The apps we ship that are still going strong five years later have one thing in common: their owners committed to maintenance from day one. The apps that quietly died after eighteen months also had one thing in common, which is that they didn't.
This is the most under-glamorised area of software, and the one with the highest ROI over the long term.
What software actually does between releases
A common misconception: an app, once built, just keeps working. It doesn't. It enters slow decay.
- Operating systems change. Apple ships a major iOS version every year. Google ships major Android versions almost as often. Each release moves the goalposts: APIs deprecate, behaviours shift, security models tighten. An unmaintained app starts to feel old within twelve months and breaks outright within two to three years.
- Dependencies move. Modern apps depend on dozens of third-party libraries, each with its own release cadence and security disclosures. Last year's library is this year's vulnerability.
- Devices change. New form factors (foldables, tablets, larger phones) and new screen densities mean layouts that looked perfect at launch start to look wrong eighteen months later.
- Compliance changes. App store policies tighten. Privacy regulation tightens. Apps that worked under the old rules silently become non-compliant under the new ones.
- Your data grows. What worked at 10,000 records doesn't necessarily work at 100,000. Lists scroll slower, queries time out, screens take longer to render.
None of this is dramatic on any given day. All of it is fatal on the timescale of years.
What good maintenance covers
Our maintenance retainers typically include:
- Monthly dependency and OS upgrades with full regression testing. The boring work most clients hate paying for and most agencies hate doing.
- Security patching within agreed SLAs. Critical: 24 hours. High: 5 business days. Medium: with the next monthly cycle.
- Performance monitoring via real-user metrics, not just synthetic tests. We watch for cold-start regression, crash-free percentage, network failure rates, and battery impact.
- Crash and error response. Sentry / Crashlytics monitoring with an on-call rotation. Issues triaged within hours, not weeks.
- Small feature work. The 10% that doesn't justify a project but matters to users. Tweaks, polish, micro-improvements.
- Quarterly reviews. Usage data, technical debt, roadmap conversations. A real check-in, not a status update.
What it costs
A reasonable rule of thumb: 15–25% of original build cost per year for active maintenance. That's the range that keeps an app modern, secure, and visibly evolving.
Less than 15% means you're going to be looking at a rebuild in three years. More than 25% means you're effectively funding ongoing development, which is a different conversation.
For an app that cost $150k to build, expect $22k–$38k a year to keep it healthy. Apps that get this investment last a decade. Apps that don't, get binned.
The two patterns we see
The good outcome
Client invests in monthly maintenance from day one. App stays compliant, gets a small feature every couple of months, accumulates user trust. By year three, it's an irreplaceable part of operations. By year five, the original investment has paid back many times over and the client is iterating on v2 features informed by real usage.
The bad outcome
Client cuts maintenance after launch to save money. By month nine, the new iOS update broke the keyboard handling. By month fifteen, three core libraries are end-of-life and unpatchable. By month twenty, the app is functional but feels obviously dated. By month twenty-four, the conversation has become "do we rebuild or kill it?" — and the answer is usually rebuild, which costs more than maintenance would have for the entire intervening period.
We've sat through this conversation often enough that it's now our standard advice on day one: commit to maintenance, or commit to a rebuild in two years. Pretending neither is required leads to the worst outcome.
What to ask your maintenance partner
Whether it's us or someone else, the questions worth asking:
- What's your monthly cadence? Weekly is usually overkill, quarterly is too slow, monthly is the right beat for most apps.
- What's your SLA for critical security issues?
- Do you provide proactive monitoring or only reactive support?
- What does the monthly report look like? Real data, or just hours logged?
- What happens when an OS update breaks something? Is fixing it included or extra?
The right partner can answer all of these without hesitation.
If your current app has been quietly drifting and you'd like an honest assessment of where it stands, book a discovery call. We'll tell you whether maintenance fixes it or whether you're past the point of return.
