ABOUT US > Consulting Project Lifecycle Models


Our Approach
Quality Assurance
Privacy Policy

Best Viewed
Employment
Contact Us
--------------------------------------

 

 

Consulting Project Lifecycle Models
--------------------------------------------------------------------------------------------------------

Appian Analytics' data intensive services and software development projects are conducted using Project Delivery Lifecycle Models that help to enhance success. A Delivery Lifecycle Model prescribes the tasks, method, and sequence in which to complete a project. There are many types of project lifecycle models, the simplest of which is sometimes referred to as the “Waterfall” approach. This model is characterized by a strict sequencing of project work steps.

In the Waterfall lifecycle, each phase of the project, including its work steps, occur in a strict time sequence. That is, for each given phase, each set of work steps within the phase are completed only after prior work stops are completed. For example, the first phase of a project might be thought of as the Planning Stage. In the Waterfall Lifecycle Model, activities in Phase Two may not take place until all the work steps in the Phase One (Planning Stage) are completed.

Oftentimes, software development, research, and other complex projects will be organized around a particular type of lifecycle model. The reason is that each lifecycle model offers certain virtues by nature of the prescribed method that work is organized and completed. The result is that each model may offer certain benefits and drawbacks for any given project.

In order to maximize project success, each Lifecycle Model must be selected with consideration of:

  1. Type of project undertaken
  2. Level of project risk and uncertainty
  3. Speed with which deliverables are required
  4. Resources available

Project Lifecycle Models

Lifecycle

Description

Benefits and Drawbacks

Waterfall

Project steps are completed in strict sequence.  Binary Milestones are identified at the conclusion of each phase.  Formal identification of Binary Milestone parameters for conclusion in order to ensure appropriately completed work phases.

 

This model helps to minimize risk and uncertainty by being very careful and methodical in the completion of project tasks.

The drawback is that this method requires a great deal of resources devoted to project management and oversight.  Projects must be maximally documented and planned. 

This method can minimize the risk of missed deadlines and cost overruns.  However it can be slower than other
methods for completing a project.

The Loop

Projects are broken up into smaller, more manageable "chunks".  These chunks are completed using a set of steps that are continually completed for each chunk until the entire project is done. 

Step categories include:  Planning, Trade-offs Analysis, Risk management, Prototyping, Evaluating, Building, and Testing.

This method is methodical but also flexible.  It requires less planning and oversight overhead while still maintain good project visibility.

A disadvantage is that it requires an experienced project manager with good knowledge of the industry.  This is because a great deal of judgment is required in order to guide the project to maximum project success.

Prototype to Solution

You develop the solution while you work.  As an example, a client would start with a research project goal in mind.  The project could begin by exploring certain initial hypotheses in the data.

This model would research the questions and then consider the results.  These results then would then feed the next hypothesis, the answers of which would direct more research.

Very practical approach with a maximum amount of flexibility.

May be a bit too "seat of the pants" and whimsical for very formal research projects especially scientific.

Reasonable approach for some small software projects where the business and end-user requirements are well understood.

Staged Delivery

Staged Delivery means that certain intermediate work products are provided to clients before the entire project is completed.

In general, this approach should be used when some level of the functionality is required as soon as possible.  Of course, this would depend upon that functionality being ready and appropriately quality assured for the use intended.

Indispensable when projects will yield immediately needed deliverables during the course of project completion.

In software development, continually maintaining debugged and executable code can be imperative to developing a software product with minimal risk of rejection by end-users.

The disadvantage would be that the project team must deliver products that are ready to use throughout the entire project process.  This adds cost and some level of added complexity.

RAD Development

"RAD" which stands for Rapid Application Development is generally considered a software development paradigm.  While this is true, the elements of RAD approaches can be quite useful to other database, research, and analysis projects.

For software projects, RAD approaches are characterized by the use of software object "building blocks."  These building blocks may be as simple as reusable code and procedure libraries, to as complex as entire application modules. These application models sometimes called, "Components" can be integrated into other software applications and make use of the features available by using interfaces that are accessed programmatically.

In other projects, this approach might leverage business analytical framework libraries, quantitative models, statistical processes, or core databases.

Certain software development approaches draw upon RAD techniques heavily. Examples include Extremeprogramming and
X Programming.

RAD has the advantage of being able to leverage reusable modules that offer a great deal of functionality quickly and cost effectively.  Under the RAD approach reasonably tailored solutions can be obtained with much less effort than that of a complete custom project.

The disadvantage is that there is a level of inflexibility to using either a general purpose module or one that was designed for some other use in a custom software or research project.  The building block may not offer the level of customization required for the most optimal solution.

Also, there is a risk that certain disparately created modules may not fit together as well as possible.

Wrap-Around

Similar in function as the RAD approach but completely characterized by the use of an analytical object such as a software application, analytical framework, quantitative model, statistical process, or core database.

The key is that the business research question fits neatly inside an analytical object.  An example might be the use of a special statistical model for predicting electricity consumption for a certain manufacturer.  This object might be used as the "Wrap-Around" solution for a similar manufacturer to fine tune an estimate of electricity consumption.

The Wrap-Around offers an even greater amount of rich features and development speed.   The reason is because the Wrap-Around solution has inherently addressed the most critical issues that the new project must address.

The disadvantage is that there is a high level of inflexibility due to using a general purpose module that must be integrated into the current project.

Also, there is a risk that certain variables unique to a client problem may not be uncovered.  This might happen because the process of developing a complete, custom solution will more faithfully uncover such issues.

Stop wasting time on data and focus on your business!
Outsourcing to Appian Analytics is a snap

 

Copyright © 2009 Appian Analytics All rights reserved.