Incremental Gain

iQGateway’s engagement model for successful business relationships

The Philosophy

Years of experience spread across various industry verticals, we have learnt a great deal in terms of partnering with our clients and building successful business relationships.

The core philosophy – We are in it together as Partners – is extremely important that we ensure the client’s interest is well-protected at all times.

Data Science projects, like any other projects, have their own risks and are prone to failure. There are many reasons why a data science project can fail. How we learn from failure and not lose any money in the bargain is most important.

What can go wrong in a Data Science project?

Data

Data is the most important input for a data science project. Without data there is no project. Some major challenges faced around data are:
1. Availability of data
2. Quality of data
3. Size of data
4. Source of data

Business Problem

The business problem, next in importance to the data, needs to be well-defined. It should:
a. Not be a motherhood statement or wish list
b. Have a binary answer like yes/no, true/false, pass/fail, etc.
c. Be defined by the person who faces it
d. Answer to the problem should have a specific business use

People

It is the people who tie data to the problem and help in getting the solution. A good team understands the problem and comprises:
a. One sponsor
b. One decision-maker
c. Cross functional team members
d. Beneficiaries of the solution

Technology

Technology is only a means to the end and not really the core issue. Decision around technology
a. Should consider that data science is not a technology problem
b. Should not hinder the solution
c. Should allow for flexibility
d. Should be taken in the end

What is incremental gain all about?

Incremental gain makes small-structured investments in a data science project with clear understanding of:

Incremental gain is a partnership-based approach where:

What outcome to expect

The client and iQGateway work for the benefit of each other.

What risks to mitigate

It provides flexibility to stop at a certain milestone, if result(s) not as per business needs.

Safeguarding the organization from further failure and financial burden.

It provides direction for future projects

Let us look at the milestones in more detail along with the deliverables.

Phase 1

Problem understanding and definition

Deliverables
A detailed report of the business problem, its impact on the business and how, on solving it, the business will be helped. The kind of data required, what is currently available and what needs to be collected in future and how long will it take for the problem to be solved.
Phase 2

Problem definition and possible solution

Deliverables
A report on the technical problem definition process, kind of data used and approximate results obtained. A root cause analysis on results obtained and what is required to make them better. Discuss issues around quantity / quality of data / incorrect business problem.
Phase 3

Modelling and Validation

Deliverables
Base model details, data management processes and a report of all findings including test cases. Reasons for not being able to scale up and possible future corrective steps. Suggest Process improvements if they can help in getting additional relevant data.
Phase 4

Pilot

Deliverables
Approach and design for integration with enterprise applications. IT infrastructure design, user requirements documentation, usability inputs, outcome effectiveness for business and test cases for production.