Buyers pay for checks and monitoring. Developers never pay Kolaboro: the evidence sets every Score. Agents never take money from developers.

Is Kolaboro a lawyer or a notary?

No. Kolaboro agents verify facts on the ground and in documents and tell you what they found. For legal advice and to sign, use an independent lawyer or notary; we can tell you what to ask them.

Do you hold my money?

Not yet. Today your agent confirms each milestone on site and tells you whether the evidence supports paying. Milestone escrow with licensed partners is on our roadmap.

We are launching in Bali first and adding cities as we train local agents. If you are buying somewhere else, tell us where: demand decides the next city.

How do you decide "go" or "hold"?

Nine checks, from land title to realistic promises. A red flag on title, permits or progress means "hold" until it is fixed, and every finding links to its evidence. The full method is public.

Can a developer influence a report?

Only with evidence. Developers can send documents and reply, but never pay us. Where something can't be verified, the report says so instead of guessing.

What if I already bought?

That's when monitoring matters most. We can verify where construction really is before your next payment.

Kolaboro / Database / Delivery-delay index

Delivery-delay index

How far tracked off-plan projects are behind the first handover date we have on record, market by market, and an honest estimate of each project's odds of handing over within 6 months of it. Every figure can be recomputed in your browser from the open data.

Read this first. History for this index starts in November 2026. The projects are the ones being marketed now, so finished projects are under-represented. Dates are what sellers advertise.

…past the first promised handover, all markets

…median months late among them

…90th percentile, months late

…projects these figures rest on

Share past the first promised handover

Markets with at least 5 projects that are not completed and have a usable handover date. Select a market to see each project's months late.

All markets

Market

Every market

"Too few" means fewer than 5 projects stand behind the figure, so we do not show it. Months late are counted from the first handover date on record to this month.

Market	Tracked	Dated, not completed	Past first date	Share	Past today's date	Median months late	90th percentile

History

One snapshot is frozen on the first run of each month, so the index becomes a time series. It starts with the first snapshot; nothing before it is estimated.

Month	Taken	Dated, not completed	Past first date	Share	Median	90th percentile

Odds of handover within 6 months

For one project: what share of comparable projects, whose 6-month window has already closed, handed over in time? Enter a project id, such as DXB-001.

…closed projects (window over)

…of them handed over in time

…pooled rate, 95% interval

Back-test: how well would this have done on closed projects?

Why the comparables are not yet matched on stage and time to handover

Every status, with counts

Status	Projects	Meaning

How it is computed

Now. The month of the open data file (…), counted as a month index: year × 12 + month.
Today's handover date. If the overdue watcher says a project is k months past its advertised handover, that date is now − k. Otherwise it is December of the advertised year. December is the latest month the year allows, so this can only understate lateness.
First handover date on record. Today's date minus the recorded months the date has moved (delay). With no recorded move we assume none, which can only understate lateness.
Months late. now − first, for projects not marked completed whose first date has passed.
Share. projects past the first date / projects not completed with a usable date. Projects with no date, and the few with a past year but no overdue reading, are left out and counted.
Median and 90th percentile. Linear interpolation between ordered values: position h = (n − 1) × p, value x⌊h⌋ + (h − ⌊h⌋) × (x⌊h⌋+1 − x⌊h⌋), the same rule as numpy and PERCENTILE.INC.
Minimum sample. A market figure needs at least 5 projects behind it.
Odds. A project's window is closed when its first date is more than 6 months ago. It counts as on time if it is marked completed and its recorded moves add up to 6 months or less. Comparables are closed projects in the same market (at least 10), else the same country (at least 10). The interval is the Wilson score interval at 95%: (p + z²/2n ± z√(p(1−p)/n + z²/4n²)) / (1 + z²/n), z = 1.96. No model is trained; this is a count.

Recompute it in your browser

Your browser downloads the open project file (about 1.4 MB), checks its SHA-256 against the one the index was built from, and recomputes every market row, every project's status and the pooled odds with the rules above.

Data: /data/delay-index.json and /data/on-time.json (CC BY 4.0), rebuilt every day from /data/bali-cases.json (SHA-256 …). API: /api/x/v/on-time?id=…. Monthly figures for journalists: press release. Related: asking-price index, transparency index.