Investment memorandum / 01Private equity modeling

Industry modeling for private equity

See what changes
the investment case.

We build executable models of an industry’s major players and their interactions, then test how changes across that system could affect a target company’s revenue, margins, and cash flow.

An industry model. A company in focus.
Relationships, assumptions, and an investment readout.

An industry event. A company consequence.

Start outside the company.
Follow the consequences in.

A competitor adds capacity. Retailers transmit price pressure. Customers shift purchases. Follow those interactions into ConsumerCo’s revenue and margins, then inspect every rule along the way.

A consumer industry, in miniatureSynthetic relationships · one annual period · not a forecast
01 / Choose a scenario
02 / Follow the relationships Select any participant to inspect
A connected slice of an industry.
Counterparties have declared roles;
only the target has a full P&L here.
03 / Read the target economics

ConsumerCo / Capacity expansion

Illustrative annual EBITDA$16.85M17.6% EBITDA margin
$3.15M below starting case
Revenue
$95.65M
Variable costs
$58.80M
Fixed costs
$20M
Net response benefit
$0M

Cash flow would require working capital, capex, tax, and financing. They are outside this demo.

Same industry assumptions. Three company outcomes.
Starting case$20M
Capacity expansion$16.85M
Operating response$17.60M

The response adds $0.75M ($1M savings less $0.25M recurring cost); $2.40M of downside remains.

Model boundary: one competitor, aggregated retailers, customers, and suppliers around one target. Fixed category demand and input unit costs; no market feedback, new entrants, capital providers, or infrastructure constraints. Parameters are chosen to expose the logic, not calibrated to an actual industry.

Questions to take back to the evidence

  • How does added capacity change competitor offers?
  • How do retailers and customers transmit that change?
  • Which participants or feedback loops could change the result?

This small example demonstrates explicit propagation, not a predictive industry engine. An engagement would map the major relevant actors, establish evidence for their interactions, and define what remains outside the model.

Existing research model

The industry is the unit of analysis.

In our existing AI-industry research model, named participants are connected through modeled commercial and financing relationships. Inspect the industry network, trace synthetic settlements between participants, and distinguish those changes from reference balances.

01

Trace a modeled relationship to exact settlements.

The relationship view and settlement table make the modeled participants, amounts, and dates available for review.

Actual research UI: modeled relationships and recorded synthetic settlement rowsOpen full money flows capture
02

Keep reported references separate from modeled changes.

The cash summary distinguishes reference cash from the changes produced within the modeled scenario.

Actual research UI: cash summary separating reference balances from modeled cash changesOpen full cash summary

Actual repository UI; synthetic research scenario. Company names identify modeled participants, not clients. These captures demonstrate inspectability, not forecast accuracy or a private-equity customer outcome.

One model. Two consequential decisions.

Keep underwriting connected to ownership.

Before you invest

What must be true for this deal to work?

Test the assumptions behind growth, margins, and cash generation. Identify the evidence that would change your view.

How much pricing pressure can the business absorb before the investment case needs to change?
During ownership

Which actions could create value?

Translate operating initiatives into assumptions about adoption, implementation cost, and cash savings.

If AI reduces the work required, how much becomes lower spending rather than capacity?

Powered by prompt-driven development

Read the assumptions.
Challenge the logic.

Methodology should be understandable to the people making the decision, with checks that distinguish arithmetic from a claim about the future.

01Evidence

Source, date, units, limits.

02Assumptions

Readable and challengeable.

03Calculations

Traceable to the output.

04Review

Limits stated beside results.