← Back to work
Case study

Visual standards at scale

Creating a repeatable visual system that kept a large and constantly changing collection of content coherent across different sizes, formats and proportions.

The situation

The business managed a large, constantly changing collection of visual content. The subjects varied widely in shape, size, proportion, colour and detail, new assets arrived all the time, and existing images were reused across pages, campaigns and customer touchpoints. While the collection was small, individual judgement kept it consistent. As it grew, that stopped scaling.

What was missing

The standard lived in people's judgement, so every new asset needed the same subjective decisions made again: how large the subject should appear, where it should sit in the frame, how much space it needed around it, whether to show it at literal scale or adjust it for balance, and when an exception was justified. Different people made reasonable calls and still produced work that didn't sit together.

It showed most when assets appeared side by side. Two items of similar real-world size could look dramatically different on screen, and unusual shapes could dominate a listing or disappear in it. Each image was fine on its own; the set wasn't, and customers found it harder to compare options.

My role

I identified that this couldn't be fixed one image at a time. I reviewed the different asset types, the contexts they appeared in and the inconsistencies that showed up when they were placed together, looking at both the physical characteristics of the subjects and their visual effect on screen. Then I defined the visual logic, tested it across formats, and turned it into standards other people could apply without relying on undocumented design judgement.

What got written down

The decisions that had been made asset by asset, captured once as standards and translated into specifications for creation, editing and quality checks.

The visual standards

  • Every asset is prepared within the same overall canvas, so images combine cleanly in grids, listings and campaign layouts.
  • Where physical size matters to understanding, relative scale is preserved. Where literal scaling would make an asset unusably small or overpowering, an agreed adjustment applies.
  • Subjects sit in the frame by defined centre points, margins and space around their shape.
  • Backgrounds, shadows, edges and other treatments are standardised.
  • Exceptions are allowed only in defined situations, and applied in a defined way.

Visual system

Subjective image decisions, turned into a scalable set of rules

A shared framework for presenting a large, constantly changing collection of visual content consistently

Individual judgement

Every asset decided in isolation

  • Scale, positioning and cropping vary from one image to the next
  • Similar-sized subjects can appear dramatically different on screen
  • Inconsistency becomes visible wherever assets sit together
  • Every new asset requires a fresh, undocumented decision

Shared visual framework

One system, applied consistently

  • A consistent presentation canvas across the collection
  • Defined scale relationships, not one arbitrary percentage
  • Standardised positioning, spacing and background treatment
  • Documented exceptions, so flexibility doesn't become inconsistency

The framework

Shared canvas

Every asset prepared within the same overall format and boundaries

Defined scale relationships

Consistency without a false impression of uniformity

Positioning & spacing

Centre points, margins and space defined for every shape

Background & treatment

Standardised so production differences don't distract from the subject

Exceptions with rules

Flexible where it's justified, without becoming inconsistent

Production guidance

Repeatable specifications for creation, editing and QA

What changed

The collection became more coherent across pages, campaigns and journeys. Customers could compare options more easily because differences in presentation no longer competed with differences between the subjects, and grids felt calmer and more deliberate. Production got easier too: new content could be made consistently, reviewed against clear criteria and added without gradually weakening the whole.

The same problem shows up wherever visual content is produced at volume without a shared system, including AI-generated imagery: Batch blindness is the same shape of failure, with a model in place of a photographer. Image management system tackles the same territory from the availability side.

Decisions that had been made asset by asset could now be settled by the standard.

---
title: Visual standards at scale
slug: visual-standards-at-scale
updated: 2026-10-03
summary: Creating a repeatable visual system that kept a large and constantly changing
  collection of content coherent across different sizes, formats and proportions.
---

If a problem like this sounds familiar, the audit is where I'd start: one workflow, one 90-minute session, and a written brief on what's missing and what to fix first.

Book the audit

Prefer to talk first? Get in touch — no obligation.