Working with Multiple Workflows in Parallel¶

You'll often juggle several workflows at once — three brands, two workflows each, plus variants in progress. Here's how to do it without dropping context or losing track of what's where.

The mental model¶

Each workflow has state at any given moment:

A current version (V0-N or V1+)
A status (testing / approved / archived)
A fan-out state (which accounts have shipped)
Where it is in the pipeline (briefed / scripted / storyboarded / generated / picked / delivered)

When you're juggling 5 workflows, you're tracking 5 sets of this state simultaneously. The pipeline's tooling exists specifically to make this manageable.

Use the master tracker as your north star¶

The Google Sheets master tracker is your at-a-glance view across all workflows. Open it whenever you're switching contexts:

"What was I doing last?" — the tracker tells you
"Which workflow needs attention next?" — sort by last_updated and look for stale-but-not-finished ones
"What's approved vs. still testing?" — sort by status

Sync the tracker after every meaningful change (tracker-sync skill) so it stays accurate.

Avoid context-switching cost¶

Switching between workflows mid-task is expensive:

You have to reload the workflow's state into your head
You have to remember where you left off
You have to re-confirm the brief, audience, sales channel, etc.

Every switch costs 5-15 minutes of state reconstruction.

Strategies to minimize switching:

Finish to a clean handoff point before switching: "I'm done with XYZG3-V0-3 testing; the next step is fan-out, but that's a 30-minute task — I'll do it next session." That's a clean handoff. Switching now is fine. Stopping mid-iteration (Scene 3 partly fixed, Scene 4 still needs work) is messy — when you come back, you have to rediscover where you were.
Batch similar tasks across workflows: If you have 3 workflows in fan-out phase, do all the fan-outs in one sitting. Repeated process = faster execution + less context cost than alternating tasks.
Daily / weekly cadence per workflow: For high-volume teams: each workflow gets a specific time slot. Mondays = research, Tuesdays = scripting, Wednesdays = generation. Rigid but it works at scale.

Tracking state across workflows¶

For each workflow you're juggling, know:

State	How to track
Current version	Version registry → `current_version`
Status	Version registry → `status`
Fan-out state	Version registry → `fanned_out_to_accounts`
Where you are in the pipeline	Mental model + last entry in version registry history
Outstanding action	Probably a TODO in your project tracker (Asana / Notion / etc.)
Last meaningful change	Version registry → `last_updated`

When in doubt, the version registry is authoritative. The master tracker mirrors it (after sync). Your mental model is fuzzy and unreliable.

The "stale workflow" problem¶

Workflows that sit untouched for weeks often need to be re-evaluated:

The product file may have updated — your workflow may now have unapproved claims
The brand's banned-word list may have updated — your script may have new violations
The platform's content policies may have shifted — your workflow may be at risk
The account's strategy may have changed — your workflow may not fit anymore

Before resuming a stale workflow:

Re-read the brand file for any rule updates
Re-read the product file for any claim updates
Re-run the compliance audit against the script
Check the platform's recent policy changes

If anything has changed, fix the workflow (surgical edits or Lvl 1-2 variant) before continuing.

Avoiding cross-workflow contamination¶

Common mistake: applying knowledge from Workflow A to Workflow B without checking it applies:

Wrong: "the hook style that worked for XYZG3 will work for XYZG5": Different products, different audiences, different brands. The hook style might transfer; it might not. Test before assuming.
Right: "the hook style that worked for XYZG3 might work for XYZG5 — let me run a quick test variant before committing": Variant cheaply tests the assumption. If it works, great. If not, the data is cheap.

Workflows aren't templates that automatically generalize. Each one has its own audience and context.

The "what was I doing?" recovery protocol¶

When you come back to a workflow after a break:

Read the version registry's history array for that workflow. The most recent entry tells you what happened last.
Open the latest .nbflow in PatchWork to see the current state visually.
Look at the compliance audit doc (if any) for what was already approved.
Check the master tracker for any external notes (deadlines, client comments, etc.).

5 minutes of reading these = clean context restoration.

Naming conventions help juggling¶

Predictable file names make multi-workflow juggling easier:

projects/may/xyz-wellness/sales/testing/XYZS1-V0-3.nbflow
                                        ^      ^
                                        |      version
                                        workflow code (brand prefix + S/G + N)

A glance at the filename tells you:

Which brand
Sales or growth
Which workflow in sequence
Current version

Without consistent naming, you'd need to open every file to know what it is.

Communication discipline¶

If you're working with a team, the master tracker is half the communication — but not all of it. Conventions:

Update the version registry's history note with what you did + why
Tag the team in the project tracker when handing off
Note client commitments somewhere visible (Slack, project tracker)
Document brand-rule updates in the brand file with a date

Workflow juggling fails most often when assumptions about state diverge between team members. Explicit communication prevents that.

When you should NOT parallelize¶

Some workflows are bad candidates for parallel work:

First workflow ever for a new brand — full attention, don't divide
Critical client deadline — finish to delivery before starting the next
Compliance-sensitive workflow — too much risk of errors when switching contexts
Lvl 4 structural change — the structural rewiring is risky; concentrate on it

Use judgment. Not every workflow needs to be in flight simultaneously.

When you're ready¶

→ Next: Speed Levers — how to actually run things faster once you understand cost and limits.