Meridian
notifications
SO

The next two quarters, today.

Friday · 13 June 2026 · forecast rebuilt nightly on closed-book actuals through 31 May.
Cash at quarter-end · 30 Sep · expected (P50)
2.58M €
2.04M3.05M  P10–P90 · ▲ 7% vs plan · base scenario
SepNovJanMarMayJun·nowJulSep
Actual Expected (P50) P10–P90 Plan
J
Jeeves · what’s moving the forecast
The base case still clears plan — but the margin is now one stalled deal wide.

Two-thirds of next quarter’s revenue is already committed or recurring, so the floor is solid. The wobble is all in pipeline: the Cobalt licensing renewal (€240K) has sat 9 days and it carries ~€90K of the P50. Slip it past Sep and the “chance of hitting plan” drops from 72% to about 58% — I’ve modelled both. Separately, collection lag crept from 41 to 48 days in May, which is what pulled the cash cone down a notch even though revenue held.

What builds the number

revenue · next 90 days · P50
function decomposition 2.71M €
1.18M
0.52
0.94M
·
Committed — signed SOWs & contracted retainers 1.18M
Recurring — renewals (churn-modelled) 0.52M
Pipeline — probability-weighted 0.94M
Seasonality — Sep ramp vs summer dip +0.07M

The committed and recurring blocks (€1.70M) are the forecast’s floor — they move only if a client cancels. Everything above that line is pipeline, and pipeline is where the band widens.

On the forecast’s radar

drift · renewals
trending_down
Licensing pacing below trend
78% of run-rate at mid-month · −€42K vs model
Drift
schedule
Collection lag widening
41 → 48 days · pulls the cash cone, not revenue
Watch
event_repeat
Halcyon retainer up for renewal
€120K/yr · 88% modelled renewal probability
Renewal
event_repeat
Northwind retainer up for renewal
€25K/mo · 94% modelled renewal probability
Renewal

Has it been right?

Full backtest →
MAPE · 6 mo
4.8%
Inside the band
5 / 6
Bias
−0.9%
Worst miss
Feb · 11%
A forecast you can’t check is a horoscope. Every projection is scored against what actually happened once the books close.

Same numbers your ERP already has. We just ran them forward — and, more importantly, showed you the spread, so “we’ll probably be fine” becomes a number with a confidence attached.