13947 Geranium Place.
$255K on the surface.
The story lives four docs deep.
A Wellington flip from $670K to $925K. Pulled from the county clerk, decomposed into market drift versus operator-attributable lift, and modeled live with the inputs only the operator knows. This is what a deal looks like once iterFact stacks the public-record layers.
Wellington, FL 33414
Four recorded instruments. The skeleton of the entire deal.
Aggregator data is a guess. County data is the answer. Pulled from Palm Beach Clerk Official Records via the book/page index – this is the actual document chain that opens and closes the flip, with instrument numbers and recording dates intact.
Alyona Davidova acquires the parcel
Conventional purchase mortgage at 87.4% LTV
End-user buyer acquires the renovated property
Buyer takes 80% conventional – an end-user signal, not another flipper
The aggregator-reported $585.4K mortgage was only medium confidence until the Clerk record was pulled. The actual instrument moved it to county-confirmed – with lender, borrower, recording date, book/page, instrument number, parcel id, and legal description on the record. This is the ingestion layer iterFact should automate across Florida: deeds, mortgages, satisfactions, modifications, liens, lis pendens.
Market drift made $18K of it. The operator made the other $237K.
A flip is a thesis: that an operator can create more value than the market can. The public-record proof starts by decomposing the gross spread into market tailwind versus residual lift – and then cross-checking it two different ways.
The proof: lot barely moved. The structure carried it.
PAPA’s 2024 valuation split the parcel 47.2% land / 52.8% improvements. Apply that split to the $670K basis and walk it forward at the actual market drift. The land contribution is small. The structure contribution is the entire story.
This was not a passive market trade. Wellington 33414 only moved 2–3% over the hold – that’s ~$18K of pure tailwind on a $670K basis. The remaining $237K of residual gross lift is where renovation scope, condition upgrade, timing, and marketing execution have to explain the spread.
$431/sqft in a $318/sqft zip. Two methods, one answer.
The waterfall said residual lift was ~$237K. A second, fully independent method – price-per-square-foot premium against the strongest local comp – lands in the same range. When two methods converge, the answer stops being a hand wave.
− $670K basis
− $18K market drift
× 2,146 sqft
= premium over best comp
Two independent methods converge on roughly $215K–$237K of gross lift after market drift. That’s the headline number. Whether it was true profit depends on renovation cost, carry cost, financing terms, and rental optionality – the inputs in the next section.
Drag the sliders. Watch the deal turn into a real number.
Public records lock the skeleton: purchase, sale, original mortgage, taxes, hold period. The operator owns the rest: renovation cost, rental income, carry costs, payoff, and closing structure. Move the sliders, and iterFact rebuilds the net P&L, return on equity, and break-even thresholds in real time.
13947 Geranium Pl · Operator P&L
Tagged inputs: PUBLIC = pulled from county / DOR / Census. EST = iterFact estimate, adjustable. YOUR INPUT = only the operator knows.
Inputs & Line Items
P&L Breakdown
A pure flip with $100K of renovation and zero rental income lands in loss territory at typical FL closing/carry costs. Add 5 months of seasonal rent at $12K/mo, and the same deal turns profitable. The key investor question becomes simple: was this only a flip, or did rental income help carry the hold?
Property taxes nearly doubled between 2024 and 2025. The new owner inherits that.
FL caps assessment increases for homestead-protected owners. A non-homestead investor purchase resets that cap. The 2025 bill on this parcel jumped +84% – a number that compounds for the next operator.
Every FL acquisition resets the assessed-value clock. Underwriting at the seller’s tax bill is a structural error. iterFact should default the post-acquisition tax line in any model to the just-value-derived rate, not the seller’s historical bill. On Geranium, that’s a $5K/year delta – over 2 years, it’s a $10K hole in the underwriting.
Where the public record stops – and the next layer should start.
Two specific gaps on this property surface a clear backlog. The 4-bed-versus-5-bed mismatch needs permits to resolve. The renovation cost needs permits and contractor records to narrow. Both are public, both are missable, both are the next data layer iterFact should ingest.
Bedroom count: official versus marketed
The Wellington EnerGov / Citizen Self Service permit portal is the source of truth for whether the bed/bath upgrade was permitted (legitimate value-add) or functional/marketing (potential disclosure issue). Same portal exposes total renovation permit value – which would narrow the renovation-cost slider above from a ±$100K range to a ±$15K range.
What this artifact tells iterFact to build next.
Every section of this teardown either uses a layer iterFact already has, or surfaces a layer it doesn’t. The list below is the ranked backlog implied by what would have made this analysis even sharper, with priority and the section it would have improved.
Six ingestion + product moves the Geranium teardown is asking for.
Each task below maps to one or more sections of this artifact. P0 items are the gaps that hurt this analysis the most. P1 items are the next moat layer.
Pulled the county chain. Decomposed the value. Cross-checked it. Made the deal modelable.
From a single address – in one engine pass – iterFact resolved the parcel, walked the recorded document chain, separated market drift from residual lift, ran the comp premium cross-check, and turned the whole thing into a live P&L modeler. This is the workflow to automate across Florida. Drop in an address, and the next teardown can come back at this depth.