Specialty Café·214 Oxford Street, Leederville WA 6007

Appendix A1

Methodology

Each report is assembled by an 8-agent pipeline (A1–A8) that runs in parallel against current data sources — Google Maps for competitors, ABS Census 2021 for demographics, and category- specific financial benchmarks. Rent benchmarks and demand scores are AI-estimated from suburb context and category patterns. Agents that fail or time out fall back to category-level benchmarks; the confidence label reflects the proportion of measured vs estimated data.

Composite score weighting

Verdict ladder

• PROCEEDEconomics and location signals support proceeding with normal diligence.
• VERIFYSplit read — economics, competition, demand, or location need verification before you commit.
• AVOIDEconomics or competition read as unfavourable on measured inputs.
• INSUFFICIENT DATACore commercial inputs are missing — verdict is provisional, not a measured rejection.

Rent is compared against a workable band of 8–18% of monthly revenue for the category — the industry-standard hospitality viability threshold, consistent with CBRE Australian Retail Benchmarks and Retail Doctor Group occupancy-cost guidelines. Inside the band means rent burden sits within the model range; above the band signals margin risk unless pricing is premium.

Appendix A2

Data sources

Locatalyze blends measured and benchmark data. The table below lists the source for each major section. The Data Quality section in the main report shows which data checks resolved for this specific report.

Appendix A3

Confidence framework

Every Locatalyze report carries a confidence label. The label is informational — content is never hidden based on confidence. A low-confidence report displays every section with the verdict dimmed and qualifiers applied. This appendix explains what each label means so a recipient can read the memo with appropriate weight.

Appendix A4

Glossary

Composite score

0–100 weighted average across the five axes (Rent, Competition, Demand, Profitability, Location). Drives the conviction number on the cover.

Conviction

The composite score, shown as a single number 0–100. High conviction is not the same as high confidence — a confident report can still rate the site as a no-go.

Rent-to-revenue

Monthly rent divided by modelled monthly revenue. Compared against the workable band for the category.

Workable rent band

8–18% — the rent burden range typical operators can absorb in this category.

Break-even

Months from opening to cumulative profit ≥ 0, given the model's cost stack and ramp curve.

Data signal

Percentage of inputs that resolved against measured external data (not category benchmarks).

Failure modes

Specific tripwires that would invalidate the verdict if observed in person.

Path to PROCEED

Sequence of verifications required to lift a VERIFY recommendation to PROCEED.

Recommendation

PROCEED / VERIFY / AVOID — the lease call shown on the hero, separate from the Location Score. INSUFFICIENT DATA when core inputs are missing.

Analysis model

Version of the Locatalyze analysis model used to produce this memo. Re-running on a newer model may shift figures within the bands.

Appendix A5

Disclaimer and report metadata

This memo is decision support, not financial, legal, or tax advice. Locatalyze synthesises public, licensed, and modelled data into a structured read; figures are model outputs subject to the assumptions and data quality described above. Recipients must conduct their own diligence before committing capital, signing a lease, or making any other binding decision.