6 · Blind spots & confidence register

⚠️ What the model does not know. Read again before any binding decision.

The principle

A model that does not show its weaknesses is misleading. These are the areas where assumptions rest on shaky ground, ranked by the cost of getting them wrong.

Points added by the investor review of 7 September 2026

• Capacity and margin need reconciling. The 1,335 m² candidate implies around 890 members under the assumption of 1.5 m² per member. The stated break-even point cannot be reproduced from the exported costs alone. See the Club economics chapter.
• Current competition. Villa M already combines a hotel, fitness and recovery in Marseille's 8th arrondissement. The competitive landscape must include substitute offers, beyond clubs advertised above €150/month.
• Replication is unproven. A second site must operate with a salaried team and acquire members beyond the founder's network.
• Three formats, three models. The compact format and the hotels must each justify their own capacity, investment, cash requirements and returns. Compact recovery remains a working assumption.

Major blind spots

1. The capture rate is calibrated on two clubs in one city

The 4–5% figure comes from Canal and Lamarck. Paris has a density, walking culture and competitive landscape found nowhere else in France. There is no guarantee that the same club in Marseille or Versailles would capture the same share of its catchment. What could address this: ask operators of premium clubs outside Paris about their actual membership-to-catchment ratio.

2. Attrition is an industry assumption

Annual attrition of 12% is an assumption to verify with operators. It must be distinguished from the separate assumption that ancillary sales equal 12% of membership revenue. Attrition is the model's second-largest driver after rent: it determines annual acquisition volume, which determines the marketing budget and therefore the CAC ceiling. What could address this: operator data or, failing that, a stress scenario at 20%.

3. The isochrone is an abstraction

15 minutes by car, on foot or by public transport? The model treats the isochrone as a uniform circle. Reality is a network with barriers, such as a railway line, a main road or a bridge, that split an apparent catchment in two. What could address this: map an actual isochrone around the proposed site for each transport mode.

4. The brand value of the social space is not modeled

It is assumed. That is honest, but it means the model cannot say whether 200 m² of social space is enough or whether 400 m² is needed. This is a design decision made outside the model.

5. The final cost of the fit-out

Repurposing an existing building is central to the real estate strategy, and it is also where budget overruns are most common. The model uses a cost per m²; it includes no contingency for structural issues.

Secondary blind spots

• Seasonality. Membership does not build evenly; January and September drive acquisition, while summer brings a dip. The monthly cash flow plan must reflect this.
• Future competition. The model freezes the competitive landscape at the time of analysis. A new entrant in the same catchment changes the capture rate.
• The mix of resident members and local workers. They attend at different times and cancel for different reasons. The model treats them as one segment.

How we improve the assumptions

Ranked by information gained relative to cost:

1. Operator interviews: actual attrition and capture rates cannot be found anywhere else.
2. An isochrone mapped around the site: a few hours of work could potentially correct 30% of the assumed catchment.
3. Test pre-registrations: the only measure of desirability that goes beyond stated intentions.
4. Fit-out quotes for an actual building: turn an assumption into a number. This is exactly the order of the three-week protocol. → See the Assumptions database for details on each parameter, its source and the confidence behind it.