PRELIMINARY STUDY

Phase PS.2 Conception 

PS.2.1 Purposes

 The purpose is to conceive the new solution in the great lines: to make proposals for

the processes.

the organization: work and responsibilities distribution, automation.

the technical architecture

the evolution scenarios

 You should stay at a "macro", conceptual level.

In particular, you should only take the 'normal' representative procedures into account, as selected during the previous phase.

PS.2.2 Realization actions

 If you are looking for a package

 The conception activities should be reduced to the essential: functions, processes and data. The purpose will be to have a reference inventory to will be used to make sure that the packages answers the requirements.

 Take preliminary contacts to be informed about the possibilities, about the functions usually performed, about the usual management rules.

 Principles & management rules

 remind the principles & management rules;

 implement the principles & management rules under a systematic way

flow charts

decision tables

 as much as possible, you should describe the rules as a set of parameters, that can smoothly develop.

   Web Site Contents 

 Design your site as a function of customers requirements, expectations and behavior

 Make your site please your customers ? According to customers profile, choose quality criteria :

• customization : let the customer define his profile and expectations, and receive only information pertinent to him
• efficiency : do not provide too much information, too many distractions ; show clearly your purpose ; allow the user to perform efficiently the process he prefers
• speed : no useless graphics, minimize the interaction between the browser and the server (e.g. form populating should be done on server-side whenever possible)
• permanency : create a lasting relation between the customers and the company

To achieve these purposes, it can be necessary to design several applications (or site areas) following the various customer types.

It is often not enough to port existing the local IT applications (e.g. ones used by the sales department) to web applications. An internet application should be intuitive; it should be successfully usable by anybody without having read any other documentation than the instructions provided on the main screens. Often the users will even not read these instructions !

Security policies should also be strenghtened for everyone's use. Avoid a customer to give a discount to himself ...

 Manage loyalty. Remote consumers are expected to be far less linked to one company. They will very easily switch suppliers. They could select a product on your site because it is well designed, and purchase it somewhere else because it is a little less expensive.

• promise only what you can keep : be sure to deliver the product on time and fully, limit your offer
• create a system to track the problems; propose a good sales support service
• • in case of problem, you will be able to propose a new try to the injured customer, with a very attractive sorry-discount
• provide loyalty incentives

 Provide the information your customers need to make a purchase. The process for a customer to arrive at this decision is typically a sequence of five phases :

• need identification
• research among available products
• product selection
• product purchase
• (buyer's remorse)

Provide your potential customers with the information they need to move through each phase.

   Web Site Contents Management 

 Contents Structure; web site map

Describe the browsing structure of the web site as a tree , or a set of trees, based on a home page.

If necessary, provide each market segment by its own, customized, content structure.

  Contents management 

 Starting from the site map, list & organize the contents to be developed; for each topic, describe :

Targeted market segment

Autorship: who will be responsible for its creation & maintenance ?

Coordination with the other topics

• Release frequency
• Contents coordination process

Information origin: select the way you will manage data and other information type

• data from a database
• • See the Relational Data Model to know how to organize your data within a database
• other information type
• • files (text documents, spreadsheets, graphics, ...)
  • texte HTML & XML

See further Web site technical architecture to learn how to best manage the information, how to mix data with other information type, and how to generate HTML code.

   Web Site promotion 

 Prepare your site referencing

  Search engines are very important tools to promote your site; prepare as from now your site referencing

  Checking your competitors on search engines will help you to refine your target and to assess the volume of your expected visitors & benefits

 Find the right keywords

• find in which expressions keywords are usually used
• find the score of similar keywords and expressions
• select the keywords and expressions that you will target
• have a look to your best competitors site

  Prepare your site to be attractive for the selected keywords

• You will probably require additional optimized 'doorway pages'
• • It is particularly important when you use data and text integration, or HTML generation, as this information cannot be referenced by search engines
• Fix your site structure: it will be very hard to change it once pages will have been referenced by external search engines and visitors

  Evaluate the expected workload, costs and benefits of the different possible referencing strategies

• free / paying
• manual / automated
• general / customized by search engine

  See also our web development useful links to find referencing tools and detailed methodologies

 

 Evaluate the opportunity to run an advertisement campaign

  Evaluate the expected costs & benefits from an advertisement campain

  Advertisement channels :

• online
• traditional advertisement
• press

 Plan the tracking of information on your visitors activities

  Make sure that activity statistics will be available

  Evaluate the expected workload to analyse the statistics

  Evaluate the opportunity to provide an internal search engine and to store the used keywords in a database

  Evaluate the expected costs and benefits of external online surveys

 Applications and systems

  Describe the main characteristics of the applications and systems

functions; functions hierarchy .

inputs: mails, forms, data (including from existing system) , documents, ...

outputs: to user agent (browser), to existing system (interface)

main actors.

Data Flow Diagram ( /Application )

 You should give a particular attention to the interfaces, this is to the objects that are the output of one system and the input of another one.

 During the systems description, the data can still be mentioned in a schematic way. (data store)

  Processes: describe the system functions

The functions description should follow the same rules as the applications description.

The same should be done at a thinner detail level.

The functions description replaces favorably the processes description [ Conceptual Processes Model ] of the Merise methodology. Indeed, the functions describe the processes to be performed, but independently from the technical and organization choices, and without requiring all operational details.

Data Flow Diagram (/Function)

To generate HTML code from data:

  Always separate information contents from information presentation

Abstraction should be sufficient to perform multiple layouts from a single information source but does not need to be as deep than in a data normalization.

 A database can also include pieces of information

HTML and XML code (free text)

links to external files

 Describe also in the Conceptual Data Model the information that will not be included in a database

  Give these pieces of information a structure

 Describe the data [ Conceptual Data Model ]

 Begin by setting up the inventory of the data to be processed (data already or not yet computerized)

 Generalize the data up to an abstract conceptual level, that will be customizable and evolutionary.

  Avoid a data model based upon particular cases or instances

 Structure the data following a pure normalized form, without taking into account, at this stage:

neither the technical constraints of distribution, access speed, redundancy, archiving, ...

nor the "psychological" aspects: human interfaces, access rights, views, etc

   Apply the Codd's normalization rules

"The basic ideas in normalization are to organize the information in a database as follows:

 Each distinct type of object has a type identifier, which becomes the name of a base relation

 Every distinct object of a given type must have an instance identifier that is unique within the object type; this is called its primary-key value

 Every fact in the database is a fact about the object identified by the primary key

 Each such fact contains nothing other than the single-values immediate properties of the object

 Such facts are collected together in a single relation, if they are about objects of the same type. The result is a collection of facts, all of the same type.

Note that this methodology makes no distinction between abstract objects and concrete objects. Furthermore, no distinction is made between entities and relationships."

Codd, "The Relational Model for Database Management", version 2, 17.5.1

Entity Relation Diagram

Entity Relation Diagram

 Organization

 Describe in the great lines, without entering into procedure details:

Automation choices

Communication ways

Impact on the organization

PS.2.3 Project management actions

 Estimate the selected solutions(s) in terms of

• costs, needed resources
• benefits
• planning
• risks

PS.2.4 Quality assurance actions

 Control by a second reading, simple or alternate.

 Key factors: make sure that all requirements find a solution

Note: at this stage, the solution is not yet completely described, the data inventory is not yet complete, etc. It is thus a first evaluation that needs being completed during the next phases:

Preliminary study - Generalization

Detailed study - conception

 control the functions & processes, by crossing them with the data model (data used by functions)

 check systematically that the requirements from the preliminary study will find a solution

 evaluate

the requirements that would not find a solution

the possible limits, particularly in organization terms (automation level).

control the data

 make sure that the data inventory is mode under a systematic form .

 output instances tables showing how the conceptual data model can represent the physical data (important to get the users understanding and motivation !).

 this inventory can be the base for the tests games and the basic customizing.

PS.2.5 Phase closing actions

 At the end of this phase, you should choose one or several solution(s)

 at the functional level

What needs being done ?

 at the organizational level

Who does what ? How ? Main characteristics.

Services in charge

Automation ways.

 This solution covers still only the normal representative cases .

 The decision can be taken when having already a good understanding of the technical, budgetary, security and planning consequences.

 Check that the selected solution corresponds to the fixed purposes (or adapt these).