Should I pause or cut my ads? (small business)

Top picks

Protocol

1. 1

   Paste facts (5–8 min).

   Assemble only what you actually know. Mark gaps [MISSING] / [NEED FACT]. Do not invent rates, BATNA, legal outcomes, or benchmarks. This small-business budget go/no-go uses pasted performance—not creative testing. Adjacent cash context: weekly-cash-position-brief. No invented industry benchmarks or “good ROAS” targets.

   Paths at a glance

   Path	Lean this way when (from your paste)
   A — Pause weak channels	Named channels miss your CAC/ROAS floor; winners can continue.
   B — Cut budget X% keep winners	Cash tight but some campaigns clear your floor—you authorize the % cut.
   C — Reallocate to sales ops	Paid underperforms vs sales activity you can fund with the same dollars per your notes.
   D — Hold spend	Metrics meet your floors; cash supports current burn; no cut needed now.

2. 2

   Lock decision-tree rules (step 2).

   Paste the rules first so the model cannot mint industry benchmarks, expected ROAS, or a cut larger than you authorize.

   Copy-paste prompt

   Copy

   Help me run a PAUSE-OR-CUT-ADS budget decision tree.
   Hard rules:
   - Use ONLY spend, CAC/ROAS, cash, and channel facts I paste. Do not invent industry benchmarks, expected ROAS, or competitor CPCs.
   - Paths: Path A Pause weak channels / Path B Cut budget X% keep winners / Path C Reallocate to sales ops / Path D Hold spend.
   - If Path B, use only the % I authorize.
   - Not financial advice. I decide.
   Confirm in one sentence, then wait for my pack.

3. 3

   Fill pack (step 3a) and generate (step 3b).

   Paste the pack, then send the generate command. The model must not pick a path for you.

   Paste pack

   Copy

   CHANNELS / CAMPAIGNS IN SCOPE: [ ]
   SPEND BY CHANNEL (period): [ ]
   CAC / CPA / ROAS I MEASURE (definitions + numbers): [or MISSING]
   MY ACCEPTABLE CAC OR ROAS FLOOR: [or NEED FACT]
   CASH / RUNWAY NOTES (facts): [or link weekly cash brief facts]
   PIPELINE DEPENDENCE ON ADS (facts): [or MISSING]
   WINNERS VS WEAK (named from my data): [ ]
   % CUT I AUTHORIZE IF ANY: [or NONE]
   SALES-OPS USES FOR REALLOCATED $: [or NONE]

   After you send the filled pack, copy the Generate command in the next box and paste it into the SAME chat so the AI runs the decision tree / produces the output.

   Generate command — copy and send this next (same chat)

   Copy

   Build the decision tree now:
   1) Facts vs [MISSING]
   2) Scorecard table: Criterion | My fact | Lean (+/0/− or NEED FACT) | Note
   3) Path A Pause weak channels / Path B Cut budget X% keep winners / Path C Reallocate to sales ops / Path D Hold spend — for each: when it fits, what must be true from my pack, 3 owner-verify bullets
   4) Do NOT pick for me. Do not invent facts.

4. 4

   Choose a path (step 4); optional wording.

   Only after you pick a labeled path. Owner-verify: facts still match your pack; [MISSING] gaps filled or accepted; any sendable wording uses only terms you authorize.

   Path wording

   Copy

   I chose Path [A/B/C/D]. Draft an internal action checklist for that path only (what to pause/cut/reallocate, review date).
   No invented benchmarks. Mark [NEED FACT].

5. 5

   Second-model check (step 5).

   In the other tool, paste the tree plus your pack:

   Second-model check

   Copy

   Here is a pause/cut-ads tree from another assistant plus my pack.
   Find: (1) invented ROAS/CAC benchmarks, (2) cuts larger than I authorized, (3) channels paused I didn't name as weak, (4) financial advice tone.
   Return only corrections.

6. 6

   Human decide and act (step 6).

   Pick one path. Fill [NEED FACT] gaps from real sources. Send or implement yourself—AI drafts are not approvals, contracts, classifications, or tax elections. Cash context lives in the weekly cash-position brief.

When not to

• Do not invent platform benchmarks or “good ROAS.”
• Brand/trademark or regulated ad categories—follow those rules separately.
• Skip if you lack spend and conversion facts—export reports first.
• Not a creative testing recipe.