How Linden calculates a wedding budget

The inputs

Seven onboarding answers seed every budget: partner names, city, wedding date, expected guest count, target total (optional), style + venue approach, and a priority ranking across 22 categories. Nothing else is required — the engine derives everything else from these.

The 11 layers

1. Resolve items — Filter the 400+-item master catalog down to the set that applies to this market, the selected cultural overlays, the style/venue approach, and the priority tier. A Copenhagen city-hall wedding draws from ~60 items; a full Hindu wedding in London, ~140.
2. Multipliers — Apply three independent cost indices: city multiplier (0.6× in Sofia to 1.8× in Zurich relative to the EU median), season multiplier (peak/shoulder/off), and day-of-week multiplier (Saturday is the reference; Friday & Sunday cut 10–18% depending on country).
3. Allocate — Distribute the total budget across active categories based on your priority ranking. The allocation honors market-specific baselines (catering is a bigger share in the US than in the EU because of service charge + gratuity stacking) and adjusts for guest-count scaling.
4. Calculate — Per-item estimates using the right pricing model for that item: per-person (catering, drinks), per-hour (DJ, photographer), per-table (rentals), per-household (stationery, postage), or flat (attire). Each model has a default that can be overridden at the item level.
5. Service charge — US-only layer. 18–25% added to food & beverage based on regional convention (Northeast typically 20%; Texas/Midwest 18%; Hawaii/West Coast 22%). Skipped entirely in the EU because service is already baked into European vendor quotes.
6. VAT / sales tax — EU VAT by country (7% in Luxembourg up to 27% in Hungary; Denmark sits at 25%). US sales tax by state + locality (Oregon is 0%; most metros land 5–10%). VAT is applied inclusive (price already contains it); US tax is exclusive (added on top).
7. Bundles — Reconcile all-inclusive venue packages against the individual items they replace. Prevents double-counting when a venue quote already includes catering, rentals and service — a common mistake in spreadsheet budgets.
8. Hidden costs — Surface warnings for items most couples forget: vendor meals, overtime (photographers and DJs after 10pm), welcome-bag printing & postage, bridal party transportation, day-of coordinator gratuity, tasting fees. Each carries a severity flag (INFO / WARN / BLOCK).
9. ML pricing — Optional calibration layer. When a market has enough anonymized community data (≥ 50 closed budgets in the past 18 months), a regression model adjusts individual item estimates toward observed prices. Currently active for DK, UK, and US-West.
10. Benchmarks — Attach percentile bands (p25 / p50 / p75) to each category so you can see where your budget sits relative to comparable weddings. Comparability is defined by city tier, guest count bucket, and overlay set — not just national average.
11. Risk score — Overall budget feasibility rating from 0–100, composed of: total vs. market baseline for the setup, unallocated buffer percentage, hidden-cost exposure, seasonal premium, and priority-to-allocation misalignment. Below 40 = high risk of overrun.

Where the numbers come from

See the companion data provenance page for the full list of seed sources per market, vendor-directory origin, refresh cadence, and how city multipliers are maintained.

What the engine cannot tell you

Three things the engine deliberately does not try to answer: (1) which specific vendor to book — the directory surfaces options but ranking is yours, (2) whether a given guest count is right for your relationship — that is not an engineering problem, (3) whether you should splurge on one category at the expense of another — the priority ranking is your call to make, not the model's.