Maaz Zindani

Loading the index…

↑↓ move↵ openesc close4072 pages

Buyer guide

Should I rebuild my old app from scratch or keep fixing it?

In short

Rewriting is rarely the first answer. Fix the app when it still earns money and the problems are specific. Consider a rewrite when the technology is unsupported, nobody can change it safely or each change costs more than the last. A phased replacement, one part at a time, often carries less risk than a full rebuild.

  • Written for founders and business owners
  • Updated
  • By Maaz Zindani, Senior Software Engineer

Who is this guide for?

Owners of an app or web system that is years old, slow to change or breaking often, who have been told by one developer to rebuild it and want an independent way to judge that advice.

Why this is hard to judge from outside

A developer who finds old code hard to read will often say the whole thing should be rewritten. That can be true, and it can also be the natural reaction to code someone else wrote. The advice alone is not evidence.

The useful question is what is actually wrong. A slow page, a missing feature and an unsupported server are three different problems with three different fixes.

Signs that point to fixing it

The app still does its job, customers still use it and the problems you can name are specific: one slow screen, one failing integration, one outdated library. Specific problems can be priced and fixed one at a time.

Another good sign is that the data is clean and understood. The data is usually the most valuable part of an old system, and keeping it intact removes the riskiest step of a rewrite.

Signs that point to replacing it

The UK government's guidance on managing legacy technology lists common triggers: maintaining the old technology costs more than replacing it, it blocks changes the business needs, supplier support has ended, or the risk of keeping it has become too great.

The same guidance names risks that make old systems harder to move: hidden dependencies, unclear data, poor documentation and too few people who understand the system. If these describe your app, a plain repair may keep failing.

A middle path: replace it in pieces

The UK Government Service Manual describes the strangler pattern. You place the interface you want around the old system, then replace what is behind it step by step without disturbing the new parts. It also warns that wrappers add complexity over time, so they must be removed once the move is finished.

For a business app this usually means keeping the old system running while one feature at a time moves to new code. You see progress early and you can stop if priorities change.

What to ask before you decide

Ask for a short written assessment before any rebuild is quoted: what is wrong, what it costs you now, what each option would involve and what could go wrong. A rewrite needs a written scope that lists every behaviour the old app has, including the ones nobody remembers.

Ask also what happens to your users and data on the day of the switch, and whether the old system can stay available as a fallback.

How I would approach it

I would read the code and the hosting setup first, list the specific problems, and give you the options in writing with the risks of each. The price is quoted after a short discovery call, against a written scope.

How do the options compare?

OptionSuitsWatch for
Fix specific problemsThe app works and the faults are specificRepairs can pile up if the base is unsupported
Replace in phasesA system that must keep running while it changesTemporary wrapper code that is never removed
Full rewriteUnsupported technology or a system nobody can change safelyMissing behaviours, data migration and a long period with no visible result
Leave it aloneStable, low-risk, rarely changedSupport ending on a server or library you depend on

What should I check before I sign?

  • You can name the specific problems, not just that the app is old.
  • You have a written assessment from someone who read the code.
  • The options include fixing and phased replacement, not only a rebuild.
  • The scope lists the behaviours the new version must keep.
  • Your data and its meaning are understood before any move.
  • You know how users are switched over and how to go back.
  • You have a second opinion if one developer is the only source of advice.
  • Backups and source code are in accounts you control.

How pricing works

I quote in writing after a short discovery call, against a written scope. Building an app from scratch? The app cost calculator gives a first range in two minutes.

What else do people ask?

Is a rewrite always more expensive than fixing?

Not always, but a rewrite has more unknowns, because you must rebuild every behaviour the old app has. Compare written options and ask what is most likely to go wrong in each.

Can I keep using the app while it is replaced?

Often yes, if it is replaced in phases. The old system keeps running while parts move to new code, and the official guidance describes this approach.

How do I know if the old technology is a real risk?

Check whether the supplier or maintainers still support the language, framework and server it runs on. Ended support is one of the common triggers for replacing a system.

Should I get a second opinion?

Yes, when the advice is to rebuild everything. A short independent review costs far less than the decision it informs.

Sources

How was this page made?

I drafted this page with AI assistance from the official documents listed under Sources. Found a mistake? Tell me.

How can we work this way together?

Next step

Want a written quote to compare?

A discovery call within 48 hours, then a scope and a price in writing.