Time Management

How to Track Design Revision Rounds Without Mixing R1 Into R3

Freelance designer revisions are where fixed-price projects go to die — not because R3 is evil, but because R1, R2, and R3 usually live on one blob called “Design.” When the invoice (or the post-mortem) arrives, you cannot tell a paid round from a free one.

You do not need a new personality to do this. You need a client → project → task shape that matches how design work actually ships, and a habit of closing the door when a round is done.

Why one “Design” task lies to you

A brand or product engagement is not one activity. Discovery, wireframes, UI, a prototype, three revision rounds, and asset export are different economic objects. Mixing them produces three failures you only notice at invoice time:

• Scope looks like craft. R3 feedback that reopens IA still posts as “making it prettier.”
• Estimates cannot learn. You quoted 12 hours for visual. The 9 hours of extra rounds never leave the same bucket, so next quote is the same 12.
• Clients cannot see the story. Non-technical stakeholders do not read Figma comments. They read line items. “Design — 41 hours” is an argument. “Revisions R2 — 6 hours” is a conversation.

Tracking revisions in general is useful; we have written about why feedback loops belong on a timesheet at all. This post is narrower: how to keep rounds from collapsing into each other.

A phase list you can copy this week

Use the client’s project as the folder. Put the work on tasks named like invoices, not like your private process:

• Discovery / audit
• UX / wireframes
• UI / visual
• Prototyping
• Revisions R1
• Revisions R2
• Revisions R3
• Asset export / handoff
• Meetings (yes, its own task)

If a round has obvious sub-pieces — homepage, interior templates, component cleanup — those can be subtasks under that round. Keep the round as the thing you lock. Do not create a subtask named “tweaks” that outlives every round; that is the blob again, one level down.

The lock is the habit, not the software

A round ends when the client accepts it — email, Slack, a recorded “looks good,” whatever you already use. That is the moment you lock or close the task so further tracking cannot land there. The next hour of “tiny tweaks” has to choose a new round or a new paid change request. That choice is the whole system.

If you skip the lock, you will still start the timer on the last thing you had open. Designers do this because the file is the same file. The economic object is not the file. The economic object is the round you sold.

Write the acceptance into the task description: date, who said yes, link to the version. Briefs, screenshots, and Figma links belong in that rich text so a future-you (or a future invoice dispute) is not reconstructing intent from memory. You are not “integrating” a design tool. You are leaving a trail a non-designer can read.

What to do when R3 wants to reopen R1

That is not a revision round. That is a new discovery spike, or a change request. Keep R3 locked to “changes against the accepted R2.” If the client wants a different information architecture, start a task named like the truth: “IA change after R2 accepted.” Then you can decide — in English, with hours — whether it is covered, billed, or traded against something else in the remaining package.

For website work especially, set a checkpoint before the round that usually explodes. If you sold a fixed-price marketing site, you want a warning when time or budget is half gone — not a surprise at 110%. A project- or task-level tracker (an estimate and a budget you watch, not the running timer) is for that risk. The round split tells you where the overrun lived. The tracker tells you when to talk.

Untracked Slack “can you look?” messages are still time. If you will not put them on a round, put them on Meetings or a Support task. Do not pretend a glance is free because it was six minutes. Six-minute rounding exists as a conversation you can have later; the hours still have to exist as entries on the right leaf.

How to show this to a client who does not speak Figma

Invoice line items should look like the phase list, not like ticket keys and not like a Kanban dump. “Revisions R2 — 4.2 hours” plus a one-line note (“two homepage options, no IA change”) is enough. You can still attach the file in email. The timesheet’s job is the money story.

If you bill a day rate or a package, you still want the split internally. Packages go wrong in silence. The round data is how you notice you sold three rounds and delivered five — which is the same lesson as why most work-management boards fail designers: the board moved cards. It did not protect the round you quoted.

Where a time tracker fits (after the method)

The method works in a spreadsheet. Spreadsheets fail because locking a phase is a social habit with no teeth. MinuteShark is built around client → project → task → subtask, with time on the task. You can lock or close a task so further tracking stops. Subtasks are available on Standard if a round needs leaf work. Project and task descriptions are rich text — briefs, screenshots, links — so the acceptance note lives next to the hours.

When a site or brand package has a number attached, a tracker watches one project or one task against an estimate and a budget. That is not the running timer. It is the “we are halfway through the money” email you wish you had sent in week two.

Reports are What × Who × When. Generate the week, pivot by task, then generate an invoice from the same report and a cost profile. If you round (off by default; original duration kept), the time rounding guide is the rulebook — not something you invent in the PDF. Line items can follow the phase names you already tracked.

MinuteShark AI can add the R2 task or log forgotten time in plain English; you still confirm the card. It will not silently merge R3 into R1. That merge is a human shortcut. The lock is how you refuse it.