Razoyo Logo

Harden or Rebuild Your Software? How to Tell Before You Pay

Harden or Rebuild · Inherited Code · Decided on Evidence

Harden or Rebuild? Six Questions That Decide It Before You Sign Anything

If the first thing a development firm proposes is a rewrite, be careful. A rewrite is the most expensive answer and the easiest to sell, and a firm that has not read your code cannot know it is the right one. The right answer comes from evidence. Sometimes the code is sound and needs hardening. Sometimes the foundation cannot carry the load and replacing it is cheaper than patching it for two years. This page shows how to tell which you have, and where we have said each.

Schedule Free Consultation Take the Keep-or-Replace Test
“

Will you just tell us to rebuild everything?

Can we harden it first and rebuild later?

What if an AI tool built it?

Do we own the answer if we do not hire you?

Razoyo Zoy assessing an inherited codebase
$2,500
Fixed-Price System Spec, Time-Boxed
7
Production-Readiness Dimensions Checked
3
Public Cases Where We Said Keep It

Is This the Right Page for You?

This is for you if: software already exists (built by an AI tool, a freelancer, an agency or a team that has left) and you need to decide what to do with it.

This is not for you if: nothing has been built yet. Start with scoping a software project instead.

Two different jobs

What Is the Difference Between Hardening and Rebuilding?

Hardening

Keeps the code and the data model and fixes what stops it running safely with real users: access control, secrets, tests, monitoring, a repeatable deploy. The product on day one is the product on day thirty, only safer.

Rebuilding

Replaces the foundation, usually because the design itself is what fails: a data model that cannot represent the business, or an architecture that hits a ceiling as volume grows. You keep the requirements, the workflows and the lessons. You do not keep the code.

How do you tell which one you need?

The Keep-or-Replace Test

Six questions, in order. Answer them about the code, not about how you feel about it.

Step 1

Does It Break at Scale, or Only Under Bugs?

Bugs get fixed. A ceiling is different: work that takes hours today and would take days at ten times the volume, because of how it is built. A ceiling points to rebuild.
Step 2

Can the Data Model Represent Your Business?

If your real rules (roles, tenants, states, exceptions) do not fit the tables, every feature fights the schema. That points to rebuild. If they fit and are only messy, that is hardening.
Step 3

Is There an Access Model at All?

No login, shared credentials, anyone with the link can see everything. This is the most common gap in AI-built prototypes and it is usually fixable in place.
Step 4

Could a Second Engineer Read It?

Structure a stranger can follow means you can extend it. Tangled, undocumented code with prototype scaffolding mixed into business logic raises the cost of every change.
Step 5

Are There Tests, and Is There a Deploy Path?

Absent is not disqualifying. Absent tests are added around code that works. What matters is whether the code is testable at all.
Step 6

Is What You Would Keep Worth More Than What You Would Replace?

Working workflows, the data and the intent behind them are worth keeping whatever you decide about the code.
Schedule Free Consultation

What Points to Hardening, and What Points to a Rebuild?

Signal Points to hardening Points to a rebuild
Failure mode Specific, findable bugs A ceiling that gets worse as volume grows
Data model Fits the business, needs cleanup Cannot express the real rules
Security Missing controls that can be added Trust boundaries in the wrong place
Structure Readable, some scaffolding to remove Business logic tangled through everything
Tests None yet, but the code is testable Cannot be tested without rewriting it
Cost to change Rises slowly Rises with every feature

Most systems are a mix. That is why the decision is made on an audit, not a guess.

Three public cases, none of which ended in a rewrite

When Have We Told a Client to Keep What They Had?

Women's POWER Pickleball League

A $5,000 fixed-price Spec produced a complete specification and an honest estimate that came back larger than she wanted to spend. She declined the build and took the document to a partner who could fund it. See the case study.

VivaTerra

The site was slow and the owner blamed the host. It was not the host. Too many database calls and a few extensions were the problem. We fixed those and left the platform alone. See the case study.

Hanger Project

A troubled Magento upgrade. We reorganized the product catalog and redesigned the pages rather than starting over. See the case study.

And the other answer

When Have We Said Rebuild?

Spider, the platform behind Lighthouse

The first version, built by an outside team, was failing almost daily and could not onboard new properties. We established first whether to extend or replace it. The answer was replace. It now serves more than 10,000 people a day, and we still run it. See the case study.

Matrix

An internal hotel sales tool, not sellable as a subscription. We rebuilt it as a multi-tenant product. It now runs hundreds of properties. See the case study.
A fixed-price System Spec, yours to keep

Decide on evidence

What Does It Cost to Find Out?

A System Spec is $2,500 fixed and time-boxed. It looks at what you actually have, checks it against seven production-readiness dimensions, and gives you a written harden-or-rebuild call with a senior-engineer estimate. The document is yours whether or not you hire us. Details on how the process works and on technical due diligence.

System Spec

Start here
$2,500 fixed · yours to keep

Time-boxed

We look at what you actually have and give you a written harden-or-rebuild call.

  • Checked against seven production-readiness dimensions
  • A written harden-or-rebuild call
  • A senior-engineer estimate
  • Yours whether or not you hire us
Book a System Spec

Free Consultation

Free 30 minutes · no obligation

Bring what you have. You will leave knowing whether it is worth a System Spec, before you spend a dollar.

  • A senior person, not a sales rep
  • A straight answer on fit
  • A scoped next step and what it costs
Schedule Free Consultation

Questions We Get Before the Call

No. The call comes from the audit. Recommending the cheaper option when it is the right one is part of how we work, and we have done it.

A System Spec is time-boxed and usually measured in days to a few weeks, depending on the size of the system.

Sometimes. If the foundation can carry the load for a year, hardening first buys time and revenue. If the foundation is the problem, hardening first only postpones the cost.

The test is the same. Lovable, Replit, Bolt.new, v0, Cursor and Bubble apps are usually hardened, because the workflow and data model are worth keeping. The gaps are typically access control, secrets, tests and a deploy path.

Yes. The Spec is yours outright and works with any team.

Software rescue, vibe-coded app to production and technical due diligence. Built with an AI tool? Start with the seven signs your AI-built app has outgrown DIY. Comparing firms rather than answers? Rewrite vs refactor an inherited codebase and who takes over software someone else built.

Pick your next step

Start With a Conversation. It Costs Nothing.

Every engagement begins the same way, with a free consultation. You will leave it with an honest recommendation, including the recommendation not to build.

1 Free

Free consultation

A working session with a senior person, not a sales rep. Always the first step.

Schedule Free Consultation
2 $2,500

Fixed-price System Spec

If it is worth doing, the Spec finds the answer and returns a real build estimate. Fixed at $2,500.

Start a System Spec
3 No email required

Read the proof first

Not ready to talk? Read the full Lighthouse case study.

Read the Lighthouse Case Study