Demand First works one chain end to end: where a topic is moving in community and search data, whether crawlers and answer engines can actually use your pages, and what a real buyer does when a price is on the table.
Each step produces something checkable. Nothing later in the chain starts until the step before it has an answer, because optimising a page nobody is looking for is the most common and most expensive mistake in this category.
Reddit, YouTube, TikTok, community threads, and search-query data are collected as exports and scored on volume, velocity, engagement, recency, and corroboration. A term is only ranked when at least two independent sources agree and it clears a volume floor. Everything else is reported as insufficient. This is a manually operated pipeline — no always-on platform API is connected.
One public page is read the way a crawler reads it, then graded across search engines, answer engines, AI crawlers, app-store links, and offer clarity. Checks that cannot be decided from the response are dropped rather than guessed. You can run this yourself, for free, before anything is agreed.
Pages get rewritten, structured data gets added and validated, answer-first content gets built against the terms the sweep found, and app-store metadata is revised when a listing exists. Every claim published on your behalf is approved by you first.
Message, channel, and price propositions are tested against an exposure floor and a stop rule written before the results are visible. Price is tested as a structure, not just a number: what is committed, when, and what it takes to leave.
What moved, what did not, what it cost, what remains unknown, and an explicit continue-or-stop decision. A stop decision at day 30 is a valid and often valuable outcome, not a failure of the programme.
The programme loop needs a live surface to measure. When the product does not exist, the founding Demand Test Sprint tests one priced proposition against a defined B2B buyer before any build is funded. The rest of this page describes that method.
A landing page cannot rescue an undefined buyer, harmless problem, inaccessible market, or founder who cannot act on a positive result.
The result must change a build, client, hiring, or product investment due within roughly 30 days.
The named buyer already loses time, money, revenue, risk capacity, or operational certainty.
An owned audience, partner source, targeted channel, panel, or approved acquisition budget can reach the cohort.
The founder or studio can build or sell a focused product quickly if the evidence earns the next step.
The signed brief separates a test from post-hoc storytelling. It records the exact claim, commercial action, minimum evidence, stop conditions, and interpretation limits.
Role, company context, painful event, geography, qualification rule, recruitment source, and exclusions.
Current alternative, promise, mechanism, visible price, objections, founder-approved claims, and beta disclosure.
Meaningful action, qualified-exposure floor, budget/date cap, BUILD/REFINE/STOP rules, and insufficient-signal condition.
Test design, proposition copy, page implementation, event measurement, agreed distribution operations, monitoring, exclusions, analysis, and readout.
Context, source access where applicable, factual claim approval, prompt feedback, third-party budget, and the real follow-up promised by the CTA.
The buyer, copy, price, actions, audience, source, spend, duration, and material disclosures require approval before exposure.
Consequential actions disclose beta status before payment details or binding information are requested. The test must be ethically fulfillable.
Every denominator travels with the result. A click from an unqualified visitor, a compliment, and a paid pilot request are not interchangeable evidence.
Did the offer reach the buyer defined in the brief, through a traceable source?
Did qualified visitors engage with the painful job and promised outcome?
Did meaningful intent continue after the commercial trade-off became explicit?
Did the right prospect request a pilot, call, proposal, LOI, reservation, or another truth-labeled next step?
Did the signal persist across enough qualified exposure after QA, duplicate, bot, and source-quality exclusions?
The pre-test truth: who, what, price, source, action, budget, duration, decision rules, and known threats.
The scoped page, copy, instrumentation map, acquisition treatment, and approved disclosure used in the run.
Qualified exposure, behavior stages, price response, exclusions, cost, timing, source context, anomalies, and data-quality notes.
BUILD, REFINE, STOP, or INSUFFICIENT SIGNAL; confidence; limitations; unit-economics implication; and the next falsifiable step.
The deepest agreed action clears the rule with sufficient qualified exposure and no disqualifying source or trust issue. This earns a pilot or smallest activation loop—not a full product by default.
The problem or reach signal is credible, but the segment, promise, price, trust, or action did not clear its rule. Change one primary variable and retest.
The named buyer, source, proposition, or economics fails the predeclared floor after sufficient exposure. Preserve the losing evidence.
The cohort was not reached, the source was contaminated, execution failed, or exposure stayed below the interpretation floor. Do not relabel missing data as weak demand.
The written scope controls the promise. Difficult recruitment, slow approvals, specialized panels, or an owned-audience test can change the calendar and cost.
Usually one working week for fit, research, brief, page, measurement, QA, and founder approval.
Up to 10 business days, bounded by qualified exposure, spend, date, and safety stop rules.
Evidence review and memo after the window closes and quality exclusions are resolved.
$1,500 service fee plus separately approved third-party test inputs. No automatic renewal.
Run the scan first. It costs nothing and shows you the same starting picture a programme would.