Guides
How much does app maintenance cost
Prepared by: improveIT.pl
Plans, hosting, updates and development: what makes up the monthly cost.
How much app maintenance costs per month: plan, hosting, monitoring, security updates and development. Prices excl. VAT and what to put in the contract.
App maintenance at improveIT costs USD 225–1,750 a month excl. VAT (PLN 900–7,000). The lower end covers updates, backups, monitoring and small fixes for a small app; the upper end covers a larger system with faster response times and development hours. Hosting, licences and third-party services are paid separately. Prices are converted from PLN at a fixed rate of 4 PLN = 1 USD and exclude VAT.
| Service | Example task | Do not automatically assume |
|---|---|---|
| Hosting | Providing application resources | Fixing source-code defects |
| Monitoring | Detecting a defined signal | Round-the-clock handling of every alert |
| Backups | Stored copies and agreed restoration tests | Zero data loss in every incident |
| Technical care | Updates and specified package work | An unlimited work allocation |
| Development | A new feature or integration | Inclusion in the hosting charge |
What maintenance includes
Maintenance is more than fixing bugs. A well-defined plan protects the app from problems before users see them.
- Framework and library updates
- Security fixes
- Backups and test restores
- Uptime and error monitoring
- Small changes and user support
- Development hours
What the price depends on
The price grows with the size of the system and the expected response time.
- Number of modules, integrations and environments
- Required response time for incidents
- Development hours per month
- Condition of code and documentation at takeover
- Security and audit requirements
Costs outside the plan
Some costs do not depend on the contractor.
- Hosting or cloud: from USD 25 a year for simple hosting
- Domain USD 15 and SSL certificate USD 3.75 a year
- Apple and Google developer accounts for mobile apps
- Third-party licences, e.g. maps, payments, AI
Taking over an app from another contractor
If someone else built the app, we start with a pre-takeover audit (USD 875–2,100). We check access to code, environments and backups, then propose stabilisation and an ongoing plan.
- Audit of access, code and infrastructure
- Rebuilding the deployment process
- A plan of fixes and stabilisation
- An ongoing maintenance plan
What to put in the contract
A good maintenance contract says clearly what you pay for and what you can expect.
- Work included in the plan and outside it
- Response times by severity
- Rules for unused hours
- A monthly report of completed work
- Ownership of code and access after the contract ends
Define what maintenance means for your business
Maintenance can mean paying for a server, updating dependencies, monitoring, user support or incident response. Do not assume one monthly amount includes every activity. Start with the system inventory and what happens if each component fails. Can staff postpone a task until tomorrow, or does the business lose the ability to process orders? Then agree support hours and required work. A small information website needs a different model from a payment-enabled operational application. An estimate should expose those differences rather than hide them under the word care. Only then can proposals be compared fairly and the package assessed against real business risk. The aim is an appropriate responsibility model, not simply the lowest invoice.
- Hosting provides resources to run software; it does not automatically include administrator or developer work.
- Monitoring detects agreed conditions, while the response to an alert must also be specified.
- Dependency updates can require tests and code changes, so not every update is a trivial task.
- User assistance explains how to operate the system; product development adds capabilities. Price them separately.
Build a monthly budget without counting the same service twice
Use four categories: infrastructure, external services, agreed technical care and development changes. Infrastructure includes server resources, databases and file storage. External services might bill for messages, maps, payments or AI usage. Care covers particular responsibilities and available team capacity. Development is a separate change list. Costs can be fixed, usage-dependent or one-off. Do not compare a care retainer alone against another supplier’s complete operating cost. Also identify overlap: backups or updates may already be included in a package. The same activity should not be charged again without a distinct additional scope. A clear inventory makes the budget easier to predict and review when traffic or business requirements change.
- For each cost, record the provider, billing period, limit, payment method and account owner.
- Include renewals in the annual budget even when invoices arrive only once a year.
- Care from PLN 49 excl. VAT per month, shown as USD 12.25 at the service conversion, is a specified website service, not unlimited application maintenance.
- Work limits, supported environments and additional-work rules should be clear before an agreement starts.
Response time is not a promise to repair every incident immediately
When an application fails, the team first confirms symptoms, assesses impact and establishes who should act. Response time describes the start of agreed handling; it does not guarantee that any underlying cause can be repaired within the same interval. Recovery may depend on a cloud operator, payment provider or access to a customer system. Define severity levels, support hours, reporting channels and progress communication. A service unavailable to every user differs from a typo or one account problem. With a modest budget, a narrower realistic commitment is better than nominal round-the-clock readiness nobody operationally provides. The rules should also be understandable to the nontechnical person reporting the incident.
- Identify who can report an incident and approve actions affecting data.
- Distinguish business-hour support from any separately arranged evening or holiday availability.
- Agree progress-update frequency and escalation when the cause is not quickly established.
- Check the support terms of external dependencies instead of promising control over another provider’s infrastructure.
A backup matters only if it can be restored
Backup enabled does not explain how the business returns to work after data loss. You need scope, frequency, retention and a restoration test. Decide how much recent work the business can lose and how long it can wait for recovery. Choose the implementation from those requirements, not the other way round. A database backup alone may be insufficient if files live elsewhere or necessary configuration was never recorded. Check access to keys, accounts and instructions too. Restore testing belongs in a controlled environment without overwriting production. Record the result as evidence rather than substituting a provider statement. This makes recovery an operational capability rather than an unchecked setting.
- Include databases, files, required configuration and secure recovery of secrets.
- Specify acceptable data loss and recovery time as business requirements.
- Assign an owner for restoration tests and review them after material architecture changes.
- Check backup-deletion permissions and separation so one mistake does not remove both production data and its protection.
Updates and security need a planned working window
Dependencies, operating systems and external libraries change even when no new features are commissioned. An update can fix a defect, address a vulnerability or adapt the application to an API change. Before release, assess impact, testing and rollback. Not every version needs immediate deployment, but deferring every update without review can create a difficult large upgrade later. Agree a review cycle and urgent-issue rules in the care plan. For an application without a test environment, creating one may be the first sensible task. Otherwise a saving in infrastructure cost transfers testing risk onto the live business. Security work should have defined scope and evidence rather than an impossible promise to prevent every attack.
- Inventory components and critical dependencies to assess whether a vulnerability notice affects the system.
- Review account permissions, integration keys and former collaborators’ access at agreed intervals.
- Check logs and alerts against defined signals rather than claiming universal threat detection.
- Separate immediate incident containment from root-cause analysis and a permanent correction.
Taking over from a previous supplier
A new team needs to understand what it is supporting. Handover includes source code, environments, integrations, accounts and known problems. It does not require rebuilding everything before the first invoice. It does require establishing whether changes can be released safely and the service restored. Where documentation is incomplete, agree a stabilisation stage and priorities. Secure access and the critical workflow first, then less urgent improvements. The customer should control repository, infrastructure and supplier agreements. Without them, a takeover may begin with an access problem that cannot be solved by writing more code. A useful plan ends with responsibilities and a practical reporting process, not just a technology inventory.
- Provide instructions for running, deploying and rolling back the application, plus the environment list.
- Prepare API documentation, supplier accounts, billing and renewals without putting secrets in ordinary messages.
- Record known limitations, outstanding fixes and the most common user symptoms.
- The first assessment should state priorities, costs and risks so the operating model can be agreed.
A maintenance report should support decisions
A useful report answers what was done, what was discovered and what requires a decision. A list of hours without context does not show whether the application is better prepared for operation. Include incidents, updates, backup checks, capacity usage and upcoming renewals. Compare the same measures over time where possible. Alert volume alone is not evidence of quality; relevance, user impact and handling matter more. Maintenance should reduce unplanned interruption and improve predictability rather than generate the greatest possible number of tickets. Development decisions should reflect actual needs and support costs. Reporting remains compact when it is organised around those business questions instead of copying raw technical logs.
- Agree a few understandable measures such as important-feature incidents, downtime and recurring problems.
- Explain infrastructure cost increases through traffic, data, configuration or a new service.
- Highlight approval-dependent work and the risk of postponement.
- Review the package when user volume, business criticality or usage changes materially.
FAQ
How much does mobile app maintenance cost?
About the same as a web app, plus updates for new iOS and Android versions and store releases. At improveIT maintenance costs USD 225–1,750 a month excl. VAT.
Can I pay only for actual requests?
Yes, hourly billing without a plan is possible. Without a plan, however, nobody keeps updates and backups current.
What if the previous contractor will not hand over the code?
We first check the contract and access to the repository and servers. If the code is unavailable, we assess whether it is faster to recover it or rebuild the key features.
Does monitoring mean an incident is repaired automatically?
No. Monitoring detects an agreed condition such as an unavailable endpoint or a failed task. Automatic recovery only covers scenarios deliberately implemented beforehand. Other cases require assessment and an accountable person. The agreement should state who receives alerts, response hours and package activities. Access to an availability chart is not equivalent to a round-the-clock on-call service or a guaranteed repair time for every possible cause.
What should I prepare before handing over application maintenance?
Prepare the environment inventory, repository, run and deployment instructions, backup description and integrations. Identify account owners and the person approving costs. Record known defects and important user workflows. Transfer secrets through an agreed secure channel. When material is missing, the incoming team should assess the state and propose a bounded stabilisation stage. Documenting known problems helps choose an appropriate support scope instead of hiding uncertainty.
Can a maintenance package include new features?
Yes, where an agreed work allocation includes them, but this should not be assumed. A new report, integration or user role changes the product. Correcting a defect restores previously agreed behaviour. Distinguish those cases before work begins. With a modest budget, prioritise one valuable change and defer the rest while preserving the operational duties needed to keep the application working safely.