Guide

How to scope an application maintenance contract

What to settle before handing over your application maintenance, and what derails a contract.

By the Ingenit team

An application maintenance contract works when everyone knows what is covered, within what timeframes, and how the service is measured. Most difficulties come from what was never written down. This guide reviews the points to settle, in the order they come up.

What is an application maintenance contract?

An application maintenance contract, known as TMA (tierce maintenance applicative) in France and Morocco, entrusts a provider with maintaining one or more applications for a set period. It specifies the applications covered, the types of work, the service levels, how requests are submitted and tracked, how the service is billed and the exit terms. It can cover an application built by the provider itself or by a third party; in the second case, a takeover phase comes before the service starts.

What application maintenance covers

TypePurposeExamples
CorrectiveFix a defectA wrong calculation, a screen that no longer loads, an incomplete export
EvolutiveAdd or change a featureA new field, a new business rule, a new report
PreventiveReduce risk before an incidentUpdating a library, a certificate or the operating system

Preventive maintenance is often left out of contracts, even though it prevents a good share of incidents and security flaws.

Set the scope

List what is covered, and also write down what is not:

  • the applications and their versions;
  • the environments: production, staging, development;
  • the interfaces with other systems and the data exchanges;
  • hosting: who administers it, who monitors it, who backs it up;
  • exclusions, for example workstations or third-party software.

Application maintenance and managed IT services often overlap on hosting: state who does what, to avoid incidents that nobody picks up.

Define service levels

Rank requests by priority, then set for each one a response time, a workaround or resolution time, and the hours covered.

PrioritySituationCommitment to set
BlockingThe application is down or a critical function has stoppedFast response and workaround, over extended hours if needed
MajorAn important function is degraded, with a workaroundResolution within a short timeframe, planned with you
MinorLimited inconvenience, no effect on operationsFix in an upcoming release

The timeframes themselves depend on your working hours and how critical the application is. What matters is that they are written down, measured and tracked: simple commitments that are kept are worth more than ambitious promises that are never checked.

Plan the takeover of an existing application

Taking over an existing application happens in stages:

  1. an audit of the code, hosting, documentation and recent incidents;
  2. knowledge transfer with the outgoing team or provider;
  3. a period where the new provider works under supervision;
  4. full takeover, with the service levels set in the contract.

This is what our application maintenance services cover.

Write reversibility in from the start

Source code, access, documentation and procedures must be able to be handed over to your team or to another provider, within a timeframe and under terms set in the contract. Also plan a knowledge transfer period and the list of items to hand over: code repositories, deployment scripts, technical accounts, operations documentation.

Govern the contract

One lead on each side, regular meetings, a request tracking tool and shared indicators. The most useful indicators are simple:

  • response and resolution times, by priority;
  • the number of requests opened, handled and pending;
  • defects reopened after a fix;
  • consumption of the change budget;
  • the status of security updates.

Fixed price, subscription or time and materials?

Application maintenance is most often billed by subscription, for a defined level of activity, with a separate budget for changes. Pay-per-intervention suits applications with little activity; time and materials suits teams that want to manage the work themselves.

Mistakes that derail a contract

  • An implicit scope that each party understands differently.
  • Service levels with no measurement or reporting.
  • No acceptance testing before changes go into production.
  • No budget for preventive maintenance.
  • Documentation that is no longer kept up to date.
  • Reversibility pushed back to the end of the contract.

For an application to build before maintaining it, see custom software development. To scope your maintenance contract, let’s talk.

A need on this topic?

Describe your context in a few lines: we come back to you with a first reading.

Book a call