149 research studies, averaging six weeks of analyst work each

Client

Date

$10B+ global electrical equipment manufacturer

09/2026

The challenge

A $10B+ global electrical equipment manufacturer runs innovation, product and venture teams across three continents, each deciding where to put money next. The evidence underneath those decisions was uneven. On one live product question the teams held competitor analysis and the opinion of their own sales force, and no customer evidence at all.

The solution

FifthRow deployed a research engine the teams configured for themselves. A market space, a customer persona or an investment target goes in; a full multi-agent study comes out, carrying its own research questions and the sources behind each answer, in the same structure every time.

36,422

analyst-hours equivalent, 58 FTEs

9,030

source-verified research questions answered

Depth per study244 analyst-hours of work, on average

Evidence base9,030 source-verified research questions answered

Investment thesis cadenceA seven-part thesis every Monday, unasked

Customer-need coverage32 reads on one method, 24 in a single day

Analyst-hours equivalent36,422 hours saved; equivalent to 58 FTEs

01 — OPPORTUNITY EXPLORATION

Before anyone commits engineering time, somebody has to say whether a space is worth entering at all. This is the read that answers it: the need, the technology, the players, the intellectual property and the size, in one pass.

Core Activities

Mapped the need, the technology and the players in a candidate space
Sized the addressable market before anyone committed engineering time
Read the intellectual-property landscape around the opportunity alongside the market
Scanned the wider forces that would move the opportunity either way
Returned every space in the same structure, so spaces can be ranked

The Results

A sized, evidenced read per candidate space that arrives in the same shape as the last one, so a portfolio conversation compares spaces against each other rather than weighing whichever case its sponsor argued best.

3,123

source-verified research questions answered

02 — CUSTOMER-NEED EVIDENCE

A product question with competitor data and sales opinion behind it but no customer evidence is still a guess. This stream reads unmet needs and jobs to be done directly, persona by persona, on one method.

Core Activities

Read unmet needs and jobs to be done, one customer persona at a time
Held the same method across every persona, so needs can be compared
Generated the interview questions the team would take into live research
Turned needs into a stated value proposition rather than a list of findings
Returned two full product scopes' worth of reads inside a single day

The Results

Persona-level evidence on what the market actually wants from a product, traceable back to the sources behind each need, on a question the business had previously answered only with competitor data and internal opinion.

customer-need reads returned in one day

03 — COMPETITIVE POSITION

Competitive work at group level tells a product team almost nothing. This stream builds the competitor set for a specific product and places that product against it on the terms customers actually choose on.

Core Activities

Identified the competitor set for a product rather than for the company
Compared offers feature by feature against the product being planned
Read market share and where each competitor is gaining or losing it
Placed the product on a strategy canvas against the whole set
Kept a standing competitor watch running on the same method

The Results

A product-level competitive picture with the gaps named, so a roadmap decision rests on where the set is weak rather than on a general sense of who the big competitors are.

2,096

analyst-hours equivalent, competitive positioning

04 — INVESTMENT THESES

The venture team's thesis was a document written from scratch each time. It became a structure: seven parts, the same seven every time, so two opportunities can be set beside each other.

Core Activities

Encoded the thesis in seven parts: landscape, rationale, criteria, risk, fit, exit, roadmap
Ran it against each opportunity, so theses arrive in the same shape
Answered around 200 source-verified research questions inside a single thesis
Put the thesis on a weekly schedule, delivered without anyone starting it
Returned it whole, rather than as seven separately commissioned analyses

The Results

A committee-ready thesis per opportunity, built to one standard and refreshed on a set day each week, which turns a thesis from something a team writes into something the team reads.

One hour

from launch to a complete investment thesis

What it proved

Eight people across innovation, product and venture teams ran 149 studies in under four months. Together they carry 471 analysis steps, 14,912 sources and 9,030 source-verified research questions, an average of 244 analyst-hours of work behind each study. The heaviest of them, the venture team's own seven-part investment thesis, returns in about an hour carrying around 200 checked questions, and arrived unasked on six consecutive Mondays.

What the teams were left holding is methods rather than documents: a thesis structure that re-runs against the next opportunity, a customer-need method that reads one persona the same way as the next, and a competitive view built to the same standard for every product. A product question that had no customer evidence behind it at all now has two dozen persona-level reads. Across those 149 studies the platform records 36,422 analyst-hours saved, the equivalent of 58 full-time analysts.

Results

Metric	Benchmark	Result	Improvement
Customer evidence on a live product question	Competitor data and sales opinion, no customer research	24 persona-level reads in a single day	A question with no evidence now answered
Depth behind one study	Desk research sized to the hours available	244 analyst-hours of work per study	Depth no longer set by capacity
Investment thesis production	Written from scratch, roughly 90% by hand	Seven-part thesis in about an hour	Thesis is run rather than rewritten
Research method	Held in each team's own heads	Encoded once, re-run across teams	One standard on every opportunity
Analyst-hours saved	Eight people researching alongside their day jobs	36,422 hours across 149 studies	Equivalent of adding 58 full-time analysts

DEPTH ON DEMAND

One launch returns a multi-agent study carrying hundreds of source-verified questions, so a funding decision rests on more than a summary.

ENCODED INVESTMENT METHOD

A thesis structure is built once and re-run, so every opportunity is judged against the same seven questions rather than whichever ones its sponsor chose.

SCHEDULED THESES

The thesis re-runs on a set day, so a portfolio view refreshes without anyone having to start it or remember it is due.

PERSONA-LEVEL EVIDENCE

Customer need is read one persona at a time on a single method, so unmet needs can be compared across an audience instead of argued.

SELF-BUILT SYSTEMS

Teams configure their own research systems, so the method matches how each function already works rather than how a vendor assumed it did.

TRACEABLE ANSWERS

Every finding arrives attached to the sources behind it, so a planning number can be cited rather than defended.

Recent case studies

See more proof

Cutting a four-month research cycle to six weeks across five validation workstreams

Client

$25B+ US energy & home-services company

Date

08/2026

6 days

intake brief to ranked report

From a weekly newsletter to a self-running innovation radar

Client

$70B+ global energy utility

Date

08/2026

284,680

analyst-hours equivalent, 71 FTEs

Matter More

Let's Talk