a few assumptions/comments emerged:
1) Visiting 3-4 accounts in a day is reasonable and productive.
2) No less than 20% of the accounts should be left with TM's.
3) All accounts are, technically speaking, national accounts, but some are being served by street representatives.
4) We don't know where salespeople begin their days – some start from home, but we
don't know their home addresses, and should not try to factor this in, as a
reconfigured set of accounts may simply force rehiring of salespeople.
5) Of the sites with National Team presence, some of these are National Chains
(segment = "NCH"), which should be excluded (or, rather, treated as sites without
regular National Team coverage) when running the analysis.
6) There are potential obstacles associated with transferring customer sites owned by
The associated file contains the locations of the client's own customers, and which division,
team and salesperson covers them.
Data definitions for the file's fields are as follows:
ID – a unique number for each row of data.
Division – refers to the geographic division of the company that manages each customer site.
Customer Number – refers to each unique customer site.
Segment – a grouping of individual accounts into categories such as IND for industry or EDU
Business Type – this speaks to whether the business is, in theory, national or local. Everything
with an "NM" in here is theoretically national business; everything with an "LMC or "LMNC" is
theoretically local business. Note: this doesn't mean that these accounts are actually served by
national vs. local sales reps.
Covered By – designates which type of rep actually covers the account ("National Team" vs.
"Street Team", and a few "Unclear" that are still worth mapping if they are NM accounts).
Salesperson ID – a unique code for each salesperson in the beverage distribution company.
Latitude – the Y coordinate for the customer location visited by the sales representative (in an
EPSG:4326 spatial reference system) as derived from the teams Geocoder.
Longitude – the X coordinate for the customer location visited by the sales representative (in
an EPSG:4326 spatial reference system) as derived from the teams Geocoder
Location type – designates the accuracy of the location (based on a geocoding process).