POC

1
POC for Halo/Lee Wayne
Prepared byScope - A Quatrro Group Company
FEBRUARY 2010
2
Scope Services for Halo/Lee Wayne
Taxonomy
Define a single organizational taxonomy
Formalize
Formalize the descriptions (without lexicons)
Dedupe
Dedupe the case description by identifiers
Classify the case descriptions into the taxonomy
Classify
Extract the technical information from the
descriptionsand case fields
Extract
Dedupe
Dedupe the case using the technical attributes
Descriptor
Build normalized descriptions for the record
3
Formalizing The Descriptions

• Input data
• Messy descriptions in several languages
• No manufacturer information
• Data to be utilized
• Lexicons
• Desired output
• Cleansed

Generates a formalized and standardized list of
tokens (a single element of data) from the
original description While formalizing the data,
the user can train the system with new examples
to enhance the lexicons and ultimately accelerate
the automatic process
DIA WH 1L1 140X9X.75 D11 1L1 140X9X.75 D91 7350114
WHL-1421 SPECIAL PLATED 320 GRIT 10X2.000X3 HOLE
M10x1mmx30mm SHCS-SS
Diamond Wheel 1L1 140 X 9 X 0.75 D11 1L1 140 X 9
X 0.75 D91 7350114
Wheel 1421 special plated 320 grit 10 X 2 X 3
hole
M10 x 1mm x 30mm SHCS Stainless Steel
4
Formalizing Screen
5
De Duping

• The Deduper is used to identify and manage
  duplicates.
• Deduping is performed on cases
• The deduping is done over all the cases
• The deduping is done based on different
  identifiers as
• Deduping by identifiers enables fast and accurate
  identification of duplicates
• Manufacturer PN
• Supplier PIN

6
Duplication Example
Manufacturer PN Description
1010180 SKA38X5 0070989900
1010180 SKA38X5 0070989867
1010180 SKA38X5 0070989900

1010180 SKA38X5 0070989900
Category Type Screw Size Pitch/TPI Head Style Length Drive Material
Screw Machine screw 10 1mm Cylind 30 HEX SS
Screw Machine screw 10M 1 Cylinder 30 Female H 304
Screw Machine screw 10 1 Straight 1-1/4 HEX 316

Screw Machine screw M10 1mm Cylindrical 30mm Female Hex Stainless Steel
7
De Duplication Screen
8
Classification

• Classification to one organizational taxonomy
  will enable
• Fast searching capabilities over all cluster
  records
• Better management of the records
• Ability to compare, review and select among
  similar products
• The refinement process also allows the
  identification of duplicate records
• The Classifier is used to classify the object
  description into any given taxonomy
• Input data
• Product descriptions (or any other object)
• Taxonomy tree
• Data to be utilized
• Knowledge bases
• Desired output
• Categorized data

9
Classification Screen
Descriptions
Classification Tree
10
Extractor
The Extractor module is used to extract data from
a free textdescription or additional fields into
tabulated, organized data

• Extracts data from case descriptions
• Uses defined patterns (or new patterns) to
  identify values within cases and extract these
  values into defined attributes

• Extracts data values from descriptions into the
  correct attribute
• cbn 2875090 number 3 5 x 1.25 inch x 0.25 inch x
  3 17/200 grit flat
• borazon wheel 1a1 125 x 31 x 6 x 76.2 b91 g86743
• sinter cbn 7734-389E 120 X 30 X 5 X 3 inch b76
  1a1
• cbn 2976-999 number 2 flat 4.5 X 1.5 inch X 0.5 X
  3 inch grit 200/23

Item Category Abrasive material Tool shape Dia. Thickness (axial) Rim depth A.H diameter Grit size Grit shape
1 Sintered CBN Wheel CBN 1A1 5 inch 1-1/4 inch 0.125 inch 3 Inch 170/200 mesh
2 Sintered CBN Wheel CBN 1A1 125 mm 31 mm 6 mm 76.2 mm B91
3 Sintered CBN Wheel CBN 1A1 120 mm 30 mm 5 mm 5 mm B76
4 Sintered CBN Wheel CBN 1A1 4.5 inch 1-1/2 inch 0.5 inch 3 inch 200/230 mesh
11
Extraction Panel
Attributes
Product Family
12
Product Comparison
13
CASE STUDIES
14
Creating an ecommerce portal
Client

• A leading publisher in North America

The Need

• To take advantage of the internet and use it as a
  channel for potential sales
• To build a commercially successful parametric
  product search based portal serving three
  industry verticals in the industrial product
  space
• To categorize, parameterize the huge repository
  of product models and categories to deliver a
  superior user experience which would potentially
  attract retain customers
• Need to focus management efforts on developing
  brand visibility marketing the product whilst
  simultaneously creating/maintaining an online
  presence

15
Creating an ecommerce portal

• 1st Phase
• Database building Scope created an online
  database of close to 200,000 product models
  across 550 different product categories across
  manufacturing, construction and life science
  sectors.
• The project included
• Identification of 15 to 20 important parameters
  for a product categories
• Selection of searchable parameters for parametric
  comparison across companies
• Normalization of units and other information
• 2nd Phase
• Editorial Support creation of meta data for
  product categories 2000 landing pages
• Quality Control Client outsourced onsite quality
  control for all their offshore vendors to Scope.
• 3rd Phase
• Marketing Support web analytics, reporting,
  traffic analysis, SEO creation of keywords,
  cross-linking directory submission services

16
The bottom line

• Financial benefit
• The client saved more than 50 of their resource
  hiring costs infrastructural support costs by
  outsourcing the data, editorial support
  marketing analytics to Scope
• Intangible benefits to the client
• Time to market building an online database with
  250, 000 companies within 4 months
• Cleansed the database removed B2C and other
  irrelevant entries
• Provided URL link to the client by which they
  were automate the process
• Freeing up client time to focus on market facing
  and other activities
• Developing organization-wide knowledge base w.r.t
  offshoring of content-related services what can
  be outsourced what cannot.
• QC support client had set up qc team in our
  premises with Scopes employees to do qc of our
  and other vendors work
• Testing capabilities of offshore vendors in
• Scaling up
• Flexibility innovation
• Handling quality control onsite
• Organizational learning developing capabilities
  in the area of searches

Client were so impressed with our service that
they recommended us for the award of service
provider of the year with DPA which we won in the
year 2005
17
Web solutions for Industrial Companies

• Client
• A leading publisher in North America
• Project
• The projects aim is to create comprehensive
  keyword enriched online content for custom
• manufacturing companies, supported by with
  in-depth research and analysis. The main purpose
• of the content is to aid users, which are mainly
  engineers and technical personnel, in making an
• informed purchasing decision. The content is
  created using the best SEO practices to also
• help improve the online visibility of the
  manufacturing company and consequently increase
  the
• number of visitors to their site. The emphasis on
  engineers and technical personnel makes the
• content quite technical in nature.
• Input from the client
• To aid us in creating the content, we receive a
  number of different sources of information, or
  input data,
• from the clients customer including
• Call Audio Files
• Customers Website
• Other Materials provided by the Customer (PDF
  Files)

18
Web solutions for Industrial Companies

• Production Process
• The process begins with extracting information
  from the customers existing site for a
  particular capability.
• This is followed by information being extracted
  from the numerous input data sources, which the
  client may have provided.
• If information is still inadequate for creating
  content, comprehensive research into the
  particular capability from sources such as
  Answers.com and Wikipedia are performed as well
  as extracting information from competitors
  websites offering similar products or services.
• Content is created for each capability, which
  includes a description and an attribute table.
• The description is essentially a keyword rich
  write up on the services or products that are
  provided by the company, including any
  specialized features that are unique to the
  customer.
• The attribute table on the other is a table
  listing the capabilities of the customer in a
  summarized format.
• This content is fact checked for technical
  accuracy, followed by editing for grammatical
  accuracy and conformity to clients style needs.
• Content is then uploaded into a client provided
  tool.

19
Web solutions for Industrial Companies

• Challenges Involved in the Custom Navigator
  Project
• Amount of information provided varies, with
  certain companies providing too much and certain
  others providing insufficient information.
• To ensure that no plagiarism occurs.
• Ensuring uniqueness in the content, especially in
  cases where various companies provide the same
  service or product.
• Issues with clients style requirements.
• Factual and technical accuracy of the content.
• Different Content Engineers, having different
  styles, clients, and requirements
• Output Data
• Using all the input sources provided along with
  additional research, we produce unique content
  for each
• capability with the corresponding keywords
  included for SEO. In addition, attribute tables
  for summarizing the
• company's capabilities are provided as well. This
  content is uploaded into a client provided
  software tool that
• directly creates web pages on the clients
  servers.