Skip to content

Case study · AI + SaaS Products

A lean, fast platform with no lock-in

The agency expected more pages and more polish. The real question was what each page actually needed to do.

What it could have been
  • Decorated, page by page
  • Heavier than it needed to be
  • Dependent on the builder to run
Client
a marketing and web agency
Problem
A web platform heading toward more pages and more polish than the job needed.
Approach
Challenged every page like engineers, shipped it lean, and handed it over owned.

The problem

The agency came in expecting the usual path to a better platform: more pages, more polish, more to maintain. It is the default assumption, that a site gets better by getting bigger.

That path adds weight the business then has to carry, often with a dependency on whoever built it. Before adding anything, Stallwart asked what each page actually needed to do.

The reality behind the problem

The instinct to add pages feels like progress because the work is visible: another page to point at, another flourish to show. It is expensive for the same reason, because every page that ships is weight the business then carries, in load to serve, in maintenance to keep current, and often in a standing dependency on whoever built it.

Most of what the agency had planned was decoration rather than a job. More polish had become the default rather than a decision, and the default is paid for continuously, long after the page that carried it was signed off.

What was assumed

More pages, more polish.

What Stallwart asked

What does this page actually need to do?

The desired outcome

The goal was not a bigger site. It was a leaner one that still did everything the business actually needed, and that the agency could run on its own.

  • Every page matched to the job it has to do, with decoration that carries no job removed.
  • Weight taken out wherever a page did not need it to work.
  • Fast to load and simple to run under real traffic.
  • Owned outright, with no standing dependency on the builder.
  • A platform the agency can change and ship from without Stallwart.

The system we designed

Stallwart challenged every page like engineers, not decorators, keeping only what each one needed to do its job. What shipped is leaner and faster than the platform the client came in expecting to build, and the agency owns all of it, with nothing about it needing Stallwart to keep it running.

How it works

  1. 01

    Challenge every page

    Each page questioned on what it has to do, like engineers rather than decorators. Anything that was decoration without a job came out.

  2. 02

    Cut to what the job needs

    The platform built down to what the work actually requires, so it ships leaner and faster than it would have by default.

  3. 03

    Hand it over owned

    Shipped so the agency owns it outright, with nothing about it depending on Stallwart to keep it running.

The approach

Leaner than it would have been, and owned.

What it could have been

  • Decorated, page by page
  • Heavier than it needed to be
  • Dependent on the builder to run

What shipped

  • Every page challenged like engineers, not decorators
  • Leaner and faster than planned
  • Owned outright, nothing needs us to run it

Architecture

What sits underneath.

The platform was organized around what each page has to do, not around a page count.

01

Pages and content

Each page is built down to the job it has to do. A page with no job does not ship, so there is less markup, fewer assets, and less for a visitor to load.

02

Rendering and delivery

Lean pages mean the platform is fast to serve and cheap to run, without the weight a decorated site carries on every request.

03

Build and deploy

The site is produced from source the agency holds, so a change can be made, built, and shipped by the team without anyone outside it.

04

Hosting and operation

The platform runs on infrastructure the agency controls, with no proprietary tie that would need Stallwart to keep it running.

The hard parts

Where the engineering judgment was.

01

Engineer the pages, do not decorate them

A page earns its place by doing a job. Treated as decoration, pages add weight and polish the business then has to carry.

The trade-off

Fewer pages to point at, and some expected flourishes left out on purpose.

02

Lean by default

A platform built down to what the job needs is faster to run and simpler to keep reliable than a heavier one.

The trade-off

It looks smaller than the site the client first pictured, by design.

03

Full ownership, no lock-in

A platform the client owns outright can be run and changed without the builder. Dependency is a cost that never stops.

The trade-off

No ongoing tie to Stallwart to lean on, which is the point.

Reliability & guardrails

How it avoids the wrong call.

A leaner platform is a more reliable one: less to run, less to break, less to watch.

Fewer moving parts

Each page does one job and no more, so there is less that can go wrong and less to check when something does.

Fast under load

Lean pages stay fast when traffic rises, because there is less to serve per request than a heavier site would carry.

Predictable to run

A platform the team can hold in their heads fails in ways they can predict. Removing the pages that carried no job removed whole categories of problem with them.

Recoverable

The site can be rebuilt and redeployed from source the agency holds, so a bad change is reversed rather than lived with.

Productionization

What turns it from a demo into software.

What kept the platform dependable in daily operation, and kept it the agency's to run.

Repeatable deploys

Shipping a change is a repeatable build and deploy the team runs, not a manual process that only the builder understands.

Monitoring

The platform reports whether a deploy succeeded and whether pages are serving, so a problem is visible rather than discovered by a visitor.

Rollback

A deploy that goes wrong can be reverted to the last good version, so a mistake is a short interruption rather than an outage.

Kept lean over time

The platform was handed over with the discipline that produced it, so pages are added by the same test: what job does this one do?

What changed

What shipped was leaner and faster than the platform the agency came in expecting to build. The pages that carried no job never shipped, so there is less to load for a visitor and less to maintain for the team.

The agency owns all of it, with nothing about its operation depending on Stallwart. The platform can be run, changed, and shipped from by the people who hold it, which is what independence from the builder means in practice.

  • From more pages to pages that each do a job.
  • From decorated to engineered.
  • From heavier than the job needed to lean and fast.
  • From dependent on the builder to owned outright.

In practice

Founder, a marketing and web agency

The founder described Stallwart challenging every page like engineers rather than decorators, and shipping something leaner than he would have built, and faster, that he owns entirely, with nothing about it needing Stallwart to keep it running.

Ownership & handover

What you receive, and keep.

The platform is yours to run, with nothing held back.

  • Source code for the whole platform.
  • The build and deploy pipeline that ships it.
  • Documentation of how each page is structured and why.
  • Runbooks for deploying and rolling back.
  • Hosting and configuration under the agency's own accounts.

What we learned

The insight that generalizes: a platform gets better by getting smaller, not bigger. Challenging every page like engineers keeps only what earns its place, which is how the work ends up leaner, faster, and cheaper to run. Ownership is part of the same idea, because a platform with no standing dependency on its builder is one the business can carry on its own.

This work connects to

See it in action in
  • AI + SaaS Products
  • High-performance web
  • Full ownership

Frequently asked

What makes a web platform high-performance?

A platform built down to what each page actually needs to do, with decoration that carries no job removed. Leaner work is faster to run and simpler to keep reliable than a heavier site built by adding pages.

Why build leaner instead of adding pages?

More pages is more to build, more to run, and more to depend on someone else for. Challenging each page like engineers keeps only what earns its place, which is how the platform ends up leaner and faster.

What does owning the platform outright mean?

The agency can run and change the platform without Stallwart. There is no lock-in and no standing dependency on the builder, so nothing about keeping it running needs us.

Will the agency need Stallwart to keep it running?

No. It was shipped to be owned outright, with nothing about its operation depending on Stallwart. That independence was a goal of the build, not an afterthought.

Carrying a platform heavier than the job needs?

We challenge every page like engineers, ship it lean, and hand it over yours.

Last updated: October 9, 2026