Installation and afterwards: conversion strategies, change management, maintenance
Moving from the old system to the new one (direct, parallel, pilot, phased), change management and training, system support, the four types of maintenance and the post-implementation review. 'Which conversion strategy' and 'which maintenance type' questions are certain on the final.
The system has been tested and is ready. But the value of a system shows up not on the day it is installed but on every day it is used. This last week answers three questions: How will we move from the old system to the new one? How will we prepare people for the change? How will we keep the system alive after it starts running? On the final, a comparison of conversion strategies and a classification of maintenance types are almost certain to come.
Golden rules:
- Conversion has three dimensions: style (direct, parallel), location (pilot, phased, simultaneous), module (whole-system, modular). A strategy is a combination of the three.
- Risk and cost are inversely related: direct is cheap and risky, parallel is expensive and safe.
- A technically correct system fails if people don’t use it. Change management exists for this.
- Four types of maintenance: corrective, adaptive, perfective, preventive.
Conversion
Conversion is the process of moving from the old system (manual or computerized) to the new system. When planning, data migration (converting the old data into the new format), hardware installation and user training are considered together.
Dimension 1: style
| Style | How | Risk | Cost | Duration |
|---|---|---|---|---|
| Direct | The old is shut down, the new is turned on | High | Low | Short |
| Parallel | The two run together for a while, the outputs are compared | Low | High | Long |
Direct: old ======|
new |=========>
Parallel: old ==========|
new |=============>
<-both->Dimension 2: location
| Location | How | Risk | Cost | Duration |
|---|---|---|---|---|
| Pilot | One location first, then the others | Low | Medium | Medium |
| Phased | Locations in turn, in groups | Medium | Medium | Long |
| Simultaneous | All locations at the same time | High | Medium | Short |
Dimension 3: module
| Module | How | Risk | Cost | Duration |
|---|---|---|---|---|
| Whole-system | All modules at the same time | High | Low | Short |
| Modular | Modules in turn (first membership, then reservation, then reporting) | Low | High | Long |
For modular conversion the modules must be sufficiently independent of each other; if not, temporary “bridge” code has to be written to make old and new modules talk, which raises the cost.
Selection criteria
- Risk: How critical is the system? What is the damage if something goes wrong? If money, health or safety is involved, parallel or pilot.
- Cost: Is the budget limited? Parallel and modular are expensive.
- Time: By when must it be done? Direct and simultaneous are fast.
Worked example: converting the FitHub branches
Situation: 5 branches. The old system has an operating cost of 12,000 TL per month. If the turnstile system crashes, members cannot enter the gym (high impact). Management wants the conversion within 3 months.
Compare the options:
| Option | Duration | Extra cost | Risk |
|---|---|---|---|
| A. Direct + simultaneous | 1 week | Low | Very high: 5 branches could be locked up at once |
| B. Parallel + simultaneous, 3 months | 3 months | (12,000 + 8,000) × 3 = 60,000 TL | Low, but every transaction is done twice |
| C. Direct + phased, 2 weeks per branch | 5 × 2 = 10 weeks | Low | Medium: one branch is at risk at a time |
| D. Pilot (3 weeks) + the remaining 4 branches simultaneous (1 week) | 3 + 1 = 4 weeks | Medium | Low-medium: problems are solved in the pilot |
Decision: D. In the pilot branch the turnstile and check-in problems are caught, and then the other branches are converted quickly. B is safe but brings an extra cost of 60,000 TL and a staffing load; C takes 6 weeks longer than D.
Change management
A new system changes the way people work. People resist change; this is natural.
Roles
| Role | Who | Task |
|---|---|---|
| Sponsor | The manager who wants the change and has authority | Makes the decision, provides resources, shows support |
| Change agent | The person who carries out the change (analyst, project team, or someone from the organization) | Prepares the users, trains them, supports them |
| Potential adopters | The people who will use the system | The target of the change |
Lewin’s three stages
UNFREEZE -> MOVE -> REFREEZE
break up move to the new habit
old habits, the new is permanent: old
explain why system, system removed,
training new procedureResistance and the solution
Resistance arises from a personal cost-benefit calculation. A system that is beneficial for the organization can make a user lose: the familiar method, expert status (“I knew the old system best”), control, time (learning effort).
Two ways to motivate change:
- Informational strategy: Explain the benefits and show the effect on the user’s own work. It works for changes whose benefit is clear.
- Political strategy: Make it mandatory with organizational authority. It is used for changes that give no direct benefit to the user but are necessary for the organization.
To make the change permanent: reward using the new system, remove the old system (leave no way back), and update procedures and performance measures.
Training
The focus of training should not be “what the system does” but “how the user does their job with the new system”. That is, task-oriented rather than function-oriented.
| Method | Pro | Con |
|---|---|---|
| Classroom | Many people, interaction | Not flexible, everyone at the same pace |
| One-on-one | Personalized, effective | Expensive, slow |
| Computer-based (video, interactive) | Cheap, at the person’s own pace, can be rewatched | Hard to ask questions, motivation may be low |
Post-installation activities
System support
- Online support: Online help, FAQ, a knowledge base, user forums.
- Help desk (level 1): Simple and frequent problems (“I forgot my password”, “how do I freeze my membership”).
- Level 2 support: Specialist support; technical problems that level 1 cannot solve.
- Cases that cannot be solved and are real bugs or new needs become a change request.
System maintenance
Maintenance is the process of changing the system after installation. Change requests come from four sources: users (problem reports, suggestions), the help desk, the system developers (known gaps), and the business environment (legal changes, a new strategy). Each change request is prioritized; large ones go through the SDLC like a small project.
| Maintenance type | Why | FitHub example |
|---|---|---|
| Corrective | Fixing a bug | The crash on names containing an apostrophe was fixed |
| Adaptive | The environment changed | The VAT rate changed; adapting to a new mobile OS version |
| Perfective | A new feature, performance, usability | “Export as PDF” was added; the report was made 2 times faster |
| Preventive | Preventing future problems | The code was refactored; a library whose support is ending was replaced |
PITFALL · adaptive versus perfective. If the user says “I want a new feature”, it is perfective. If the system is adapted because something outside (a law, a platform, a partner system) changed, it is adaptive. The question: “Who triggered the change, a user request or the external environment?”
Project assessment
A two-part post-implementation review:
Project team review: The team evaluates its own process. How accurate were the estimates (function points, schedule)? Which techniques worked? Which mistakes must not be repeated? The results are added to the organization’s knowledge; the next project’s estimates get better.
System review: Did the system deliver the promised value? The benefits and costs in the week 2 feasibility study are compared with the actual ones:
| Item | In the feasibility study | Actual (year 1) |
|---|---|---|
| Benefit | 70,000 TL | 64,000 TL |
| Operating cost | 20,000 TL | 23,500 TL |
| Net | 50,000 TL | 40,500 TL |
The difference is analyzed: why did the benefit fall short (was adoption slow?), why was the cost high (the cloud bill?). That way the loop closes: the assessment can turn into a new system request (a new version, a maintenance project).
Common mistakes
| Wrong idea | The truth |
|---|---|
| Parallel = the cheapest | The most expensive, the safest |
| Pilot = phased | Pilot is one place first, phased is places in turn |
| Direct is always bad | It makes sense for low-risk, small systems |
| Training explains functions | It explains tasks: how you will do your job |
| Resistance is bad faith | It is a personal cost-benefit calculation; it can be managed |
| A legal change = perfective | Adaptive |
| Refactoring = corrective | Preventive (it does not fix a bug, it prevents) |
| The project ends with installation | Support, maintenance and assessment continue |
Test yourself
Exam-style questions. Solve them on paper first, then reveal the answer.
Which conversion style turns off the old system and starts the new one at the same moment?
- Parallel
- Direct
- Pilot
- Phased
Show answer
Direct conversion: the old system is shut down and the new one is turned on. It is the cheapest and fastest but the riskiest: if the new system crashes, there is no working system to go back to.
Which conversion has the lowest risk but the highest cost, and why?
Show answer
In parallel conversion both systems run together for a while and the outputs are compared. If the new system makes an error, the old system continues (low risk). But every transaction is done twice, with extra staff and the operating cost of two systems (high cost).
The old system costs 12,000 TL per month to run and parallel running needs 8,000 TL per month of extra staff. Parallel conversion lasts 3 months. What is the extra cost of running the old system during conversion?
Show answer
The monthly extra cost is 12,000 + 8,000 = 20,000 TL. 3 months × 20,000 = 60,000 TL. This amount is like an insurance paid against the risk of direct conversion.
A gym chain installs the new system in one branch first, fixes problems, then rolls it out to the remaining branches. Which location strategy is this?
- Simultaneous
- Pilot
- Modular
- Direct
Show answer
Pilot conversion: try it in one location, solve the problems, then spread to the others. Phased, on the other hand, means converting the locations in turn, in groups. Simultaneous converts all of them at once.
Classify the maintenance type: 'The tax rate changed by law; the invoice module must use the new rate.'
- Corrective
- Adaptive
- Perfective
- Preventive
Show answer
Adaptive maintenance adapts the system to a changing environment: law, operating system, an external API, a business rule change. There is no error in the system; the world has changed.
Classify: (a) a crash when a member's name contains an apostrophe is fixed, (b) a new 'export to PDF' feature is added, (c) old code is refactored to make future changes easier.
Show answer
Corrective: fixing a bug. Perfective: a new feature or a performance/usability improvement. Preventive: work done so that problems do not arise in the future (refactoring, updating documents, replacing an aging library).
In Lewin's change model, what are the three stages?
Show answer
Unfreeze: break up the existing habits, explain the reason for the change. Move: the transition to the new system, training. Refreeze: make the new habits permanent (remove the old system, formalize the new procedures).
Who is the person in an organization who wants the change and has the authority to make it happen, and who is the person who leads the change effort day to day?
Show answer
The sponsor is the person who wants the change and has the authority (usually a senior manager). The change agent is the person who carries out the change and supports the users (often the analyst or the project team). The potential adopters are the people affected by the change, who will use the system.
Why might users resist a new system even if it is better for the organization?
Show answer
Resistance comes from an individual cost-benefit calculation, not an organizational one. A system that is good for the organization can make a user's job harder or reduce their power. Change management changes this perception through training, communication and incentives.
A pilot branch takes 3 weeks, then the remaining 4 branches convert simultaneously in 1 week. Compared with a phased rollout of 2 weeks per branch for 5 branches, how much time is saved?
Show answer
Pilot + simultaneous: 3 + 1 = 4 weeks. Phased: 5 × 2 = 10 weeks. The difference is 6 weeks. But in the simultaneous part 4 branches carry the risk at the same time; the lessons of the pilot reduce this risk.
What are the two parts of project assessment after installation?
Show answer
Project team review: what the team learned, how accurate the schedule and estimates were, what will be done differently in the future. System review: did the system deliver the expected benefit; the benefits and costs in the week 2 feasibility study are compared with the actual ones.
A user calls because she forgot how to freeze a membership. Who should handle this first?
- The development team
- The help desk (level 1 support)
- The project sponsor
- The database administrator
Show answer
Support is tiered. Level 1 (the help desk) solves frequent and simple problems. What it cannot solve it passes to level 2 (specialist support). If it is a real bug, it goes to the maintenance team as a change request.
Verification: The conversion cost and duration examples were computed with node (verify/SEN2022/14.mjs). The conversion strategy, change management (Lewin's model), maintenance type and project assessment questions were compared with the standard definitions (the Dennis, Wixom and Tegarden framework and the common software maintenance classification).