Home / Experiments
Experiments
Each experiment is a real property, tested against a real market.
An experiment either proves there is a business worth building, or it tells us to stop early and cheaply. Both outcomes are useful, so both are kept.
Live properties
These are online now. Every property is a separate project with its own owner, codebase and deployment. DatumHarbor observes them; it does not modify their code or product logic.
-
TEST LIVEMoney Machine Intelligence
money.datumharbor.com
How real businesses create value and generate revenue.
- Days Running
- —
- Signals
- —
- Inquiries
- —
-
TEST LIVEEmerging Industries Intelligence
industries.datumharbor.com
Tracking new industries, markets and technologies before they become crowded.
- Days Running
- —
- Signals
- —
- Inquiries
- —
-
TEST LIVECommercial Failure Intelligence
failure.datumharbor.com
Why commercial efforts fail, and what is reusable from that.
- Days Running
- —
- Signals
- —
- Inquiries
- —
-
TEST LIVEProblem Discovery Engine
problem.datumharbor.com
Which problems are commercially real rather than assumed.
- Days Running
- —
- Signals
- —
- Inquiries
- —
-
TEST LIVEQianmen — Defensive Scam Intelligence
scam.datumharbor.com
How scam patterns can be recognised and defended against.
- Days Running
- —
- Signals
- —
- Inquiries
- —
-
TEST LIVEEnterprise AI Work Intelligence
aiwork.datumharbor.com
Where real enterprise AI deployment actually stalls.
- Days Running
- —
- Signals
- —
- Inquiries
- —
-
TEST LIVEGlobal Innovation Radar
radar.datumharbor.com
Which innovation signals are worth acting on.
- Days Running
- —
- Signals
- —
- Inquiries
- —
-
TEST LIVE生意标本 (GSBI)
case.shengyibiaoben.cn
真实小生意的赚钱机制是否可以复盘复用。
- Days Running
- —
- Signals
- —
- Inquiries
- —
All eight properties are checked for reachability at every review. Lifecycle shows the publicly verifiable state; signal columns stay blank until first-party measurement exists.
The same properties as a record
The card view and the table below describe the same eight properties. The table is the flat record; the cards are the readable one.
| Property | Domain | Question being tested | Status |
|---|---|---|---|
| Money Machine Intelligence | money.datumharbor.com | How real businesses create value and generate revenue | Live |
| Emerging Industries Intelligence | industries.datumharbor.com | Tracking new industries, markets and technologies before they become crowded | Live |
| Commercial Failure Intelligence | failure.datumharbor.com | Why commercial efforts fail, and what is reusable from that | Live |
| Problem Discovery Engine | problem.datumharbor.com | Which problems are commercially real rather than assumed | Live |
| Qianmen — Defensive Scam Intelligence | scam.datumharbor.com | How scam patterns can be recognised and defended against | Live |
| Enterprise AI Work Intelligence | aiwork.datumharbor.com | Where real enterprise AI deployment actually stalls | Live |
| Global Innovation Radar | radar.datumharbor.com | Which innovation signals are worth acting on | Live |
| 生意标本 (GSBI) | case.shengyibiaoben.cn | 真实小生意的赚钱机制是否可以复盘复用 | Live |
Formal per-experiment records — hypothesis, target user, commercial question, lifecycle state, decision and dated timeline — are maintained in the control platform and are not published on this site yet. Until a record has evidence behind it, this page shows only what is directly verifiable: that the property exists and is live.
What a registered experiment will contain
When records are published, each one will state:
The claim
What is being tested, for whom, which problem, and the commercial question it should answer.
Where it is
One lifecycle state, from idea through live testing to proven — or closed.
What was observed
Search, traffic, behaviour and commercial signals, each tied to the evidence behind it.
The decision
Continue, hold, or stop — with the reasoning and the date it was taken.
A stopped experiment is still a result
We expect to stop things. Stopping early is the point of testing cheaply. What must never happen is losing the reason: a stopped experiment stays in the dataset, and its evidence is never deleted.