Software maintenance services India: what it costs

Most mid-market Indian companies do not buy software maintenance so much as inherit it: a dealer portal built in 2019, an operations dashboard a former employee wrote, an ERP add-on that two people understand. The bill arrives quietly — a licence lapses, a payment gateway retires an endpoint, a GST schema version moves — and a working system becomes a Friday-evening production incident. The companies that handle this well are not the ones with the biggest IT team; they are the ones who put a name, a scope and a monthly number against maintenance before something broke.

What do software maintenance services in India actually cover?

Maintenance is the ongoing work of keeping a live system correct, secure and compatible — it is not new feature development. In practice, an Indian annual maintenance contract (AMC) or monthly retainer usually bundles five things: defect fixes against a severity ladder, security patches and dependency upgrades, hosting, backups and uptime monitoring, a fixed monthly allowance of small change requests, and documentation so the system is not hostage to one person. What sits outside is anything that moves the product roadmap — a new module, a new integration, a redesign. Those should be quoted as separate projects with their own timelines, because folding them into a retainer is how maintenance budgets quietly double.

Worked example: what ₹18 lakh a year actually buys

An ₹18 lakh annual retainer typically buys a small team's part-time attention: an 8x5 severity-based SLA, a monthly patch cycle and a few days of change work each month. The figures below follow a pattern we see often rather than one named client. A Gurgaon auto-components distributor with roughly ₹40 crore turnover runs a dealer order portal and an internal order-management tool: 260 users across six branches, Laravel and MySQL, two application servers and one database on AWS Mumbai, integrations with Tally Prime, the GST e-invoice IRP, a payment gateway and WhatsApp order confirmations. The original vendor left in 2022.

Three quotes came back. A freelancer at ₹40,000 a month, part-time, no SLA. A mid-size agency at ₹1.5 lakh a month with the 8x5 SLA above and one small enhancement per quarter. An in-house developer at ₹14–18 lakh CTC plus roughly 15% in overheads and several months of hiring. They took the agency retainer — ₹18 lakh a year. Year one produced three Sev-1 incidents, two of them from an upstream API schema change, with a median time to restore of about four hours.

What drives the price of an AMC in India

Four variables set the price: stack age, integration count, SLA window and included scope. A stable Laravel or Node codebase on managed cloud typically lands between ₹60,000 and ₹1,50,000 a month for a mid-market business system; older ASP.NET Web Forms, heavily customised ERPs or systems with no documentation sit materially above that. In our experience the gap between the cheapest and dearest quote for the same system is often 3–4x, and the cheap quote is cheap because it excludes something.

Watch for these in any quote:

• Response time versus resolution time — a 4-hour response with no resolution commitment is not an SLA.
• Whether third-party API upgrades are in scope: WhatsApp Cloud API versions, payment gateway deprecations, GST schema changes.
• Whether you get the repository, deployment scripts and a working staging environment at the end.
• What the notice period and handover clause say. Exit should be boring.

Why do unmaintained systems fail all at once?

Most production failures are scheduled, not random: they land when a dependency, certificate or upstream API reaches end-of-life. A TLS certificate expires because nobody owns the renewal. A framework version stops receiving security patches. The payment gateway retires an endpoint on 90 days' notice sent to an old email address. Underneath all of it is people risk — the one developer who knew how the deploy worked, and left.

The fix is unglamorous and cheap: an asset register listing every system, its owner, its stack, its integrations and its renewal dates. If you do not have one, make it the first deliverable you ask a maintenance partner for, and read our notes on application architecture before you sign anything.

How should you scope and monitor a maintenance contract?

Put the scope in writing, with severity definitions and a monthly report you actually read. The contract should state what counts as Sev-1 (orders or revenue blocked), response and resolution targets per severity, the monthly ticket allowance, the change window, and who owns backups and how restores are tested. Then hold the partner to a one-page monthly report: tickets opened and closed, patch status, incidents with root cause, and hours used against the allowance.

Two practical additions. Ask for a staging environment from day one so fixes are not tested in production. And keep hosting visible — if the partner controls the cloud account, you cannot move. Our server and hosting notes cover the handover checklist we use.

Where to start

Start with the asset register and one month of honest ticket data; SLA, price and scope all follow from those two things. Maysan Labs runs a small number of maintenance retainers for systems like the one above, usually alongside building whatever replaces them.