Scenario planning: best, base & worst case
Planning off a single "here's what will happen" forecast is how a good month becomes overconfidence and a bad month becomes panic. Model three cases on the few variables that actually move your business, and pre-decide what you'll do in each before emotion gets a vote.
Fastest path: one prompt, end to end
🤖 AI prompt — paste into ChatGPT / Claude
You are a DTC financial planner. Use MY numbers only; invent nothing.
Base-case monthly inputs: revenue [$], revenue growth [% MoM], blended CAC [$], COGS [% of revenue], fixed costs [$/mo], current cash [$].
Build a 3-case, 6-month model on just the variables that matter (revenue growth, CAC drift, COGS change):
1. BASE: my inputs as given.
2. BEST: growth +[X] pts, CAC flat, COGS flat.
3. WORST: growth -[X] pts, CAC +[Y]%, COGS +[Z] pts.
For each case show monthly revenue, contribution, net, and cash balance, and the month cash would run out (if ever).
Then propose ONE pre-committed tripwire per case (e.g. "if worst-case revenue holds 2 months running, cut [what]").
If I omit an input, ask; do not assume a value. If you cannot browse for a benchmark, use only my numbers.
Output: a 3-column table (best/base/worst) with cash runway + the tripwires.
Or do it in 4 steps
- Pick the 3-4 variables that actually swing the outcome. For most stores that's revenue growth rate, blended CAC drift, and COGS/gross-margin change. Ignore the twenty inputs that barely move the number; model the few that do.
- Build three cases, not twenty. Base is your honest expectation. Best is realistically good (growth up, costs flat). Worst is a plausible bad run (growth down, CAC up, COGS up). Resist doomsday and hockey-stick extremes; they're useless for decisions.
- Compute cash runway under each case. Revenue and profit matter, but cash is what kills a store. For each case, project the monthly cash balance and mark the month it hits zero. Worst-case runway is the number that should set your risk appetite.
- Pre-commit a tripwire per case. Decide the action now, while calm: "if worst-case revenue holds two months running, cut ad spend 40% and pause the next inventory order." Pre-committed rules beat in-the-moment panic every time.
Worked example (labeled): Base: 8% MoM growth, $33 CAC, 60% gross -> 5 months of runway. Best (12% growth, costs flat) -> cash grows, reinvest into inventory.
Worst (2% growth, CAC $41, COGS +4 pts) -> cash hits zero in month 4. Tripwire: two straight worst-case months triggers a 40% ad cut and a hiring freeze. Now a bad quarter is a plan, not a crisis.
Update the three cases whenever a key variable actually moves, and keep the tripwires visible. Use pnl-forecast-12mo as your base-case model, then branch best and worst off it.