From a six-month working group to defined challenges in under an hour

Client

Date

$20B+ global workforce group's social innovation lab

09/2026

The challenge

A global workforce group's social innovation lab chooses what to work on by scanning labour markets for signals and turning them into defined challenges. That scan ran as a six-month working group on a largely manual process, with no codified method anyone else could pick up and repeat.

The solution

FifthRow deployed a signal-to-challenge engine. A region, a population group or a focus area goes in; sourced signals, prioritised trends and defined challenge statements come out, with the lab's own drivers, filters and theory of change encoded in the method rather than held in people's heads.

77,974

analyst-hours equivalent, 72 FTEs

12,757

source-verified research questions

Challenge cycle timeUnder an hour, signal to challenge

Standing coverageSix scans, every two weeks, unattended

Encoded methodology15 drivers and 5 filters, run as a system

Research questions12,757 answered against sourced evidence

Analyst-hours equivalent77,974 hours saved; equivalent to 72 FTEs

01 — SIGNAL-TO-CHALLENGE ENGINE

The lab's scan asks one question in three moves: what is happening, what of it matters, and what should we work on. The method existed as a published document and as a working group's practice. It did not exist as anything that could be run.

Core Activities

Encoded the lab's own published scan method – its drivers, its filters, its theory of change – as the logic the engine runs on
Runs three linked stages: sourced signal detection, trend surfacing and prioritisation, then problem framing and challenge definition
Takes a region, a population group or a focus area as an input, so the same method points anywhere without being rebuilt
Searches local-language sources first and English second, so evidence from the markets under study is not lost before it is read
Tags weak signals rather than discarding them, and reports what was excluded alongside what was kept
Pulls institutional, academic and media evidence together instead of leaning on one class of source

The Results

Each cycle returns a defined challenge: the population it affects, the geography, the structural barrier underneath it, why it matters now, and how well it fits the lab's own model – with the evidence behind every claim traceable to where it came from.

~690

analyst-hours equivalent per challenge cycle

02 — SCHEDULED SCAN CYCLES

Coverage that depends on someone having a free fortnight is not coverage. Six standing scans now run every two weeks across the lab's priority regions and population groups, and arrive as a summary in the inbox.

Core Activities

Set six scans on a fixed fortnightly cadence across the priority regions and population groups
Delivers an executive summary to the inbox with a link through to the full evidence, so nothing has to be started by hand
Reports the delta each cycle – what changed since the last run, rather than the whole picture again
Accumulates context run over run, so a later cycle starts from what the earlier ones established
Cross-triangulates outputs across geographies to show what holds everywhere and where the pattern breaks

The Results

A fortnightly read on every priority region and population group, arriving whether or not anyone had time to ask for it, and getting richer each cycle instead of starting from zero.

43,116

analyst-hours equivalent, delivered unattended

03 — MULTI-COUNTRY EVIDENCE DIAGNOSTICS

Some questions need more than a scan. A single diagnostic programme now runs a nine-stage, forty-one-step evidence process across a whole region in about three and a half hours.

Core Activities

Fixes a shared ontology and query architecture first, so every country is searched against the same definitions
Harvests evidence in parallel lanes: country by country, regional institutions, and the employer and platform side
Runs a dedicated disconfirming lane that goes looking for evidence against the picture forming
Synthesises across social, technological, economic, environmental, political and legal lenses
Scores and selects what survives, with a rescue lane for evidence that nearly cleared the bar
Closes with an adversarial review testing every selection for soundness, duplication, currency and coverage, then a final package audit

The Results

A scored evidence base on what shapes labour-market access across a whole region, with the reasoning for every inclusion and every exclusion visible, produced in an afternoon rather than a research quarter.

11,765

sources analysed in one diagnostic run

04 — IN-HOUSE SYSTEM BUILDING

A research lead joined the lab in the summer and was given platform access. Three weeks after a first onboarding call, the region-wide diagnostic above had been built in-house, with no external build effort behind it.

Core Activities

Onboarded the lab's research lead on editing templates, variables and datasets
Handed the engine over as an editable template rather than a fixed deliverable
The lab's own researcher assembled a nine-stage pipeline end to end, and a separate adversarial review system beside it
Ran joint working sessions to tighten how each stage feeds the next as the pipeline scaled
Kept every system in the lab's own catalogue, so a repeat study starts from a working method rather than a blank brief

The Results

The lab builds and changes its own research systems. New coverage no longer waits on an outside build queue, and the methodology stays in the lab's hands rather than in a supplier's.

3 weeks

from first onboarding call to a self-built pipeline

What it proved

Ninety complete signal-to-challenge cycles have run across Southeast Asia, sub-Saharan Africa, Latin America, Europe and North America, answering 12,757 source-verified research questions. Sixty-one of those cycles ran to a fortnightly schedule with nobody starting them. One region-wide diagnostic ran forty-one stages of evidence work in a single afternoon. Three people did all of it.

The lab now holds its own scanning capability: a published method turned into a system it can run, edit and point at any region, population group or focus area it chooses. Looking at one more country is no longer a budget line and a working group's calendar. Across 368 runs the platform records 77,974 analyst-hours saved, the equivalent of 72 full-time analysts over the period, produced by three people.

Results

Metric	Benchmark	Result	Improvement
Challenge cycle time	Six-month working-group scan phase	Under an hour, signal to challenge	Months to under an hour
Scan cadence	Time-bound, once per programme cycle	Six standing scans every two weeks	Continuous coverage, started by no one
Regional evidence depth	Manual desk research and expert calls	41 stages of evidence work in an afternoon	Region-wide depth inside one working day
Methodology	Published document, no runnable form	15 drivers and 5 filters encoded	A method that runs, not one that is read
Analyst-hours saved	Three people doing the work manually	77,974 hours across 368 runs	Equivalent of adding 72 full-time analysts over the period

ENCODED METHODOLOGY

A published research method runs as a system, so the standard behind a challenge statement is the same one every time it is produced.

LOCAL-LANGUAGE SOURCING

Evidence is gathered in each market's own language before English, so signals from the places being studied survive to the analysis.

WEAK-SIGNAL TAGGING

Faint signals are labelled and kept rather than filtered away, and what was excluded is reported next to what was kept.

DISCONFIRMING EVIDENCE

A dedicated lane searches for evidence against the picture forming, so a conclusion has to survive a hostile read before it ships.

SELF-BUILT PIPELINES

Researchers inside the lab assemble multi-stage systems themselves, so new coverage arrives without waiting on outside build effort.

STANDING CADENCE

Priority regions and population groups are re-read on a fixed schedule, so coverage stops depending on who had time that month.

Recent case studies

See more proof

Finding the next high-growth US metro in five days

Client

$5B+ Global Design and Engineering Firm

Date

08/2026

5 days

naming the next growth market

Making consultants 3–5× more productive on the research behind their work

Client

$30B+ Global IT Services and Consulting Group

Date

08/2026

3–5×

productivity on consulting delivery

Matter More

Let's Talk