Agreed, though keep the labels few. Too many categories create false sophistication. The minimum distinction is whether a number is evidenced, estimated or missing.
For investment modelling, test the variables that can actually change the decision: purchase price, finance cost, renovation budget, vacancy period, rent assumption and exit value. Formatting twenty minor inputs will not rescue weak core assumptions.
And record what would make you reject a unit. A model is more useful when it identifies a stopping point than when it can be adjusted until every property appears viable.
Photographs taken at different listing dates can reveal condition changes or an incomplete fit-out. Keep them associated with their dates rather than assuming every image shows the property at completion.
Agent descriptions are useful evidence of how a unit was marketed, but not independent verification. Put promotional statements in a notes field until another source or inspection supports them.
The management section should distinguish owner responsibilities from occupier responsibilities. If that allocation is unclear, do not quietly assume the cheaper version in the cash flow.
A decision journal might help: why the unit entered the shortlist, which facts changed, and why it stayed or was removed. That creates better learning than a folder of disconnected listings.
There is a risk of overengineering this before Luca has a real target. Ten rows with carefully checked facts will teach more than a hundred-column template filled mostly with blanks.
Fair. My minimum row would be location, tenure, occupancy, floor area, condition, first asking price and date, later price and date, completed figure if known, plus evidence and uncertainties.
It sounds as though a pinned Manchester retail comparison thread could work if contributors use a shared core format. Local observations could stay readable without forcing everyone into an elaborate model.
Thanks all. I’m not under offer on a unit; this is groundwork before narrowing a first retail purchase. I’m comparing vacant and tenanted options, but will separate them. I’ll begin with a small Manchester set, preserve listing dates and reductions, and mark every cost or price as evidenced, estimated or unknown.
That scope is sensible. When you build the first version, could you post only the column headings and one anonymised example structure? Members can challenge the method without speculating about a particular seller or property.
Since this is pre-purchase work, add the point at which legal input would be sought. A cheap-looking unit can consume time if basic access, repair or occupancy questions remain unresolved too long.
Utilities and building services are another early question, especially where parts are shared. You do not need to solve everything during screening, but unknown arrangements should be visible in the renovation and management cases.
Take dated snapshots rather than constantly overwriting one row. Otherwise you lose the difference between what you believed at first viewing and what later information established.
For finance comparisons, record when each illustration was obtained and which assumptions you supplied. Offers gathered at different times or on different property scenarios should not be presented as direct alternatives.
Also model finance fees separately from the interest calculation. Combining them into one unexplained rate makes it harder to compare options or see the cost of a shorter holding period.
Insurance belongs in the questions list too, but the relevant cover and cost will depend on the building and occupancy. Use a placeholder rather than copying a residential assumption into a retail model.
My three-case suggestion for Luca’s vacant options: quick usable condition, delayed opening after works, and a more difficult delay. The point is not prediction; it is seeing whether modest disruption changes the decision.
For tenanted options, run a separate interruption case rather than merely reducing annual income by a tidy percentage. Timing matters when finance and works continue while income does not.