How Revision Rounds Work in Custom Figure and Toy Projects

Revision rounds are where custom programmes quietly gain or lose weeks. The engineering work involved in a sculpt adjustment is rarely enormous; what makes revision expensive is the cycle time — each round costs a render, a review, a decision and a re-issue, and rounds run sequentially rather than in parallel.

Buyers who understand the mechanism get more from fewer rounds, negotiate cost allocation from a clear position, and avoid the most common trap of all: approving a stage while leaving an open question that reopens later.

What counts as one round

The definition sounds pedantic and is in fact commercially decisive, because contracts are written in rounds. Getting the definition agreed before the first review avoids the entire class of dispute that arises when a factory bills for a second round and the buyer believed it was part of the first.

Why the definition is worth arguing over in advance

A round is the unit in which engineering effort is billed and in which schedule is measured. If the definition is vague, both sides will apply their own, and the disagreement will surface precisely when the buyer is under pressure to approve and the factory is under pressure to protect its margin. Settling it at kickoff costs a sentence.

Where the definition goes wrong in practice

Disputes about revision counts almost always trace to one of three ambiguities. First, whether comments sent in separate messages constitute separate rounds or one round. Second, whether a correction caused by factory error consumes a buyer round. Third, whether a change requested before a stage is approved counts as a revision or as part of the original work. Each of these is easily resolved at kickoff and painfully resolved mid-programme.

The per-review-cycle definition is the one to aim for. It aligns the incentives on both sides: you are motivated to gather comments thoroughly, and the factory is motivated to address them completely.

The arithmetic that makes rounds expensive

Rounds are sequential, so the cost of review is not the number of comments but the number of cycles. Each cycle carries fixed overheads regardless of how much changes: a revision of the model, a fresh render, an internal review, a communication round-trip, and a decision. Five rounds of one comment each cost far more than one round of five comments.

Where rounds occur by stage

Different stages generate different kinds of feedback, and it helps to know in advance which ones are likely to iterate.

Sculpt and deco are the two stages that iterate most, for different reasons: sculpt because proportion and likeness are genuinely hard to specify in words, and deco because colour perception varies and matching is interpretive until there is a physical reference.

Why sculpt iterations are legitimate rather than wasteful

There is a tendency to treat sculpt iterations as evidence of a poor brief. Sometimes they are. But proportion and character are genuinely difficult to specify precisely in words or even in drawings — a two-millimetre change in head width can alter whether a figure reads as intended, and that judgement is easier to make with a rendered or physical model in front of you than on a drawing. Two to three rounds at the sculpt stage is normal, healthy iteration, not failure.

Why deco iterations are different in character

Decoration iterations are usually about reference rather than judgement. Once a physical colour standard exists and the placement is drawn, deco becomes comparatively objective — the sample either matches or it does not. Deco rounds that run long almost always trace back to a missing colour standard or an unplaced artwork, both of which are brief problems rather than deco problems.

Typical round counts

The figures below are practical planning norms rather than guarantees. They assume a reasonably complete input pack; a thin pack multiplies them.

What drives a project above the typical range

Projects that exceed their planned rounds usually share one of a small number of characteristics. Recognising them early allows the round budget to be adjusted before the project is under pressure.

Two observations from these numbers. First, the sculpt and deco allowances are the ones worth negotiating upward if your design requires high fidelity to a reference. Second, rounds spent at the pre-production stage are the most expensive to recover from, because tooling is already cut and the schedule is committed.

How to plan around these numbers

Treat the figures as a budget rather than a target. If you know your design requires close fidelity — a licensed character, a portrait likeness, a specific mechanical arrangement — negotiate more rounds at the stage concerned rather than discovering the shortfall mid-project. Running out of rounds at the sculpt stage is a commercial problem with a schedule consequence.

What triggers a round

Not every piece of feedback requires a structured round. Knowing the trigger points keeps the process proportionate.

The most common trigger is the buyer's own process

It is worth being candid about where most extra rounds come from. In practice the largest single source is not factory error or engineering difficulty — it is feedback that arrives in fragments, or stakeholders who were not consulted before the first review and object afterwards. Both are entirely within the buyer's control.

• A change to the brief — anything that alters what was agreed at kickoff turns the iteration into new work.
• A dimension change — resizing a feature usually propagates into structure and tooling.
• A proportion adjustment — small on a render sheet, significant in sculpt and mould.
• A colour or material substitution — affects cost and compliance, not only appearance.
• A deco reposition or rescale — touches artwork, print tooling and registration.
• A compliance discovery — a restricted substance or test failure that forces redesign.
• An unexplained factory deviation — the sample does not match the approved drawing.

Distinguishing a revision from new work

The commercial distinction between a revision and new work is whether the change was within the agreed scope. A revision refines something already agreed; new work introduces something that was not. This distinction matters because revisions consume the round allowance while new work is separately chargeable — and because the two are frequently confused when a brief was never clear in the first place.

Who pays for which change

This is the table to have in front of you during a revision discussion. The principle is simple: the party whose decision created the rework carries the cost.

Recording the allocation decision

Whichever way an allocation is decided, record it. An allocation agreed verbally and later disputed is effectively undecided, and the second conversation will be harder than the first. A short written note stating the change, the decision, the responsible party and the effect on the round count is sufficient, and it protects both sides.

The last row is worth pausing on. If your specification allowed a range and the part falls inside that range, there is nothing to correct — the fix belongs in the specification for next time, not in a claim against this run.

The batching discipline

Batching is the highest-leverage habit in the whole revision process. Its arithmetic is unforgiving: rounds are sequential, so the number of rounds determines the calendar, not the number of comments.

Why batching is the single highest-leverage habit

Almost everything else in revision management is administrative. Batching is the one habit that changes the arithmetic of the programme, because it collapses the number of sequential cycles rather than the amount of work inside each one. A buyer who batches consistently will complete a programme in a fraction of the calendar time of a buyer who does not, with the same designed product at the end.

Templates that make batching easy

Batching is easier when it has a fixed structure. A simple comment table, used for every review, removes the effort of deciding how to present feedback and makes the factory's response straightforward to track.

Practical batching rules: collect comments in one place rather than in a chat thread; number them so they can be answered point by point; group them by part or by view; and mark each comment as must-fix or nice-to-have so the factory can prioritise rather than treat everything as critical.

• One channel — a single document, not scattered emails and messages.
• Numbered points — each comment referenced so responses can be tracked.
• Prioritised — separate must-fix from optional, and say so explicitly.
• Annotated visually — mark the render itself, not only describe the issue in text.
• Internally agreed first — consolidate your own team's views before sending, so conflicting comments do not arrive together.

Version control and timestamps

A revision history is only useful if the current version is unambiguous. Adopt a simple, strict convention and apply it from the first file.

The failure mode batching prevents

Without batching, comments arrive one at a time and the factory addresses each in isolation. The result is a design that has been revised eleven times against eleven separate instructions rather than once against a coherent position — and it is common for a later comment to contradict an earlier one, producing rework of rework. Batching forces the buyer to resolve internal contradictions before the factory spends effort on them.

What a good change log records

A change log is a short, factual record. It does not need to explain reasoning in depth; it needs to make the sequence of decisions traceable so that any feature can be traced back to the stage that introduced it.

The most expensive version-control failure is not a lost file — it is two parties working from different files. Confirm the version in writing with each exchange, and never accept a revised deliverable that does not state which version it supersedes.

Running a review that produces decisions

A review that ends without a decision consumes a round without advancing the programme. Structure the review so that it terminates in either approval or a consolidated revision list.

• Review against the brief — check the deliverable against what was agreed, not against evolving taste.
• Use a fixed checklist — the same checklist each time, so nothing is examined inconsistently.
• Decide the disposition — approve, approve with minor corrections, or revise.
• Name an owner for each comment — so responses cannot stall waiting on the wrong person.
• Set a return date — an unowned deadline is not a deadline.
• Record the outcome — one line per stage in the approval log.

Making version control effortless

Version control becomes easy when the convention is fixed once and never varied. The temptation to name a file "final", "final2" or "latest" is precisely what destroys traceability. Sequential integers with a date and a change log are dull, unambiguous, and sufficient.

Setting a review turnaround you can hold

Review turnaround is the most controllable part of the schedule and the one most often allowed to slip. Committing to a turnaround — and holding it — is worth more than any negotiation about factory speed, because the factory cannot proceed without your decision and every day of delay is a day added to the total programme.

Closing a stage so it stays closed

A stage is only genuinely closed when three things are true: the deliverable is approved, the approval is recorded, and every open question has been resolved or explicitly deferred with a written note. Anything less leaves a thread that can pull the stage back open.

• Written approval — in the same channel as the deliverable, with a date.
• Explicit confirmation — that the factory has accepted the approval and recorded the version.
• Open items resolved — or deferred in writing, with the stage at which they will be revisited.
• Basis recorded — what was approved, so a later comparison is possible.
• No side agreements — nothing agreed verbally that is not in the written record.

What reopens a stage, and why it is expensive

Why a review without a decision is not a review

A review that produces impressions rather than a disposition has consumed engineering time and calendar time without advancing the programme, and it will have to be repeated. Insisting on one of four dispositions — approve, approve with corrections, revise, or hold — converts a discussion into a gate.

The three conditions for a genuinely closed stage

Closing a stage is not the same as finishing a review. It requires the deliverable to be approved, the approval to be documented, and every open question to be either resolved or formally deferred. Skipping the third condition is what quietly reopens stages weeks later, when the deferred question finally has to be answered by someone who assumed it had been settled.

Recording an approval properly

An approval is only useful if it identifies precisely what was approved. "Approved" attached to an email thread containing several attachments is not a record; it is a source of ambiguity. State the version number and the date alongside the approval, in the same channel as the deliverable.

Making deferrals work

Deferring a question is legitimate when the information needed is not yet available. It becomes harmful when the deferral is unstructured, because the question resurfaces at a random moment. A structured deferral names the question, the stage at which it will be answered, and the person responsible.

The pattern is that reopening is rarely caused by a technical discovery. It is caused by a decision deferred to a later stage than it belonged. Each stage exists to settle one class of question, and moving a question forward moves its cost with it.

Negotiating revision terms

Revision terms are easier to agree at the outset than to argue about mid-project, when both sides have already invested.

Why reopening is usually a decision problem

It is tempting to attribute reopening to late technical discoveries. In practice the cause is more often a decision that was pushed to a later stage than it belonged. Each stage has a class of question it exists to settle; carrying one forward does not remove it, it simply moves its cost to a more expensive stage and a busier moment.

Handling a round overrun mid-project

When the round allowance is exhausted, the productive response is not to argue about the count but to establish two things: what is causing the additional rounds, and what it will cost to finish. If the cause is a factory quality issue, the additional rounds are the factory's responsibility and the allowance is irrelevant. If the cause is a buyer-side change, the additional rounds are chargeable and the sooner that is acknowledged, the sooner the work can proceed.

The rate for extra rounds is worth setting in advance even if you hope never to use it. Disputes about revision cost are usually disputes about count rather than rate, and having the rate settled lets the conversation focus on the substantive question of what caused the extra round.

How revision behaviour differs between a first order and a repeat

A repeat order starts with a tool, an approved specification and a golden sample — which removes most of the stages that generate rounds in the first place. Revisions on a repeat are usually limited to colour batch matching and any deliberate change to the product, and the round count should reflect that. Applying a first-order revision budget to a repeat is unnecessary; applying a repeat budget to a first order is a source of friction.

The hidden cost of a late-arriving stakeholder

One of the most expensive patterns in revision management is a stakeholder who joins after stage one. Their feedback arrives late, often disagrees with decisions already approved, and forces a return to a stage that had been closed. The remedy is organisational rather than contractual: assemble the complete review panel before the first stage begins, and confirm in writing who has authority to approve.

What good feedback looks like

Feedback quality determines round efficiency. Vague feedback produces a round that does not converge, because the factory has to interpret the comment before acting on it.

Revision discipline checklist

Handled well, revision rounds are the mechanism by which a design becomes manufacturable — each one narrows the gap between intent and reality. Handled badly, they are the mechanism by which a fixed programme becomes an open-ended one. The difference is almost entirely about how the feedback is batched and how the stages are closed.

Building revision discipline into the contract, not the conversation

Revision management works best when the rules are agreed before they are needed. Once a round is in dispute, both sides have invested and both sides have a position to defend; a rule agreed at kickoff, when neither party has anything at stake in a specific case, is far easier to apply than a principle negotiated under pressure.

The rate for additional rounds is worth fixing even if you intend never to pay it. A dispute about revision cost is almost always a dispute about the number of rounds, and having the rate settled removes the incentive to argue about count for financial reasons. What remains is the substantive question — what caused the extra round — which is the question that actually improves the programme.

Finally, it is worth noting that revision discipline is largely a buyer-side capability. Factories revise when asked; the number of rounds is determined by how thoroughly the buyer prepares, how completely the buyer batches, and how decisively the buyer closes. Those are all within your control, and they are the levers that turn a revision process from an open-ended cost into a predictable stage of the programme.