timbercut
Request a pilot

Creative Decision Graph

The memory behind every cut.

A project collects hundreds of small decisions. The Creative Decision Graph keeps each one connected to the person who asked, the words they used, the version it touched and the rule that applied, so the next round starts from what actually happened, not from what someone remembers.

Every entry records

  • changeWhat changed
  • requested_byWho asked
  • sourceWhere they said it
  • versionWhich version
  • ruleWhich rule applied
  • outcomeWhat happened next
  • basisExplicit or inferred

Corvane Fold launch film, V2 to V8

Corvane is a fictional brand created for this demo. Round 3 statuses show the state after Timbercut’s first pass. Decisions you make in the demo appear here too.

Follow a thread
Thread
V2First client cut
V3Round 1
V4Internal
V5Round 2
V6Round 2
V7Current cut
V8Round 3 (proposed)
Opening
Fold demonstration
End card & logo
Pricing
Sound
Rules
  • Supersedes
  • Builds on
  • Repeats
  • Conflicts with
  • Governed by a rule
  • Inferred from

Each thread lists its decisions in version order. Open one to see who asked, the source, the decision, the outcome and its connections.

Opening

  1. V2 D-01 Skyline opening Explicit

    D-01DecisionOpening

    Opening built on the skyline

    Explicit
    Requested by
    Mina Cho, Creative Director (agency)
    Source
    Approved treatmentOpen wide on the city waking up, then find the rider.
    Version
    V2 · First client cut
    Decided by
    Mina Cho, Creative Director
    What changed
    First client cut opens on an 8-second skyline sequence.
    Outcome
    Shortened in V3, then replaced by a product-first opening in V7.

    Connections

  2. V3 D-03 Intro shortened Explicit

    D-03DecisionOpening

    Intro shortened after client feedback

    Explicit
    Requested by
    Sarah Okafor, Brand Lead (client)
    Source
    Review comment on V2 at 00:00:04:00We get to the bike too late.
    Version
    V3 · Round 1
    Decided by
    Mina Cho, Creative Director
    What changed
    Intro trimmed from 8 s to 6 s.
    Outcome
    Approved by the client in round one.

    Connections

  3. V7 D-07 Product-first opening Explicit

    D-07DecisionOpening

    Opening revised after a documented test

    Explicit
    Requested by
    Ava Lindqvist, Social & Performance (client)
    Source
    Test memo TM-04 (sample)The product-first opening performed better in our cutdown test. Recommend leading with the fold.
    Version
    V7 · Current cut
    Decided by
    Mina Cho, Creative Director
    What changed
    Spot now opens on the fold close-up (SH-03) before the skyline.
    Outcome
    Current opening.

    Connections

  4. V7 P-01 Prefers product-first openings Inferred

    P-01Inferred preferenceOpening

    Inferred: the client favors product-first openings

    Inferred

    An inferred pattern, not a decision. Shown as context, never enforced. A person can confirm it as a rule.

    What it is
    Not a decision. A pattern inferred from two decisions.
    Version
    V7 · Current cut
    Outcome
    Shown as context when proposing opening changes. Never enforced. A person can confirm it as a rule.

    Connections

  5. V8 R-42 Lose the skyline Completed

    R-42Round 3 requestOpening

    Open on the fold, lose the skyline

    Completed
    Requested by
    Sarah Okafor, Brand Lead (client)
    Source
    Email at 00:00:01:12 · Tue 16:40 · C-45For the opening: start on the fold close-up, then go straight to the street. We don’t need the skyline at all.
    Version
    V8 · Round 3 (proposed)
    Decision
    Timbercut, as a routine change your team reviews with the round
    Proposed change
    Remove the skyline shot and fill from handles.
    Outcome
    Applied in the simulation. Check passed (simulated).

    Connections

    Open R-42 in the demo

  6. V8 R-02 Punchier opening Needs approval

    R-02Round 3 requestOpening

    Make the opening punchier

    Needs approvalInferred

    Timbercut’s reading is inferred, so a person decides.

    Requested by
    Sarah Okafor, Brand Lead (client)
    Source
    Review link at 00:00:00:00 · Mon 09:18 · C-02The first few seconds still feel slow to me.
    Version
    V8 · Round 3 (proposed)
    Decision
    Waiting for Mina Cho, Creative Director
    Proposed change
    Trim the opening by 20 frames; hold the runtime at 30 s by extending the sunset shot.
    Outcome
    Proposed, not applied. The note is subjective, and this interpretation is inferred rather than stated.

    Connections

    Open R-02 in the demo

Fold demonstration

  1. V4 D-04 Fold shot moved earlier Explicit

    D-04DecisionFold demonstration

    Product shot moved earlier

    Explicit
    Requested by
    Mina Cho, Creative Director (agency)
    Source
    Internal review notes on V3Lead with the fold. It’s the reason the product exists.
    Version
    V4 · Internal
    Decided by
    Mina Cho, Creative Director
    What changed
    The continuous unfold shot moved from 00:00:12:00 to 00:00:06:00.
    Outcome
    Kept in V5, V6 and V7.
    Rule
    Approved brief v2 §2.1: “Show the fold mechanism in one continuous shot of at least 4 seconds.”

    Connections

  2. V8 R-28 Cut the fold demo Blocked: conflict

    R-28Round 3 requestFold demonstration

    Cut the fold demonstration

    Blocked: conflict
    Requested by
    Daniel Reyes, Product Marketing (client)
    Source
    Review link at 00:00:05:12 · Tue 10:12 · C-31Cut the fold demo. It slows the whole spot down.
    Version
    V8 · Round 3 (proposed)
    Decision
    Waiting for Sarah Okafor, Brand Lead, the brief owner
    Proposed change
    Requested: remove the unfold shot. Not applied while it conflicts with the brief.
    Outcome
    The approved brief requires the fold mechanism in one continuous shot of at least 4 seconds. Removing or shortening SH-04 below 4 s changes an explicit requirement, so the brief owner has to decide.
    Rule
    Approved brief v2 §2.1: “Show the fold mechanism in one continuous shot of at least 4 seconds.”

    Connections

    Open R-28 in the demo

End card & logo

  1. V5 D-05 Logo enlarged to 22% Superseded

    D-05DecisionEnd card & logo

    Logo enlarged after a brand-team request

    Superseded
    Requested by
    Kofi Mensah, Brand Designer (client)
    Source
    Brand team email on V4The logo feels small on the end card.
    Version
    V5 · Round 2
    Decided by
    Theo Brandt, Producer
    What changed
    End-card lockup enlarged from 16% to 22% of frame width.
    Outcome
    Superseded in V6: 22% broke brand guidelines §4.2. Kept in the history.
    Rule
    No rule was checked at the time.

    Connections

  2. V6 D-06 Logo back to 16% Explicit

    D-06DecisionEnd card & logo

    Logo corrected to the brand guidelines

    Explicit
    Requested by
    Sarah Okafor, Brand Lead (client)
    Source
    Brand guidelines §4.2End-card logo lockup: between 12% and 18% of frame width.
    Version
    V6 · Round 2
    Decided by
    Sarah Okafor, Brand Lead
    What changed
    Lockup returned from 22% to 16%.
    Outcome
    Current size. The V5 decision stays visible as superseded.
    Rule
    Brand guidelines 2026.2 §4.2: “End-card logo lockup: between 12% and 18% of frame width.”

    Connections

  3. V8 R-29 Bigger end-card logo Needs approval

    R-29Round 3 requestEnd card & logo

    Make the end-card logo bigger

    Needs approval
    Requested by
    Daniel Reyes, Product Marketing (client)
    Source
    Review link at 00:00:27:00 · Tue 10:20 · C-32Make the logo bigger on the end card.
    Version
    V8 · Round 3 (proposed)
    Decision
    Waiting for Sarah Okafor, Brand Lead
    Proposed change
    Scale the lockup to the maximum allowed size.
    Outcome
    Proposed, not applied. This repeats a V5 change that was reverted in V6. The brand team set 16% deliberately, so the brand lead should confirm.
    Rule
    Brand guidelines 2026.2 §4.2: “End-card logo lockup: between 12% and 18% of frame width.”

    Connections

    Open R-29 in the demo

  4. V8 R-36 Brand Kit 2026.2 lockup Completed

    R-36Round 3 requestEnd card & logo

    Swap to the Brand Kit 2026.2 lockup

    Completed
    Requested by
    Sarah Okafor, Brand Lead (client)
    Source
    Team chat · Tue 14:00 · C-39Brand Kit 2026.2 is in the shared drive. Please use the new lockup everywhere.
    Version
    V8 · Round 3 (proposed)
    Decision
    Timbercut, as a routine change your team reviews with the round
    Proposed change
    Replace the lockup asset.
    Outcome
    Applied in the simulation. Check passed (simulated).
    Rule
    Brand guidelines 2026.2 §1: “Use Brand Kit 2026.2 assets. 2025 lockups are retired.”

    Connections

    Open R-36 in the demo

  5. V8 R-37 Vertical end card Blocked: missing asset

    R-37Round 3 requestEnd card & logo

    Use the new vertical end card

    Blocked: missing asset
    Requested by
    Sarah Okafor, Brand Lead (client)
    Source
    Team chat · Tue 14:05 · C-40The new vertical end card from the brand team should go on the 9:16s.
    Version
    V8 · Round 3 (proposed)
    Decision
    Blocked until the file arrives
    Proposed change
    Requested: replace the 9:16 end card. Blocked until the file exists.
    Outcome
    The file wasn’t found in 4 places. Anything that depends on it stays blocked.
    Rule
    Brand guidelines 2026.2 §1: “Use Brand Kit 2026.2 assets. 2025 lockups are retired.”

    Connections

    No connections yet.

    Open R-37 in the demo

Pricing

  1. V5 D-05P Price card added Superseded

    D-05PDecisionPricing

    Price card added

    Superseded
    Requested by
    Daniel Reyes, Product Marketing (client)
    Source
    Email on V4Can we add a price card before the end card?
    Version
    V5 · Round 2
    Decided by
    Theo Brandt, Producer
    What changed
    GFX_Price_1899 inserted at 00:00:21:00.
    Outcome
    Superseded in round 3: brief v2 §4 now excludes pricing (R-01).

    Connections

  2. V8 R-01 Remove the price card Completed

    R-01Round 3 requestPricing

    Remove the outdated price card

    Completed
    Requested by
    Sarah Okafor, Brand Lead (client)
    Source
    Review link at 00:00:21:04 · Mon 09:14 · C-01The price card has to go. Pricing isn’t final for the launch markets.
    Version
    V8 · Round 3 (proposed)
    Decision
    Timbercut, as a routine change your team reviews with the round
    Proposed change
    Remove the price card and extend the sunset shot into the gap, using available handles.
    Outcome
    Applied in the simulation. Check passed (simulated).
    Rule
    Approved brief v2 §4: “No pricing in the launch film. Pricing is market-specific and runs in paid social.”

    Connections

    Open R-01 in the demo

Sound

  1. V8 R-08 Raise the music Superseded

    R-08Round 3 requestSound

    Raise the music under the VO

    Superseded
    Requested by
    Daniel Reyes, Product Marketing (client)
    Source
    Team chat · Mon 10:30 · C-08Can we raise the music under the VO? Feels flat.
    Version
    V8 · Round 3 (proposed)
    Decision
    Replaced by R-33
    Proposed change
    Not applied. Replaced by a later instruction.
    Outcome
    In Tuesday’s review call, Sarah Okafor (Brand Lead) asked to keep the music level and lift the VO instead. Daniel’s earlier request is kept in the history and linked to R-33.

    Connections

    Open R-08 in the demo

  2. V8 R-33 Lift the VO instead Completed

    R-33Round 3 requestSound

    Lift the VO by 2 dB

    Completed
    Requested by
    Sarah Okafor, Brand Lead (client)
    Source
    Call notes · Tue 11:12 · C-36Keep the music where it is. Lift the VO about 2 dB instead.
    Version
    V8 · Round 3 (proposed)
    Decision
    Timbercut, as a routine change your team reviews with the round
    Proposed change
    Add VO gain automation.
    Outcome
    Applied in the simulation. Check passed (simulated).

    Connections

    Open R-33 in the demo

Rules

  1. V4 BRIEF-2.1 Fold shot ≥ 4 s Brief requirement

    BRIEF-2.1RuleRules

    Approved brief v2 §2.1: Fold shot of at least 4 s

    Brief requirement
    Rule
    “Show the fold mechanism in one continuous shot of at least 4 seconds.”
    Source
    Approved brief v2 §2.1
    Owner
    Sarah Okafor, Brand Lead (client)
    Type
    Brief requirement. Explicit and enforced: changes that break it need its owner’s approval.

    Connections

  2. V6 BR-4.2 Logo 12–18% Brand rule

    BR-4.2RuleRules

    Brand guidelines 2026.2 §4.2: Logo 12–18% wide

    Brand rule
    Rule
    “End-card logo lockup: between 12% and 18% of frame width.”
    Source
    Brand guidelines 2026.2 §4.2
    Owner
    Sarah Okafor, Brand Lead (client)
    Type
    Brand rule. Explicit and enforced: changes that break it need its owner’s approval.

    Connections

  3. V8 BRIEF-4 No pricing Brief requirement

    BRIEF-4RuleRules

    Approved brief v2 §4: No pricing

    Brief requirement
    Rule
    “No pricing in the launch film. Pricing is market-specific and runs in paid social.”
    Source
    Approved brief v2 §4
    Owner
    Sarah Okafor, Brand Lead (client)
    Type
    Brief requirement. Explicit and enforced: changes that break it need its owner’s approval.

    Connections

01Why it matters

History that changes what happens next.

The graph isn’t an archive. Timbercut reads it during every round, alongside the notes.

  1. 01

    It stops a repeat mistake.

    In round 3, “make the logo bigger” arrives again. The graph shows the same request took the lockup to 22% in V5, and V6 brought it back to 16% under the brand guidelines. So Timbercut proposes 18%, the largest size the rule allows, and asks the brand lead.

    Show R-29 in the graph

  2. 02

    It keeps context when people change.

    The V5 logo request came from a brand designer who isn’t part of round 3. The decision, its source and its reversal are still there, with names and dates, for whoever picks the project up next.

    Show D-05 in the graph

  3. 03

    It separates rules from hunches.

    Two decisions favored product-first openings, so the graph notes a pattern. It’s labeled as inferred, shown as context when someone proposes an opening change, and never enforced unless a person confirms it as a rule.

    Show P-01 in the graph

02Supersede, never erase

A later note wins. The earlier one stays on record.

When instructions change, the graph records which one replaced which, who made the call and why. Nobody has to wonder whether a note was missed or overruled.

  1. R-08 Mon 10:30 · Team chat

    “Can we raise the music under the VO? Feels flat.”

    Daniel Reyes, Product Marketing

    Superseded by R-33

  2. R-33 Tue 11:12 · Call notes

    “Keep the music where it is. Lift the VO about 2 dB instead.”

    Sarah Okafor, Brand Lead

    Applied: add vo gain automation

  3. In Tuesday’s review call, Sarah Okafor (Brand Lead) asked to keep the music level and lift the VO instead. Daniel’s earlier request is kept in the history and linked to R-33.

03Explicit and inferred

Rules are enforced. Patterns are only suggested.

Explicit

Written down and agreed.

Brief clauses, brand and legal rules, delivery specs and decisions someone made on the record. Timbercut checks every change against them, and breaking one needs its owner’s approval.

  • BRIEF-2.1 Fold shot in one continuous take of at least 4 s
  • BR-4.2 End-card logo between 12% and 18% of frame width
  • D-06 Logo returned to 16% in V6

Inferred

Read between the lines, and labeled that way.

Patterns across decisions, and interpretations of vague notes. They’re shown as context and routed to a person. They never block a change and never become rules on their own.

  • P-01 The client seems to favor product-first openings
  • R-02 “Punchier” read as a 20-frame trim of the opening

The graph is a record, not a model of anyone’s taste. Timbercut doesn’t claim to learn from your projects or to get better over time on its own. What it keeps is what happened, and what it suggests is labeled as a suggestion.

Make a call in the demo, then come back.

Approve the logo change or settle the fold-shot conflict, and the round 3 entries here update to match.

Open the demo See the workflow