MEDIAN ≠ GRADE
Distribution context without winner and loser labels
Every metric shows reported coverage, median, range, and below/at/above context. No composite score, league table, causal claim, or automated recommendation.
For community & property management portfolios
See the pattern. Keep the context.
PortfolioBench compares finalized community-period aggregates against one human-reviewed scorecard—showing coverage, median, and range without resident records, a composite grade, or a leaderboard.
Aggregate only · minimum cohort 3 · no leaderboard or automated ranking
NORTH REGION · JUL 2026 · 5 FINAL
Community A · 7 daysbelow cohort median · context onlydone
Community B · 11 daysat cohort median · context onlydone
Community C · 15 daysabove cohort median · investigatelive
0resident, owner, unit, or account rows stored
3+finalized communities before comparison
$49planned flat monthly rate
The daily grind
A spreadsheet can compare numbers that do not mean the same thing.
One community counts calendar days, another counts business days, a third changed its queue definition, and the dashboard still paints them red, yellow, and green as if the difference proved something.
01
Definitions drift
The metric label stays the same while source, unit, bounds, or start/stop event changes.
02
Periods do not line up
Month-end snapshots and rolling exports appear side by side without a warning.
03
Missing data becomes zero
Optional coverage disappears into an average instead of remaining visibly absent.
04
Context becomes a verdict
A descriptive outlier becomes a staffing, vendor, employment, or governance decision without investigation.
What it does
A cadastral scorecard, not a performance leaderboard.
PortfolioBench locks metric meaning before collection, freezes each reviewed community period, requires a privacy floor, and makes the next experiment a named human decision.
#
Definitions first
A named human reviews the exact label, definition, unit, direction context, required flag, bounds, and source note.
▦
Periods finalized
Community snapshots share one scorecard and exact period, carry a source fingerprint, and freeze after named review.
◇
Privacy floor
At least three finalized communities are required. PortfolioBench stores aggregate values, not resident or account events.
∅
Missing stays missing
Optional metric coverage is explicit. PortfolioBench never turns an unreported value into zero or silently imputes it.
→
Experiments stay owned
A human chooses the community, metric, hypothesis, owner, and due date. The software never opens a corrective action by itself.
How it works
Define. Finalize. Compare carefully. Assign deliberately.
STEP 01
Activate the scorecard
Review metric meaning, unit, source, bounds, required status, and comparison limitations.
STEP 02
Finalize each period
Record only aggregate values from a preserved source fingerprint; a community reviewer freezes the exact set.
STEP 03
Build the cohort
Use finalized snapshots from the same scorecard, group, and period after the privacy floor is met.
STEP 04
Choose an experiment
Investigate mix and conditions first, then name a hypothesis, owner, and due date if action is warranted.
Pricing
Less than one misdirected initiative, resident-data warehouse, or dashboard verdict built on mismatched definitions.
Start with one region, one monthly period, three to five communities, and three metrics your team can define precisely.
Pilot
$0
Run one controlled cohort.
- One reviewed scorecard
- Up to five community snapshots
- Coverage, median, and range report
- One human-owned experiment
Founding rate
$49/mo
Portfolio context without the resident warehouse.
- Unlimited scorecard versions and periods
- Aggregate-only community registry
- Append-only corrections and final review
- Cohort privacy floor
- Deterministic reports and fingerprints
- Owned experiment history
- Rate locked for life
Questions
Before you ask.
Does PortfolioBench connect to our product databases?
No. The pilot records approved aggregate values plus an external source reference and SHA-256. It does not receive source credentials or read DuesChaser, AppFolio, Buildium, or another live database.
Does it store resident, owner, unit, or account data?
No resident-level rows are part of the schema. Community business references, group, aggregate unit count, period, source fingerprint, and aggregate metric values are the boundary.
Does above the median mean a community is underperforming?
No. Position is descriptive context. Definitions, workload, portfolio mix, seasonality, staffing model, data quality, and external conditions must be investigated before a human decides anything.
Can it rank managers or recommend staffing or vendors?
No. PortfolioBench does not rank people, generate a leaderboard, infer causation, or make staffing, vendor, employment, enforcement, or governance decisions. Those require the ordinary authorized human process.
Get started
Find the pattern without losing the operating reality behind it.
Join the early-access pilot with one region, one exact period, and a metric dictionary your team is prepared to defend.
No spam. Unsubscribe in one click, honored immediately.