Skip to content
Tutorials Center
My learning

Topics

Lesson 20 of 36 ยท Implementation and Change

Implementation Approaches: Big Bang, Phased and Parallel

Compare big bang, phased, parallel and pilot rollout strategies for ERP and CRM, with their risks and costs, and score the options with a small weighted Python model.

Once the design is clear, a basic question remains: how do you switch the company over? Do you flip everything on one weekend, go site by site, or run old and new systems side by side? Each choice trades speed, cost and risk. This lesson explains the main rollout strategies and how to choose between them.

The product is not available here, so we model decisions in Python. The numbers are illustrative, not benchmarks.

The main strategies

Big bang

All at once Every module and site switches on one date.
Fast One cutover, one training wave.
High risk Any failure hits the whole company.

Needs strong testing and a tested rollback.

1 / 3

Big bang

All modules and locations go live on the same date. Advantages: one integrated system from day one, no temporary bridges between old and new, a single training and communication push. Disadvantages: everything can go wrong at once, the organization has little capacity to absorb the change, and a bad cutover can stop shipping or invoicing. Big bang suits smaller companies with simple structures or situations where the old system has to be switched off by a hard deadline (an expiring licence, a platform end of life).

Phased rollout

The scope is split into waves. Two common cuts:

  • By module: for example finance and procurement first, then inventory and production, then sales and CRM. Modules share data, so temporary interfaces to the legacy system are needed.
  • By site or business unit: the first wave is a template (a standard design for the whole group), later waves roll out the template to other countries or plants with local adjustments.

Phasing reduces the risk of each go-live and lets the team learn. The price is a longer total duration, duplicated effort on interfaces and a longer period of running two systems.

Parallel run

For a period (often one or two month-end closes), users enter transactions in both the old and new system and compare results. It is very safe for finance, where totals must match, but doubles the workload and tempts users to trust the old system. Most companies limit it to critical processes such as payroll or month-end close.

Pilot

A single department, store or subsidiary goes live first. A pilot proves the design with real users and real volumes. The main danger is picking a pilot that is unrepresentative (too small or too simple).

CRM specifics

CRM projects are usually lighter than ERP projects: fewer integrations, less financial control. A common CRM approach is by team: marketing first, then sales, then service, or a pilot with one sales region. User adoption matters more than technical cutover, because a CRM full of empty fields has no value. See change management and training.

A weighted decision model

Choosing a strategy is a multi-criteria decision. Give each criterion a weight that reflects your situation, score each strategy from 1 (poor) to 5 (good), and compute the weighted sum.

Python Score the rollout options (stand-in model)

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Now change the weights to describe a company with a hard deadline (speed 5, low risk 2) and see how the ranking moves. The point of the exercise is not the exact number but making assumptions visible so the steering committee can argue about weights instead of opinions.

Cutover timing and interim costs

Phased strategies have one cost that big bang avoids: interim interfaces and dual running. The next stand-in compares total cost of three strategies using a simple model: a base project cost, extra cost per wave for the cutover team, an interface cost for each additional wave, and the expected loss from a failed cutover (probability times impact).

Python Expected cost with failure risk

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The lesson from the model: more waves cost more fixed effort, but each failure is smaller and less likely. Whether big bang is cheaper depends heavily on the failure probability you honestly believe, which is why rehearsals and testing (testing, cutover and go-live) are so valuable: they lower that probability.

Choosing in practice

Prefer big bang when the company is small, processes are simple, a hard deadline exists and strong rollback plans are in place. Prefer phased by site when you have many similar units and a template design. Prefer phased by module when the organization cannot absorb all change at once or when the modules are technically separable. Add a pilot whenever the design is new to the company. Use parallel running selectively, where numbers must be proven.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. Which is a typical drawback of a phased rollout?

    Phasing lowers the risk of each go-live but needs bridges to the legacy system and stretches the project.
  2. 2. What is a parallel run?

    Parallel running is safe but doubles the data entry effort, so it is usually limited to critical processes.
  3. 3. A subsidiary goes live first to prove the design with real users. What is this?

    A pilot validates the design in a small, real environment before wider rollout.
  4. 4. A company has a legacy platform that is switched off by the vendor on a fixed date and a simple process landscape. Which strategy is a reasonable candidate?

    A hard deadline and simple structures favour a single cutover, provided testing and rollback are strong.

Summary

  • Big bang is fast and simple structurally but concentrates risk.
  • Phased rollouts (by module or by site) reduce risk per go-live at the cost of interfaces and time.
  • Parallel runs prove numbers but double the work; pilots prove the design.
  • Make the trade-offs explicit with weights and expected-cost models.
  • Rehearsals and testing change the odds more than the choice of strategy.

Next: Requirements and Fit-Gap Analysis