How the Desk works

Every status on this dashboard is a judgment call, not a score spat out by a model. Here is exactly how those calls get made, so you can decide how much to trust them.

The four states

Each topic sits in one of four states. The state is the fastest thing to read on the dashboard, so it always means the same thing.

OpenThe gap is real and there is room. Attention is low to moderate and no dominant solution has taken hold.

Heating UpAttention is rising fast. New entrants, spiking discussion, or fresh signals in a short window. The window is opening.

CrowdedMultiple credible players are now chasing the same fix. Still a real problem, but the easy version of the opportunity is going or gone.

ImprovingThe underlying system itself is getting fixed, not just built around. A platform change, a policy shift, or regulation is closing the gap.

What actually moves a status

These events are what trigger a re-score of a topic. Once its scores change, the status follows from the rule further down this page, and the read explains the call.

Open to Heating Up: two or more approved signals in a rolling two week window, or one signal showing clear new entrants or a jump in attention.
Heating Up to Crowded: a competitor ships a direct solution, or several signals in the same window point at the same fix rather than the same problem.
Anything to Improving: a signal shows the system itself changing. Platform policy, an official feature, regulation.
Back to Open: rare, but a signal showing the fix failed or reversed can reopen a gap.

The scorecard

Every topic carries four numbers, each 0-10, each with a one-line justification behind it (tap the (i) on the card to see it). None of these are outputs of a model — they're a judgment call, scored against what's actually documented in that topic's evidence, builders, and signals. A topic with no viable builder and no product-shaped fix scores low on Feasibility even if the problem itself is severe. A topic gets a high Timing score only when there's a real, dated catalyst behind it, not because the problem feels urgent in the abstract.

OpportunityHow large and provable the market is. A confirmed exit, real revenue at scale, or a large quantified market pulls this up. An unvalidated, purely theoretical market keeps it low.

PainHow severe the harm or frustration is for whoever has this problem, based on what's actually measured. Quantified financial or physical harm scores highest; a mild, occasional annoyance scores lowest.

FeasibilityHow buildable an actual fix is. Live, working builders with real traction push this up. A problem that requires new legislation, a platform's cooperation, or another company's incentives to change is a structural blocker, and scores low regardless of how painful the problem is.

TimingWhether there's a specific, dated catalyst happening now — new regulation, a fresh lawsuit, a platform change, new research — versus a problem that's simply been true for years with nothing new moving it.

The bands under each number (Exceptional, Strong, Moderate, Limited, Minimal for Opportunity; similar scales for the other three) are read directly off the 0-10 score — they're a label, not a separate judgment.

How status follows the scores

The status badge and the scorecard are meant to agree, so status is worked out from the scores and the evidence rather than picked separately. Timing decides whether attention is rising; the builders on the page decide whether the field is crowded.

OpenTiming is 6 or below and no crowded field. The gap is real and nobody dominates it. A low Opportunity or Feasibility score still explains why: room isn't the same as a business.

Heating UpTiming is 7 or above: a fresh rule, lawsuit, launch, or study is pulling attention to the problem right now.

CrowdedEnough live builders that the easy version is gone: four or more with Feasibility 5+, three with Feasibility 5+ and Timing 6 or below, or two with Feasibility 6+ and Timing 6 or below.

ImprovingThe system itself is closing the gap, such as prices falling or a binding rule taking effect, rather than anyone building around it.

Two calls override the formula because a hand-written read says otherwise: Amazon Reviews is marked Crowded and Luggage Claims is marked Open.

Signals and sources

A signal is a single dated, sourced entry attached to a topic. Studies, articles, competitor moves, forum threads, posts. Every signal carries its source and the date it was published, and every source stays visible on the topic page so the whole trail can be audited. Nothing is auto generated, and nothing gets added without review.

The read is the product

The aggregated signals prove the work is grounded. The read is the part that isn't a database anyone can rebuild. It is one person's interpretation of what the signals mean and where the opening is, written by hand, and it is always the thing you are actually paying for.

Update cadence

The Desk is reviewed weekly. Topics are updated when something genuinely moves, not on a schedule for the sake of it. A quiet week on a topic is information too. The date on each topic is the date of its most recent signal, so freshness is shown by the data rather than claimed.

Pattern, not a snapshot

We track how it moves, not a single snapshot. A number without a trend behind it is a press release. Every dataset on the Desk is built to show change over real, dated intervals: monthly, quarterly, year over year, so you can see whether a problem is getting worse, getting fixed, or just getting louder.