One analytical move: turn a market into a number you could be wrong about, and label where every input came from. Numbers without provenance are decoration.
The top-down move - "the French proptech market is worth X billion, we need 0.1 percent of it" - is unfalsifiable and therefore useless. The bottom-up move counts units:
Every step is now attackable, which is the point. A number nobody can argue with taught you nothing.
| Fact | Source | Status |
|---|---|---|
| About 30,000 real estate agencies in France | industry panorama, 2025 | trade aggregation |
| iad France over 14,000 conseillers; SAFTI over 6,000 | immo2.pro ranking, 2026 | trade press |
| Mandataire networks did about 25 percent of existing-home transactions in 2023, up from 1.75 percent in 2010 | industry panorama, 2025 | trade aggregation |
| FNAIM has over 10,000 member establishments, roughly a quarter of all agencies | fnaim.fr | primary, undated |
| About 921,000 transactions in existing homes in the year to September 2025, up 11 percent, after a trough near 780,000 | Notaires de France, via trade press | primary, reported |
| No official national count of registered agents commerciaux immobiliers exists in public sources | - | not found |
Two readings fall straight out. First, the fastest-growing part of this market is individuals, not agencies, which is why the seat-cap warning in Lesson 4 was not academic. Second, the market is recovering hard from a two-year trough - transaction volume is back to 2017 levels - so agents have money and appetite again, which is a better condition to launch into than the last two years were.
Get in the habit of labelling every number as you write it down:
| primary | The party that owns the fact says it: a regulator, a vendor's own pricing page, a statistics agency |
| primary, reported | Trade press reporting a primary source. Usable; go to the original if the number decides something |
| trade aggregation | An industry blog or panorama assembling numbers from elsewhere. Fine for order of magnitude, never for a board slide |
| unsourced | Repeated with confidence and no origin. Treat as a placeholder and mark it as one |
The document under review contains this passage:
No source could be found for any of those three figures as stated for this sector. The closest real thing is the France Num baromètre 2025, a government survey of 11,021 firms, and a secondary blog claiming 66 percent of French small firms spend under 300 EUR a year on digital - a threshold, not a median, and covering all SMEs rather than estate agencies.1
Learn the smell rather than the individual correction. Three warning signs, all present here:
Note also the giveaway phrasing: "if a dev challenges it, your chain is". That is the author preparing to win an argument rather than to be right. When you catch yourself assembling numbers for the challenge rather than for the decision, stop.
The document says "November" repeatedly without saying why. It is almost certainly RENT (Real Estate & New Technologies), the French proptech trade show: 4-5 November 2026, Paris Expo Porte de Versailles, about 12,000 professionals and over 400 exhibitors.2 The next date after that is the FNAIM congress on 7-8 December 2026.3
Confirm it with the author, because the entire schedule argument depends on it, and because it converts a vague deadline into a hard one: from late August 2026 that is about ten weeks. Any decision about what is shippable has to be argued against that number and not against a season.
Before researching any number, ask: would a different value change the decision? If the answer is no, write "order of magnitude only" and move on. Most market research is performed on numbers that could not have changed anything, while the one number that decides the outcome - here, the conversion rate - goes unexamined.
Skim the France Num baromètre 2025 summary. It is the actual source people are half-remembering when they quote French SME digital statistics, and knowing what it does and does not measure inoculates you against the whole genre.
Want the beachhead sized properly, with the numbers traced to primary sources rather than trade blogs? That is a focused research session. Ask your teaching agent.