What Is Turnaround Time in Custom Data Collection?

A buyer planning a training run needs a date, but “how long will it take” can mean the first sample, the first usable batch or the last batch of a program. Without a defined start point, a defined end point and a stated condition such as coverage depth, a quoted turnaround cannot be compared between vendors or written into a plan.

Firsthand states two turnaround figures, both measured from a signed spec to the first validated batch: a median of 48 hours where coverage is already deep, and three to four weeks where it is thin. Brief to spec takes the same week. Turnaround for later batches, delivery cadence, minimum term, SLA penalties and prices are not stated.

Turnaround time is the elapsed time between two events: a spec the buyer and provider have agreed, and data that has passed QA and been delivered. It is a property of a collection against a spec, not of a provider in the abstract, which is why a useful figure always comes with its conditions.

The start point matters as much as the end point. On this site the published clock starts at the signed spec. Discovery, the working call that turns a model failure into a problem statement, and specification, where the skill spec is written and signed, happen before it. The end point is validated data, not raw footage that still needs a separate review pass.

These terms are often used as if they were one number. Time-to-first-batch is the turnaround of the first validated batch, and it is the figure Firsthand publishes. Pilot timing is the time to a small, fast first batch against the same spec, run to prove the spec before volume. Delivery cadence is how often later batches arrive after that.

The site states that batches are rolling and delivered per validated batch, and that the pipeline is already proven after the first pilot. It does not state a cadence interval, the volume per batch or a turnaround for the second and later batches. Treat those as questions for the spec conversation, not as assumptions.

The site names five drivers. The first is coverage depth for the named population: deep coverage reaches a first batch in a median of 48 hours, and thin coverage takes three to four weeks because vetting operators is the bottleneck. Examples of deep coverage are home kitchens, retail handling and light assembly; thin examples are construction, clinical, agriculture and hospitality.

The other four are spec completeness, whether a spec pilot is run first, the number and rarity of condition cells, and the reject rate against the acceptance test. Items that fail the reject rule are recaptured rather than shipped, which protects delivery quality but adds time in proportion to how often it happens.

Most of the levers sit on the buyer side of the clock. Bring a complete spec: task, geometry, condition budget, reject rule and acceptance test written before it is signed, since a missing element causes rework and disagreement mid-collection. Keep the condition budget to the cells the model needs; more and rarer cells take longer to fill.

For a thin-coverage spec, start specification and sourcing early rather than planning against 48 hours. For an unfamiliar spec, a spec pilot catches a coverage or reject-rate problem at low volume. A buyer with a familiar spec and a well-served population often moves straight to full volume. In every case, write the quoted timeline into the contract as a term, not a verbal estimate.

On-demand collection points an already-warm pipeline at a new brief, so contributors, rigs and reviewers are in place when the spec lands. That is why the first validated batch can arrive in a median of 48 hours where coverage is already deep. Speed does not come from skipping consent, per-batch QA or anonymization: every batch clears the same reject rule and acceptance test.

Where coverage is thin, the same page states the real staffing time, three to four weeks, rather than a hopeful one. There is no minimum program, and top-ups run against the original spec and accept criteria.

The stated timeline is short. The spec conversation happens the same week. From a signed spec, the first validated batch arrives in a median of 48 hours in deep coverage and three to four weeks in thin coverage. A Spec Pilot is a small, fast first batch and is optional.

Not stated: turnaround for later batches, batch cadence interval, volume per batch, minimum term, SLA penalties, rework turnaround as a number, and any guaranteed delivery date. The figures above are medians and typical ranges, not guarantees. Ask for any of the unstated items in the spec conversation.

What is turnaround time in custom data collection?

The elapsed time from an agreed, signed spec to delivered, QA’d data. It depends on the spec, the coverage already in place and the QA outcome, so it should always be quoted with its conditions.

How fast is the first batch with Firsthand?

A median of 48 hours from a signed spec where coverage is already deep, and three to four weeks where contributors must be recruited first. Brief to spec takes the same week.

Does the turnaround include writing the spec?

No. Discovery and specification happen before the published clock starts. The figures measure from a signed spec to the first validated batch.

What is the difference between turnaround and delivery cadence?

Turnaround is the time to deliver validated data against a spec. Cadence is how often later batches arrive. The site states rolling delivery per validated batch but not an interval.

Why do some environments take three to four weeks?

Thin-coverage environments such as construction, clinical, agriculture and hospitality need operators vetted against the spec before capture can start. That vetting is the bottleneck.

Does a faster turnaround mean QA is skipped?

No. Every batch, fast or slow, clears the same reject rule and acceptance test, and consent and anonymization still apply. Speed comes from an already-screened contributor pool.

Is turnaround for later batches or an ongoing program stated?

Not stated. The site gives no turnaround for second and later batches, no cadence interval, no volume per batch and no minimum term. Ask for them in the spec conversation.

Can I get a delivery date written into the contract?

The site advises treating a quoted timeline as a contract term rather than a verbal estimate. It states no guaranteed dates or SLA penalties, so request a custom dataset and agree them in the spec.