I've written the memo to `fy2026_scenario_memo.md`. In the Downside, FY2026 operating income falls to a $6K loss, roughly breakeven.

| | Base | Upside | Downside |
|---|---:|---:|---:|
| Revenue | $11.20M | $11.94M | $10.44M |
| Operating income | $303K (2.7%) | $501K (4.2%) | $(6)K (–0.1%) |
| Change vs Base | — | +$197K | –$309K |

- **The downside is bigger than the upside:** the Downside loses $309K against Base, while the Upside gains $197K.
- **Losing the three credit-watch accounts is the largest single risk, at –$203K.** Each customer brings in about $67.6K of gross profit, and at Base economics the business breaks even at about 39.5 customers.
- **Smaller downside hits:** the March vendor cost increase that isn't passed through costs $78K, and the fuel surcharges cost $28K. Passing the vendor increase on to customers would bring the Downside back to about +$72K.
- **Upside drivers:** the two Enterprise wins add $135K and the 2% price increase adds $62K.

The CSV only lists the drivers, so I set up the P&L myself: revenue is customers × revenue per customer, gross profit applies the margin %, and operating income subtracts headcount cost and other opex. The memo states this assumption, along with these points:
- I didn't apply any extra month-by-month timing to the annual drivers.
- The file gives no probabilities, so the memo has no probability-weighted outcome.
- Costs are the same in every scenario, so the Downside doesn't include any cost cuts in response.