ModvenRun an Opportunity ID →

Strategy & innovation for the old economy

The growth partnerfor the old economy.

We partner with private equity firms and mid-market distributors, manufacturers, and service businesses to unlock the value already sitting in their records, and we build what captures it.

Opportunity ID: $500, ten business days, credited against any build.

Modven POV

Old models are where alpha hides. Because nobody hunts there.

The old economy is the prime opportunity. Manufacturing, distribution, transportation, industrial services. Businesses that have recorded their own behavior for decades while the operating model stayed the same. The records quietly disagree with the assumptions. That gap is the alpha.

Until now, testing an assumption meant a study, a budget, and a year. Agents have cut that to weeks and a fraction of the spend. The assumptions haven't been tested. The cost of testing them has collapsed.

The result is asymmetric. Small tests can expose outsized value.

The domain changes. The arithmetic doesn't.

Signals — extraction run
extract.pyassumptions.pysources.sqlMODVEN / SIGNALS
1# Read what the business already records. Write nothing back.
2from modven.sources import Export, Ledger
3from modven.models import SignalExtractor, AssumptionMap
4from modven.alpha import size, grade, StopCondition
5
6sources = Export.load("client_2024-2026.zip") # structured export, nothing connected
7tables = ["quotes", "invoice_lines", "shipments", "contracts", "returns"]
8
9extractor = SignalExtractor(window="24M", grain="line")
10ledger = Ledger()
11assumed = AssumptionMap.from_operating_model(sources)
12
13for t in tables:
14 recorded = extractor.read(t) # what actually happened
15 expected = assumed.project(t) # what the model says should
16 gap = extractor.diverge(recorded, expected)
17
18 for signal in gap.signals(min_strength=0.6):
19 alpha = size(signal, horizon="90d")
20 alpha.stop = StopCondition.write(signal) # before anyone builds
21 ledger.add(alpha, grade(alpha))
› export 24M window · 5 tables · 2.3M rows
› quotes 2.1M lines · assumption: price follows list · divergence 0.71
› shipments 184K loads · assumption: freight is billed · divergence 0.83
› contracts · assumption: resets track the index · divergence 0.64
› panel buyers ×40 · operators ×20 · skeptics ×20 · 3 assumptions failed
› 3 signals above threshold · stop conditions written · next: assumptions.py

On record

We analyzed more than twenty innovation projects, from USAA, Steelcase, and Delta Faucet to a half-dozen ventures we’ve launched. The same opportunity DNA showed up in every sector.

Two of them, an innovation project and a venture.

KISS
A new category of buyer

The ask · What we delivered

Increase the value of the band. Brand value rose 57% in 45 days. Our idea positioned KISS with a new category of buyer, and the IP sold for $300M+.

brand value

IP sale

from ask to result

United MileagePlus
The market we created

The ask · What we delivered

Burn down a mile liability measured in the tens of billions. Program contract value. A live auction where miles-rich customers competed for scarce experiences. The business was acquired.

program contract value

Live

auction for scarce experiences

Acquired

business exit

Every engagement runs on

Opportunity DNA™

Assumption

Name it.

It starts with what the team believes.

Every hunch gets written down with a number and a range, before anyone opens the data.

Signal

Find it.

Then it reads what the records say.

Years of quotes, orders and invoices test every hunch, and surface the gaps nobody raised.

Alpha

Size it.

It sizes the gap and why it survived.

A dollar range from your own data, and the incentive that kept anyone from closing it.

Holdout

Prove it.

It has to hold up twice.

Found in one stretch of records, then confirmed in a stretch nobody looked at.

Mechanism

Build it.

Then we build what captures it.

A startup, a product or a workflow. About ten weeks, with the stop condition written first.

What ships

What ships is a working thing, never a PowerPoint. It usually takes one of six forms.

Pricing

A pricing engine. Quotes priced from your own win history instead of a cost-plus sheet. Your estimator sees a second number and decides.

Operations

A workflow run by agents. The process that eats a workday, like quoting, order entry, or reconciliation, run by agents and checked by your people.

Capital

A forecast. Demand and inventory forecasts that free working capital and stop the expensive guesses.

Revenue

A new product line. Something your customers already need, sold by the team they already trust.

Market

A new market or buyer. A market for something you were giving away, or a category of buyer nobody had pitched.

Venture

A new company. When the mechanism is worth more to your whole industry, it spins out. You hold the equity; we bring the investors.

It runs

Live in your operation, on real records and real decisions.

It’s yours

Code, models, and data structures. No license, no dependency.

It’s measured

Against the size of prize written in the OID, every week.

It can stop

The stop condition is agreed before the build starts.

Design, launch, and growth ship with every build. Hire any of them on their own.

Cover of Where Value Hides, a Modven field guide

The field guide

Where Value Hides lays out our whole method in fifteen pages. Every number in it is sourced.

Why this market, how the opportunity DNA was found, how an Opportunity ID proves value before anyone builds, a primer on graph engineering, and the decision rules behind robust portfolios. See what’s inside.

We’ll send nothing you didn’t ask for. Justin may follow up once, personally.

Built with

Anthropic Claude
Figma
Databricks
Snowflake
Vercel
Python

The Operating System

The DNA we found across those ventures became our operating system. It has four moves.

01

signals

Every business continuously records its own behavior. Quotes. Invoices. Contracts. Redemptions. Routing. When an operating model goes unquestioned for decades, those records quietly disagree with the assumptions. We read them.

02

alphas

A signal becomes an alpha candidate when it can be sized. It names a specific gap, puts a dollar range and an evidence profile on it, and states the test that would prove it wrong. It becomes a proven alpha only when that test is run.

03

shipping

We don't hand over findings. We build the mechanism that captures the alpha, whether that is a pricing engine, a market, a buyer, or a product, and it is working inside your operation within weeks.

04

portfolio

For owners and investors who want more than one bet. A managed portfolio of opportunities, each gated, each killed or scaled on evidence, run like a production line.

Who We Serve

We work with old-economy businesses and the groups that own them.

Graph engineering

The mechanisms are built as graphs. Agents run them.

Every engagement leaves a working system, not a deck. That is the only reason these timelines are possible.

Recent project

agents running concurrently

days

to a signed Opportunity ID

in value identified

See how we build →

Services

Every engagement begins with an ID.

01 · $500 · 10 business days

Opportunity ID

One export from one business comes back with the three to five opportunities worth testing, each with a dollar range, an evidence profile, and a written stop condition. If the evidence isn’t there, we say so. The $500 is credited against a build.

Run an Opportunity ID →

02 · 10 weeks per build

Build Engagements

Pick one opportunity or several. We build the mechanism that captures each one, live in your operation. Ten weeks per build. Builds can run at the same time. Fixed price, stop condition in writing.

See how it works →

03 · Sponsors and holding companies

Portfolio Alpha

For sponsors and holding companies. We run the operating system across every company you own: an OID per company, ranked by size of prize, killed or scaled on evidence. We can also run it during diligence, before you close.

See the model →

Start here

See what’s already in your records. Tell us your hunch. Ten days later it comes back with numbers on it.

10 business days · $500, credited against any build