The three ideas everything in the flagship analysis stands on, drawn simply. Full detail: the report · the top three.
The factory sells to our store. The store sells to you. The store's profit is the difference.
We looked up these two price tags for every mattress sold in the last six months — 99% of them checked (a real example above: one Hypnos Stapleford). On average, the store keeps 70 fils of every dirham of mattress sales before rent, ads and wages.
The trap we avoided: every sale is written in the books twice — once by the factory, once by the store. Count both and the numbers double; use the factory's cost and you measure the factory's profit, not the store's. We only ever counted the store's copy.
Full method: report §06b · dossier appendix
Cars drove past every day. Sales only moved when the adverts did.
In April 2026 the adverts were switched almost completely off. The store sat on the same Sheikh Zayed Road, with the same traffic — and sales stayed 60% down until the adverts came back. Over the whole year, sales rise and fall with the ad budget almost in lockstep. So we buy our customers with adverts, not with an expensive address — which changes what rent is worth paying.
The full evidence (r = 0.87, month by month): report §06
The mattress shop can pay the same rent anywhere. Extra rooms must earn their own keep.
Give the mattress shop the same profit target everywhere — Eiffel's 77,000 a month. It can then afford 79,000 of rent wherever it sits. At Al Joud that covers the whole building with money left over. At Umm Suqeim, the extra rooms must earn 32,000 a month on top — about a Festival-Plaza's-worth of sofa sales — just to keep up. That, in one picture, is the whole decision.
The full model: the Top Three page, Part 2b