Every board meeting gets a new curve. Nobody puts the old one next
to it. A forecast that is never scored is a mood. Klarschicht GmbH
is a fictional company — demo company, not a client engagement —
and this is what scoring it looks like, month by month, against the
books.
The curve that moved a little to the right
In January the board sees an EBIT curve for the next two years. It
climbs, gently, the way curves on board slides do. In March the
board sees the curve again. It still climbs. It looks, if anything,
exactly like January's — perhaps nudged a little to the right.
Everyone reads that the same way: on plan. It looked like this last
time.
It did not. Put a pin in one date — say 31 December — and read both
curves at that point. January's curve and March's curve are two
different promises about the same day. The gap between them is the
most informative number in the meeting, and it is on no slide,
because each slide only ever shows the latest curve.
At Klarschicht, the January forecast had fiscal year 2026 closing
with EBIT of +€475k. The March forecast said −€225k. The September
forecast says −€666k. Nobody lied. Each curve was honest on the day
it was drawn. The year just moved by €1.1M, and the slides were
drawn so that you could not see it.
Weather, not prophecy
Nobody expects a weather service to know on Monday whether it rains
on Saturday. Monday's forecast for Saturday is wide. Friday's should
be sharp. That is the deal, and it gives you two things to judge:
does the forecast tighten as the day comes closer, and was it honest
about how wide it was while it was still far away.
Meteorology has scored its own forecasts for decades. Every forecast
is kept, and every forecast is held against what actually happened.
Company forecasting mostly does not. The old curve is overwritten by
the new one, and with it the only evidence of whether the
forecasters know what they are doing.
A Tracker Report keeps the old forecasts. Every month, right after
the books close, the company is forecast again — the full range of
outcomes, not one line — and that forecast is left standing. Then
the books arrive and stand next to it. It is not a one-off opinion.
It is a running record, for the one to three numbers this particular
business is steered by: cash, revenue, EBIT, EBITDA.
That record answers three questions.
One date, forecast nine times
The first question is the pinned date. Fiscal year 2026 has not
closed yet, but it has been forecast nine times, once after every
monthly close.
Exhibit 1The pinned date
FY 2026 EBIT, as forecast month by month
One column per monthly close, January to September. Dot: median
forecast. Bracket: P20–P80 range. The year has not closed.
Klarschicht GmbH · fictional demo company · 1,000 runs per forecastbrenwick
Read it as a story about the team, not the company. The median went
+€475k, +€85k, −€225k, −€635k, −€465k, −€630k, −€638k, −€663k,
−€666k. It took three closes to accept that a profitable year was a
loss-making year. April overcorrected; May walked half of it back.
Since June the number has barely moved, and the range around it has
shrunk from €1.3M wide in January to €140k now.
How fast a forecast moves, and how far, says more about the people
than about the business. A team that finds €1.1M of bad news within
five closes and then holds its number is a team that knows where it
stands. A team whose year-end drifts by a few percent every single
month, always in the same direction, is telling you something too —
it just never tells you all at once. Whoever signs the forecast owns
its trail.
September, as seen from January
The second question is the one meteorology asks: now that the day
has come, how did the forecasts approach it?
Exhibit 2Landed
Cash at 30 September 2026
Forecasts made after the January to August closes. Grey bar: the
actual. Number: share of that forecast's simulated outcomes below
the actual — 50% is dead centre, near 0% or 100% is a clear miss.
Klarschicht GmbH · fictional demo company · 1,000 runs per forecastbrenwick
September has closed. Cash on 30 September was €5.76M. The January
forecast had said €6.60M, and the actual sits at its 2% mark — fewer
than one in fifty of January's simulated futures came in that low.
That is not a wide forecast being unlucky. That is a forecast being
too sure of itself.
Then the forecasts close in, but not in a straight line. February
put the actual at its 24% mark, March at 12%. April swung to 92% —
suddenly too gloomy. May swung back to 9%, too sunny again. From
June on the actual sits in the middle of every range: 57%, 42%, 37%.
That is the shape you want to end in. But the April–May whipsaw is
the useful part, and it gives a board a precise question to ask:
what did you learn in April that you unlearned in May?
Without the old forecasts on the page, that question cannot even be
phrased.
Yesterday said sunshine
The third question is the strictest. Two years out, a forecast is
allowed to be vague. One month out, it is not. The forecast made
after the August close should know September almost exactly — the
people, the contracts and the pipeline are all already there.
Picture a weather app that said sunshine for today, yesterday
evening, and it is raining. Once is weather. After the third time
you delete the app. But you can only judge it if you hold
yesterday's forecast next to today's sky, line by line. Almost
nobody does.
Exhibit 3One month out
September 2026, line by line
€ · September actuals against the forecast made after the August
close. An excerpt of the report's full statement.
At Klarschicht, September looks perfect at the top. Revenue was
€1,433,697 against a forecast of €1,435,000 — off by €1,303. Closing
cash landed inside its range.
One level down, it is less tidy. Enterprise contracts came in at
€1,004,712, above the top of their range: two deals closed where the
forecast expected none. The AI add-on — the line the plan's growth
rests on — came in at €95,385 against a forecast of €118,000, below
the bottom of its range. Other operating expenses ran €19k over the
forecast and above their range. The total landed because two misses
cancelled each other.
One month like that is noise. Three months like that, in the same
lines, is a pattern — and a board that sees it in month three is
having a different conversation from a board that sees it in month
nine.
Forecast, forensics, tracker
Steering a company takes three instruments, and each answers a
different question.
The Forecast Report says where the company is
likely heading: not one line, but the range of futures and how
likely each part of it is. The Forensic Report says
why — which assumption carries the outcome, and where the plan
breaks. The Tracker Report says whether the last
forecasts were right, and how the company got from there to here.
A forecast without forensics is a number you cannot defend. A
forecast without a tracker is a number you cannot check. Together
they turn a monthly board slide into something a board can steer by
— and hold someone accountable to.
Ask for the old curve
At the next board meeting there will be a new curve, and it will
look a lot like the last one, a little to the right.
Ask for the last one next to it. Ask for the one date where they
should agree, and how far apart they are. And ask for last month's
forecast of this month, line by line.
If nobody has them, that is your answer.
Source
Klarschicht specimen Tracker Report, books closed 30 September 2026.
Each monthly forecast is 1,000 simulated futures, started from the
books as they stood right after that month's close. Ranges are
P20–P80. Demo company. Not a client engagement.
SPECIMEN · Demo company
The report for your portfolio company
Brenwick runs the Tracker Report every month after the close, next
to the forecast and the forensic reading. Request the full specimen
report on a confidential diligence call.