Coleman LLC · Colorado Springs

Joseph Coleman

I get handed the things that don't exist yet.
Staying with us?  Start here →
Scroll
Video: NASA/JPL-Caltech
Act I

The Story

Spacecraft don't forgive optimistic math. Neither do marketplaces.

Chapter 01 — Pasadena
It starts at JPL

I owned the number a mission got committed on

$350M owned

Mars 2020 — the rover in the video above — ran roughly $2B through build and closer to $3B once operations were counted. I owned the mission systems area, about $350M of it, and the scope grew from there. I held the work breakdown structure for Europa alongside it and ran JPL/SpaceX contracts ahead of schedule.

The analysis went up through the program to NASA's head of science. At that altitude nobody is checking your arithmetic — they are checking whether the decision it points to holds. Two awards came out of those three years; the durable thing was learning to build a number someone can commit a mission to.

Chapter 02 — Dog Years
Sixteen months at a Caltech spinoff

I sold the thing I had just finished building

75% of company ARR

Startup time runs about seven to one — these were the dog years. I wore whatever hat the week needed while building the B2B motion from nothing: the pricing, the pitch, the process, the handoffs. Then I sold it into Fortune 500 companies and global banks, where a signature meant carrying five to seven stakeholders across procurement, security, IT and the C-suite. One of five reps; three quarters of the company's revenue; SaaS up 350% year over year.

Lenovo Royal Bank of Scotland Royal Bank of Canada Facebook Collins Aerospace

I have since spent years on the buying side of that same table, approving vendors for a marketplace rather than selling into one. Watching how a purchase actually gets signed off taught me more about enterprise selling than any deal I closed.

Chapter 03 — More From Less
Before the launch

The audience stopped growing. The revenue didn't.

3× revenue

Running the lending marketplace meant growing a business whose audience wasn't growing with it. Waiting on more traffic was never going to work, so I went after more value from the audience already there — taking lender revenue from $420K to $1.26M a year.

Category expansionOpened the marketplace past hard money into DSCR and conventional-alternative products.
Lead bundlesPackaged unsold inventory into custom agreements — dead stock became billable revenue.
Direct sellingWorked the pipeline personally instead of waiting on inbound to fill it.
Chapter 04 — Zero to One
Building from nothing

I scoped it before it had a name

$0 → $3M ARR

An entirely new marketplace product, built from the ground up. I scoped the legal, technical, and operational groundwork before a line of it existed, shipped an MVP, then iterated hard — designing the go-to-market, the sales motion, and the partner onboarding as it grew. Eight months after launch it was running at $3M in annual recurring revenue. The build itself gets its own section below.

Chapter 05 — Scale
Building the second one

Twelve people, hired to sell something that didn't exist yet

$800K+ / yr

Investor Concierge started as a test and now runs past $800K a year. I recruited and trained the twelve-person contractor team behind it, across the US and overseas, and built the lead routing and scoring that decides who gets called and when — Claude, the OpenAI API and Zapier doing triage that would otherwise cost more headcount than the line could carry. Two further marketplace verticals launched underneath it, inside an $8M unit.

Chapter 06 — The New Channel
Currently shipping

Every audience has a ceiling. I went around ours.

1,000+ leads / mo

Every marketplace eventually hits the ceiling of its own audience. Powered By BiggerPockets breaks through it — a partner channel that sources investor demand from across the industry instead of relying on site traffic alone. I scoped it as a staged test, proving the economics on a small footprint before asking anyone to fund the real thing, then scaled it into a channel producing thousands of leads a month.

Act II

Projects

Anyone can quote a number. These are the machines that made them.

Lender Finder · marketplace product

It decides which lenders get to see your deal

An investor describes the deal — where the property is, what kind of loan they need, what the numbers look like — and the marketplace matches them with lenders who actually operate in that market and write that product: hard money, DSCR, non-owner-occupied conventional, and conventional-alternative. Lenders pay for the matches they receive.

Property location
Colorado Springs, CO
Loan type
DSCRHard moneyConventional
Purchase price
$385,000
Find lenders
Matched lenders
Lender ADSCR · lends statewideMatch
Lender BDSCR · 30-yr fixedMatch
Lender CDSCR · investor focusedMatch
Illustrative reconstruction of the matching flow — not the live product.

The part nobody sees is the pricing.

I built the pricing model from scratch in a spreadsheet that is still in use today. It starts with the average property price in a given market and works forward: what a typical loan looks like there, what that loan is worth to the lender writing it, and therefore what a single lead in that market can honestly be priced at. Then it runs the other direction — how many lenders a market can support before the leads spread too thin and everyone churns — to land on the supply each market actually needs. The output isn't a price list. It's the supply and pricing combination that optimizes revenue.

Run it yourself.

01Average local property price$385,000
02Typical loan at 75% LTV$288,750
03Worth to the lender, two points$5,775
04Defensible price per lead$72
05Lenders this market supports15
06Monthly revenue at that supply$8,663 / mo

8 lenders, 15.0 leads each. Healthy — enough volume for everyone to renew.

Assumes 75% LTV, two points to the lender, a 5% lead-to-close rate, and a floor of eight leads per lender per month before renewals stop. Change the assumptions and the shape holds: revenue is set by leads and price, never by adding lenders.

Mars Program Office · NASA JPL

Rebuilding the reporting around decisions

I came into the Mars program office and inherited a reporting apparatus that had grown by accretion — reports produced largely because they had always been produced. I rebuilt it around one question: which of these actually changes a decision? What survived that test got sharper and faster. What didn't got retired.

Fewer reports, better used. It remains the clearest lesson I've had in the difference between measuring something and managing it.

Reports produced Reports that changed a decision
Illustrative — the ratio, not a literal count.
Rotational Program · NASA JPL

Mapping how the whole machine ticked and tied

Rotating through functions gave me an unusual vantage point: I could see how the organization's systems actually connected, rather than how the org chart claimed they did. I built a visualization of that data architecture — where numbers originated, what transformed them along the way, and where they finally surfaced to be decided on — and socialized it across the organization.

Understanding a machine end to end is what makes it possible to fix causes instead of symptoms. Nearly everything I've built since has started with the same exercise.

Sources Transforms Decisions
Schematic of the pattern — sources, transformations, and the decisions they feed.
Act III

Principles

How I operate, written down so it can be argued with.

Optimize for the long arc.

I want a new bet to pay itself back in about six months — and then I optimize the next five to twenty years. Short-term thinking is a tax on compounding: it buys a quarter and sells a decade. The discipline is holding both horizons at once. Prove it quickly, then build it to last.

Fix the machine, not the output.

A bad number isn't the problem. It's the output of a machine that produced it. Chase the number and you get a patch that holds until next quarter; trace it to its cause and you get a fix that holds permanently. Almost every recurring problem I've met turned out to be a design flaw wearing a costume.

Data has to earn its place.

Before a report, a dashboard, or a recurring meeting is allowed to exist, it has to answer one question: what decision changes because of this? If the honest answer is none, it's theater — and theater is expensive, because it costs attention you can't get back. I've retired more reporting than I've built, and every organization was better for it.

Give authority, not just responsibility.

Responsibility without authority produces a team that waits. I hand people the decision rather than the task, and hold them to the outcome rather than the method. It's uncomfortable at first and it is the only thing that scales past the reach of one person.

Make small bets before big commitments.

Before asking anyone to fund the real thing, prove the economics on a footprint small enough to be wrong cheaply. Crawl, then walk, then run. Most of what I've launched began as a test deliberately designed to fail fast if it was going to fail at all — which is what made the eventual commitment easy to defend.

Act IV

Passions

Interest rates, long train rides, and twenty years behind a drum kit.

Economics

Rates, and why they explain almost everything

My degree is in business economics, and the thing that caught me was interest rates — I spent college pulling apart how they move and what moves with them. Macro and micro have been a running obsession ever since, and they turn out to be relentlessly practical. Pricing a lending marketplace is really just microeconomics with a deadline: supply, demand, and the cost of money explain more of the business world than most business frameworks manage to.

Short end Long end
A curve, drawn for the pleasure of it.
Passion project · 2019

Overland on the Silk Road

The idea landed in 2012, on a flight home from China, when it struck me that I had no real notion of what filled the enormous space between China and Europe. It took seven years of saving and planning to answer that — most of them spent at a desk in Pasadena, daydreaming through Jim Rogers and Paul Theroux between budget cycles. On the 1st of November, 2019, I finally left.

Tokyo first, then Hong Kong, and from there by high-speed rail across China through Wuhan to Xi'an — the most historically astonishing place I have ever stood. On to Urumqi, out where the map stops being familiar to most Americans. Over the border to Almaty, down to Shymkent by Kazakh train, across the Caspian to Baku, landing on Thanksgiving night to a planeload of applause. Then Georgia.

I travelled it the way I'd read about it — through business, macroeconomics, geopolitics and music — and wrote the whole way, mostly from trains.

There are no things, only process.— from the first dispatch, 2019
Tokyo Hong Kong Wuhan Xi'an Urumqi Almaty Shymkent Baku Georgia N
Stylized route plot — east to west, China to Europe.
Read the dispatches →
The rest of it

Off the clock

25+
countries, and counting
20 yrs
behind a drum kit — Berklee-trained
Garden of the Gods
hosting guests minutes from the red rocks in Colorado Springs

Contact me

Partnerships, business development, or a stay near the red rocks.

joseph@colemangroup.co