sourcing protocol

Scanner Network

A short standfirst that frames the artifact — one or two lines that tell the reader what the document is and why it matters.

format

Sourcing + QA network, decision log

scope

National capture · ~8 yrs

ownership

Built & owned end-to-end

01 what i built

When I started, every artist's work shipped out to a small set of dedicated scanners, and shipping physical work around the country was expensive. The unit economics were underwater before a file was ever captured — the cost scaled on the wrong variable, distance rather than value.

The problem was simple to name: we couldn't keep shipping every piece to a handful of scanners. My background in the arts meant I saw early that capture quality was underserved and would only matter more as the collection grew. The fix was to distribute — source capture close to where the artists were, mostly city-specific and sometimes regional, so the work didn't have to cross the country. That handled the logistics, and then growth changed the problem. We were selling more, which meant recruiting more artists and bringing on more artwork — and at that volume, scanner-to-scanner variation that low volume had hidden started to show. It mattered because a curator building a multi-artist collection can't have one piece sharp and the next soft. That second problem came out of solving the first and growing past it; catching it early kept the logistics fix from becoming the whole design. What it needed next was a capability layer.

02 the artifact

NETWORK

Digital Capture

sourced by artist geography · qualified on output

Two problems, in order

PROBLEM 1

Logistics

Reduce the overhead to get each piece captured, without shipping everything across the country.

SOLVED BY sourcing capture close to the artist rather than routing it to a handful of central scanners.

solving 1 and growing past it produced 2 ↓

PROBLEM 2

Output quality

Once the per-capture investment was worth making, fidelity had to hold from one artist to the next.

SOLVED BY qualifying scanners on the output they delivered, not proximity alone.

SOURCING METHOD — HOW A SCANNER GOT GREENLIT

01 · INTERVIEW

Equipment and process. Vendors describe the same work differently depending on background — fine-art reproduction versus commercial graphics — so the conversation adapts to get an honest read on what the equipment can actually hold.

02 · FOLLOW THE GEAR

Track where museum-copy scanners were being sold; the shops buying them are the ones working at the level the collection needed.

03 · TEST

Trial captures on the pieces that actually break a capture — varied sizes, heavy texture — not the easy work.

04 · CONTRACT

Formal agreements with the vendors that cleared the test.

SETTING AND HOLDING THE STANDARD

There was no capture standard to inherit.

Work came in as lossy JPEGs — fine for a thumbnail, useless for reprinting at scale — with no colour space, colour temperature, resolution floor, or shared definition of an acceptable file. I built one: TIFF capture, colour space and temperature fixed at the scanner so they weren’t corrected in post, and minimums that reproductions had to clear.

It ran on feedback in both directions. Toward the vendors, every capture was a data point — what held up, what didn’t, which scanners delivered and which to move the work toward as we could. Toward the artists, the same reads became coaching: this piece captured beautifully, this one fought the scanner, and — in broad strokes to the whole community — these materials and surfaces reproduce well, these don’t.

The documented standard tightened and expanded as the collection grew, keeping pace with the program rather than running ahead of it. Where written specs didn’t land, I built visual explainers — showing a vendor why a piece bound for large-format product needs far more capture data than its size suggests, rather than asserting it.

PER-CAPTURE ECONOMICS

Demand ran toward large artwork and large reproduction sets, much of it above what the existing scans could hold. Raising the standard for every accepted capture made sense on the margin: a few strong pieces in a batch cover the capture cost of the rest, so the investment held up across the collection even though plenty of individual captures never paid for themselves. That let the program plan capture volume against a consistent baseline instead of judging it scan by scan. The limit is diminishing returns — a piece at a given size holds only so much resolution, and past that, more investment buys nothing.

GROWTH-STRATEGY CALCULUS

Capture cost factored into where the program could grow. A new state with no qualified scanner in range roughly tripled the cost of onboarding artists there, so capture feasibility fed the call on whether a market was worth entering.

NETWORK

Digital Capture

sourced by artist geography · output qualified

Two problems, in order

PROBLEM 1

Logistics

Reduce the overhead to get each piece captured, without shipping everything across the country.

SOLVED BY sourcing capture close to the artist rather than routing it to a handful of central scanners.

solving 1 and growing past it produced 2 ↓

PROBLEM 2

Output quality

Once the per-capture investment was worth making, fidelity had to hold from one artist to the next.

SOLVED BY qualifying scanners on the output they delivered, not proximity alone.

SOURCING METHOD — HOW A SCANNER GOT GREENLIT

01 · INTERVIEW

02 · FOLLOW THE GEAR

Track where museum-copy scanners were being sold; the shops buying them are the ones working at the level the collection needed.

03 · TEST

Trial captures on the pieces that actually break a capture — varied sizes, heavy texture — not the easy work.

04 · CONTRACT

Formal agreements with the vendors that cleared the test.

SETTING AND HOLDING THE STANDARD

There was no capture standard to inherit.

There was no capture standard to inherit. Work came in as lossy JPEGs — fine for a thumbnail, useless for reprinting at scale — with no colour space, colour temperature, or resolution floor. I built one: TIFF capture, colour fixed at the scanner rather than corrected in post, and minimums every reproduction had to clear.

It ran on feedback in both directions. Every capture was a data point on the vendor — what held, what didn’t, where to move the work next. The same reads became artist coaching: this piece captured beautifully, this one fought the scanner, these materials reproduce well and these don’t. Where written specs didn’t land, I built visual explainers. The standard tightened as the collection grew, keeping pace with the program rather than running ahead of it.

PER-CAPTURE ECONOMICS

Demand ran toward large artwork and large reproduction sets, above what the existing scans could hold. Raising the floor on every capture penciled out across a batch — a few strong pieces cover the cost of the rest — so the program could plan capture volume against a baseline instead of judging it scan by scan. The limit is resolution: past what a piece of a given size can hold, more investment buys nothing.

GROWTH-STRATEGY CALCULUS

Capture cost shaped where the program could grow. A new state with no qualified scanner in range roughly tripled the cost of onboarding artists there, so capture feasibility fed the call on whether a market was worth entering.

↳ the standard is the asset — everything else is how it got sourced and held

03 What it proved

15,000+

assets managed across the collection's life, multiple formatted versions per piece included (up from ~3,500 at the start)

~20–30

scanners sourced by artist geography — qualified on output quality, not proximity alone

~$50 → $150

per capture — investment scaled to the large-format margin it unlocked

The capture standard was the asset. It defined what the collection could reproduce and sell, and set the bar every scanner and every artist worked to. As the business grew, the standard grew with it.

← Previous artifact

Commission Intake → Delivery SOP

← Read artifact

Next artifact →

Submission Workflow

Read artifact →