War films are one of my favourite genres, as I have a keen interest in military history. These evocative cinematic experiences are traditionally defined by the deafening rattle of machine guns, dog fights in the sky and the visceral gore of the front lines, whilst telling a heroic story.
Yet director Anthony Maras’s Pressure, which I saw last week, reminds us that high-stakes military battles are often enacted in quiet, intense rooms where people make agonising choices on strategy and tactics.
Pressure strips away glorification of combat to focus on an unlikely variable that nearly derailed the Allied invasion of Normandy: the weather. The result is a claustrophobic, cerebral, and highly enjoyable thriller that transforms barometric readings, isobar charts, and meteorological forecasts into matters of life and death.
Set over the crucial 72-hours leading up to June 5, 1944 – the original scheduled date for Operation Overlord – the narrative follows Group Captain James Stagg, Chief Meteorological Officer for the RAF. Stagg is tasked with giving Supreme Allied Commander Dwight D. Eisenhower the final go/no-go forecast for the largest amphibious invasion in history. Stagg’s expertise and resilience under pressure undoubtedly changed the outcome of the War.
The amphibious invasion of D-Day required some key environmental conditions:
- Tides & Moon: Low tide at dawn to expose German beach obstacles, preceded by a full moon to guide airborne landings;
- Wind & Sea: Light winds so landing craft wouldn’t swamp;
- Cloud Cover: High, light clouds that permitted accurate Allied naval bombardment and air support without obscuring targets.
Because of tide and moon requirements, there were only three suitable days in early June 1944: 5, 6, and 7.
In late May 1944, weather across the English Channel was unusually mild. Colonel Irving Krick, a brash American Army meteorologist also advising Eisenhower, relied on historical analogue forecasting, matching current weather patterns to historical data, He confidently predicted clear skies, recommending go-ahead for June 5.
Tension flared between Stagg and Krick into a fierce ideological and methodological conflict. Stagg, monitoring pressure reports from isolated weather ships and stations in the North Atlantic, utilising dynamic modeling and synoptic charts, detected a series of deep low-pressure systems sweeping rapidly eastward: a severe Atlantic storm was headed for the Channel that threatened to turn it into a churning graveyard for Allied landing craft and men.
On the evening of June 3, Stagg presented his forecast to Eisenhower and senior officers: June 5 would bring high winds, rough seas, and heavy cloud cover. Launching under these conditions ran a massive risk of swamping landing craft and neutralising air support.
If Stagg was wrong and Eisenhower delayed, the invasion armada risks being exposed to German intelligence, exhausting supplies and collapsing troop morale; if Krick is wrong and Eisenhower launches, thousands of soldiers will drown before reaching Normandy. With the clock ticking and generals demanding action, Stagg stands his ground against immense geopolitical pressure. Overriding Krick’s optimistic predictions, Eisenhower chose to postpone the invasion.
As gale force winds battered England, Stagg brought crucial news to the 4am briefing on June 5: he’d identified a brief, transient ridge of high pressure moving behind the main storm system. While conditions on June 6 would remain rough and far from ideal, there would be a 24-36 hour break in the storm. The window was narrow and marginal: Eisenhower pondered the decision before giving the order: OK, we’ll go.
Stagg was intensely stoic, operating on the brink of exhaustion. He was a quiet, deep thinker, carrying the psychological weight of knowing that an error in his calculations will send thousands of men to a watery grave. The film captures the pressure – atmospheric, political and psychological – and the extraordinary calm of the man who stood firm on all three.
Stagg’s forecast proved remarkably accurate. German Chief Meteorologist Major Heinz Lettau and his team observed the same series of low-pressure systems hitting Western Europe the morning of June 6. However, they failed to spot the brief ridge of high pressure trailing behind the main storm system. They assumed the bad weather would continue. As a result, many German officers left their posts, leaving defences unprepared when the assault began on the 6 June.
Taking analogies from Stagg, founders need to know the prevailing conditions and headwinds they’re facing to know when to be bold and invest, and when to hunker down and ride out a turbulent period. Founders face the same challenges as Stagg re uncertainty in data, limited availability, accuracy and timing criticality of decisions.
For a founder, waiting for the right data before building a financial forecast is a mistake. Early startups operate in a low-data environment characterised by handfuls of closed deals and raw, preliminary metrics. Attempting to apply modeling to a dataset of twenty customers can yield noise, not insight.
However, founders operating without a forecast have nothing on which to anchor decisions. Managing burn rate requires forward-looking visibility. Early stage forecasting is not about statistical significance, but about founders using small datasets to build bottom-up operational models that reflect the mechanics of their business, and for them to apply judgment to the numbers: What are they telling us?
So what’s a sensible approach to a start sales forecast? Here are some thoughts:
1. Adopt a bottom-up approach. Traditional forecasting relies on top-down assumptions, taking a percentage of TAM, and extrapolating historical run rates. For a startup with minimal historical data, top-down forecasting is purely speculative.
Small-data forecasting requires a bottom-up approach rooted in unit economics and capacity constraints. Rather than asking What will our revenue be in Q4?, founders should break the revenue generation process into a sequence of steps:
- Lead generation capacity: How many contacts can we meet or canvas per week?
- Funnel conversion rates: Based on the last 30 leads, what percentage moved from discovery to demo, to proposal, to closed-won?
- Cycle velocity: How many days elapse, on average, between first contact and cash in the bank?
By tracking these basic milestones across even a small sample, you create a deterministic revenue pipeline model informing cash inflows projected directly from the number of demos scheduled.
2. Qualitative context. Large datasets smooth out individual anomalies through volume. Small datasets are distorted by outliers, such as one large contract or a client who closed in three days. To prevent skewed projections, layer qualitative context over quantitative metrics:
- Opportunity rating: Rather than applying a blanket probability stage, evaluate small deal pipelines using weighted stage criteria: technical fit, and pricing agreed.
- Customer segmentations: Group small datasets by customer persona or deal size rather than treating them as a single pool. Five deals of £50k behave completely differently than fifty deals of £1k.
- Outlier stripping: When calculating average sales cycles or ARPU, exclude anomalous mega-deals to establish a more realistic baseline.
3. Cash vs. sales vs. revenue. A critical distinction is that sales are events that convert into revenue, which is an accounting metric, while cash is a survival metric. Small datasets often obscure the timing gap between the three. Converting a sales forecast into a revenue (billing) event into a cashflow forecast requires mapping contract payment terms onto phased calendars.
4. Sensitivity analysis. When input data is sparse, confidence intervals are wide. You cannot rely on a single model. Instead, small-data forecasting requires building three different scenarios:
- Expected case: Uses current baseline conversion rates, sales cycle duration, and typical receipt timing.
- Downside case: Test vulnerability by reducing top-of-funnel leads by 25%, extend sales cycles by 30 days, and introducing a 10% churn or bad debt rate.
- Upside case: Assume 10% improvements in driver metrics – conversion, AOV, deal progression – to model the impact of exceeding baseline drivers.
The primary objective of the Downside stress case is to identify the cash position under adverse conditions, to inform action over protecting or extending runway.
5. Establish a continuous feedback loop. Small-data models are dynamic instruments that must be updated constantly. As new data points arrive, the model must be recalibrated. Maintaining a rolling three-month forward looking picture, updated on a weekly basis, is a valuable exercise and management tool. Through this disciplined iteration, small datasets compound into actionable insight. The founder moves from guessing based on intuition to steering the business through calibrated financial mechanics.
Like Stagg, founders navigating the early stages of their venture require balancing intuition with cold, hard information. Without a clear data strategy, founders risk chasing vanity metrics, not having a handle on their performance and consuming cash.
To build an adaptable, resilient startup, founders should harness four categories of data to keep an eye on the prevailing conditions and forecast to make informed decisions.
1. Customer & product data. Product data tells you what users actually do, rather than what they say they do. Primary qualitative data validates your hypothesis, while quantitative behavioral data proves scalability.
Key Metrics:
- Retention & Activation rates
- Funnel Conversion Rates
- Debtor Days
2. Financial & operational data. Financial data is your survival instrumentation. It tells you how long you have to live (runway), how efficiently you’re growing (unit economics), and whether your core business model works (route to profitability).
Key Metrics:
- Burn Rate & Cash Runway
- Unit Economics (LTV to CAC)
- Gross Margin
3. Market & competitor intelligence. You cannot operate in a vacuum. Understanding the size of your opportunity (TAM/SAM/SOM), tracking competitors, and benchmarking to peers prevents mispricing and strategic blind spots.
Key Benchmarks:
- Market Share & Growth
- Competitive Moat Mapping
- Competitor Price Checking
4. The North Star Metric. While managing multiple metrics, founders should identify a single North Star Metric, you primary indicator that measures core customer value delivered.
- For Airbnb, it was Nights Booked.
- For Slack, it was Daily Messages Sent Within a Team.
- For a Fintech app, it might be Daily Active Users.
When secondary dashboards conflict, your North Star Metric provides directional clarity, ensuring that every team member builds toward sustainable customer value creation rather than being distracted by short-term wins.
Summary As a founder, imagine yourself as a pilot, who needs a core of flight parameters to tell him what the aircraft is doing in real time currently and for the remainder of the flight:
- Airspeed: Speed relative to the surrounding air mass;
- Altitude: Height above sea or ground level;
- Heading: Actual flight path across the ground;
- Position & Routing: Distance/ETA to destination;
- Tactical Weather: Wind speed and route forecast;
- Power Output: Engine RPM;
- Fuel Metrics: Burn rate and calculated fuel remaining at destination;
- Temperatures & Pressures: Engine metrics;
- Electrical & Hydraulics: Mechanical metrics;
- Airspace Boundaries: Flight plan airspace sectors.
The focus is combining Leading indicators – what information is saying about the future journey, and Lagging Indicators – what information is saying about the journey complete to date. From this example, build a balanced scorecard of data needed to help guide your decision making on short, medium and lon-term decisions in your startup.
Stagg had no computers to help him assimilate the data in June 1944; Eisenhower acted on the best available judgment available rather than waiting for a certainty no forecast could give him. It’s an amazing true story of decision making and judgment under pressure which offers useful insight for founders.
Check out this video about Stagg and events that made Overlord possible: How 1940s Meteorologists Predicted The Weather For D-Day.