Skip to main content
GullySystem

From MVP to Production Hardening

Production hardening is the work done after an MVP proves the idea, so the product can carry more users with less risk. Reviews, tests and written records follow.

Why hardening waits for proof

An MVP answers one question fast. The shortcuts are deliberate, such as hard-coded settings, thin error handling and one server doing everything, and they were sensible while the idea was still unproven. Customers then depend on it. Hardening repays the shortcuts in order of danger, not in order of how untidy the code looks to a developer reading it cold.

What gets added

Security review

The dangerous ones go first. Logins, permissions, stored secrets, file uploads and third-party packages are checked against known weaknesses, and each finding is ranked by how much harm it could do.

Performance and capacity

Timings replace hunches here. Heavy database queries and slow pages are measured under realistic load, and fixes follow those measurements instead of guesses about which screen feels slow to the founder.

Backups and error tracking

Backups gain a restore drill. Errors are grouped and sent to the people who can fix them, so a customer need not report each fault before anyone notices it.

Documentation and tests

Setup steps and decisions are written down. Automated tests then cover the paths that earn money, so a later change cannot quietly break checkout without somebody being told straight away.

Gaps in an MVP

A prototype-grade product assumes that every user behaves well. Real traffic does not.

Account safety

Password reset, session expiry and limits on repeated login attempts are missing or thin. Attackers and careless users both find these gaps early.

Failing hard

When a payment or email provider is down, an MVP shows a blank error. Hardened code falls back to a clear message, queues the work and retries.

Hidden manual steps

A founder quietly fixing records by hand each night is a risk. These tasks are found, either automated or written into a runbook someone else can follow.

One person knows everything

If the original developer is the only one who understands deployment, a holiday becomes an outage. Documentation and a second pair of hands close that gap.

How a hardening pass is run

  • Start with a readiness review that lists every concern and its owner
  • Rank concerns by harm and likelihood, and fix the top of the list first
  • Record anything left unfixed as a known risk with a reason
  • Ship fixes in small releases, each tried on a staging copy
  • Repeat the review after the fixes to confirm the list shrank

What hardening cannot fix

Hardening improves the product you have. A weak design stays weak. If the data model is wrong, or the structure cannot grow, rebuilding that part may cost less than patching it again and again, and a review will say so. Skipping hardening carries the opposite risk. That risk surfaces on the busiest trading day of the year, when a slow page or a failed payment costs the most.

A production hardening pass, week by week

Hardening is a sequence of small reviews rather than one big event. Three people take a turn: the developer, the founder and sometimes an outside reviewer. Here is how the weeks look. The product has already proved the idea with real users.

Developer reviews sign-in

The developer reads the code for sign-in, password reset and sessions. Weak points are listed with a reason. Each one is fixed or accepted by the founder, with the decision written beside it. Silence is not approval.

Founder reads the incident plan

The founder reads a one-page plan saying who is called when the product stops. The plan has names and phone numbers of the day, not role titles alone. Gaps are filled before sign-off. Numbers are tested by calling.

Reviewer tries odd inputs

A reviewer types strange values into every form and tries pages they should not reach. What breaks gets a ticket. Issues touching money or personal data get fixed first.

Test run on a copy

The team restores last night’s data to a spare copy of the system and opens it. If it works, the date goes into the log. If not, the fix comes before anything else. This step is never skipped.

Make an informed decision

Agree the evidence before signing off this part.

Use a representative scenario to decide what this part of the system must show and where review is still needed.

Security review

Can an ordinary account reach a record outside its allowed scope?

For “Security review”, agree two accounts with different permissions and test direct requests. Record the expected refusal and include exported or downloaded data in the assessment.

Background reading: OWASP: authorisation checks

Backups and error tracking

Has recovery been timed using an application the team can actually use?

Use the “Backups and error tracking” review to name the business operation that must resume. Compare a witnessed restore with the agreed downtime and data-loss limits.

Background reading: AWS: recovery objectives and tests

FAQ

Frequently asked questions

When is the right moment to harden an MVP?

Harden once users have shown repeat use or paid, and you plan to grow, because before that point polish is effort spent on an unproven idea. After it, shortcuts cost real customers.

Should we rewrite the MVP instead?

No, because a rewrite repeats old work and adds new bugs, and your customers keep waiting for the features they were promised. Rebuild only the broken parts.

Can hardening happen while customers keep using the product?

Yes, in small releases that are tried on a staging copy first, with a prepared way back if a change goes wrong in front of users. Database moves wait for quiet hours.

What do we receive at the end?

You receive a written review of findings, the fixes made, a list of what remains and why, and documentation your next developer can follow without asking anyone. Everything stays in your name.

What is a production readiness review?

It is a structured check of whether a service can be run safely by the people responsible for it. The idea comes from site reliability practice. It asks about monitoring, recovery, capacity and ownership, and unanswered questions are recorded as risks.

Is a security review the same as a penetration test?

No. A review checks the code, settings and design against known weaknesses. A penetration test actively tries to break in. Start with the review and add a test later when the stakes rise.

How long does hardening take?

Duration follows the product’s size and the number of shortcuts taken. The review comes first and gives a ranked list, so you can see the effort before agreeing to fix everything.

Talk to us

Tell us what you need.

Tell us what your team does today, where work gets stuck and what you want to change.

What happens after you share your requirements

  1. We review your brief and the systems you already use.
  2. We clarify the workflow, missing information and the work to consider first.
  3. The proposed scope sets out the deliverables, responsibilities and dependencies.
  4. Review the scope and commercial terms before deciding whether to proceed.

Start with your requirements

Describe your current process, the people involved and the change you need. Start with the details you know; we can clarify the rest together.

Share your requirements