A financial model is a decision tool. It is worth what the decision it supports is worth, and nothing more
Most models fail for the same three reasons. They answer a question nobody asked. They are built so that only the person who built them can change an assumption. Or they carry a rule that stopped being true. We build models that a board, a lender or an investor can interrogate, with the assumptions on the surface where they can be argued about, and with every regulatory input checked against its source rather than carried forward from a template.
- Corporate, project, investment, fundraising and industry models, built in Excel and handed over with the logic documented rather than locked.
- Every regulatory figure on these pages carries a source and a date it was checked. Where we could not read a source, the page says so rather than guessing.
- We are explicit about what a model is not. It is not a valuation report, it is not assurance, and no model output is a certification of anything.
Five kinds of model, and the decisions each one exists to serve
Three statement model
The operating core: profit and loss, balance sheet and cash flow, tied together so that a change in one flows through the other two. Used for budgeting, rolling forecasts, covenant headroom and board reporting.
Project and infrastructure model
Built around a debt schedule rather than a profit line. Feasibility, project finance, public private partnership structures, cost benefit analysis and the sizing of debt against cover ratios.
Investment and transaction model
Discounted cash flow, comparable company and precedent transaction analysis, merger models and leveraged buyout structures. The analysis behind a price, which is a different thing from a valuation report.
Investor and cap table model
The financials behind a raise: unit economics, burn and runway, the cap table, convertible instruments, option pool and the dilution arithmetic across rounds.
Industry specific models
Real estate and property development, banking and financial services, healthcare and pharmaceuticals, manufacturing and supply chain, technology and e-commerce. The shape of the model follows the shape of the cash.
Model review and rebuild
An existing model, tested rather than replaced. Formula integrity, circularity, hardcoded values inside formulas, broken links, and whether the assumptions still match the rules in force.
Five guides, in the order the work usually runs
Each guide stands on its own. If you only read one, read the first: getting the model type wrong is the mistake that cannot be corrected later by better arithmetic.
Choosing the Right Financial Model for the Decision You Face
Which model answers which question, what makes a model decision grade rather than merely arithmetically correct, and the three questions a reviewer asks before reading a single number.
Three Statement Models, Forecasting and Budgeting
The operating core: how the three statements tie, how to choose drivers that a business actually manages, what the accounting basis does to a forecast, and the integrity checks that catch most errors in under ten minutes.
Scenario, Sensitivity and Risk Analysis
Turning one number into a defensible range: scenario architecture, what sensitivity does and does not tell you, covenant and debt service headroom, and how to put a range in front of a board without it collapsing back into a point estimate.
Fundraising Models: Investor Financials, Cap Table and Dilution
The model as an investor reads it: burn and runway, the cap table, convertible instruments under Indian exchange control, employee option pools and the tax that attaches to them.
Project Finance and Infrastructure Models
A different model shape: bankability rather than profitability, the Reserve Bank's project finance regime as it stands since October 2025, debt sizing and cover ratios, and where viability gap funding fits.
Five steps, and a handover you can actually use
Scope the decision
We start with the decision, not the spreadsheet. Who will rely on this, what will they do differently depending on the answer, and what would make them reject it.
Agree the architecture
Inputs, calculations and outputs kept separate. One assumptions sheet. No input buried inside a formula. Agreed before anything is built, because retrofitting structure into a finished model is a rebuild.
Build and check
The model is built in blocks, each one checked as it lands. Balance sheet ties, cash flow reconciles, no circular references left running, every external figure traced to a source.
Test it
Scenarios and sensitivities run against the base case, and a deliberate attempt to break the model with values at the edges of what is plausible.
Hand it over
The model, an assumptions log, a short written note on what the model does and does not do, and a walkthrough. Unlocked, so you can keep using it.
What a model is not, said plainly
A model is not a valuation report. A valuation report is a signed professional deliverable, and Indian law is specific about who may sign one and in what circumstances one is required. We do that work, and we describe it separately on our valuation pages. Building a discounted cash flow inside a model does not make the output a valuation.
A model is not assurance. Nothing in a model is audited, certified or reviewed in the technical sense those words carry in professional practice. A forecast is not capable of being audited, because it describes something that has not happened.
A model is not a substitute for judgement about the assumptions. We will tell you when an assumption looks unsupportable, and we will build the model so that anyone can see what the answer does when that assumption changes. The assumption remains yours.
Three things we do that are not standard
Regulatory inputs are checked, not carried forward
Tax rates, thresholds, cover ratios and statutory dates inside a model age silently. Every one we use is traced to the instrument it comes from and dated. Where the source could not be read, the model says so instead of assuming.
Both tax regimes are labelled
India is running two income tax statutes at once. Financial year 2025-26 is governed by the 1961 Act; tax year 2026-27 onward by the Income-tax Act 2025, with every section renumbered. A model that does not label which Act a rate comes from will be wrong for one of the two years.
Built to be handed over
No hidden sheets, no locked logic, no dependency on us. If the model cannot be maintained by your own finance team after the walkthrough, it has not been delivered.
Questions we are asked before an engagement starts
Can you build the model in something other than Excel?
Usually the answer should be no. Excel is what a lender's credit team, an investor's analyst and your own board can all open and interrogate. A model in a proprietary tool is a model most of your readers cannot check, which defeats the purpose.
Will the model tell me what my company is worth?
It will tell you what a set of assumptions implies, which is a different sentence. If you need a number that can be relied on for a statutory or contractual purpose, you need a valuation report. See who can sign a valuation report in India.
How long does a model take?
A focused three statement model for a single entity is usually two to three weeks from a complete data set. A project finance model with a full debt schedule, or a group model with intercompany eliminations, takes longer. The variable is almost never the building; it is the time taken to settle the assumptions.
Can you review a model we already have?
Yes, and it is often the better first step. A review tells you whether the problem is the model or the assumptions, and those have very different fixes.
Send an enquiry
Tell us what decision the model has to support, who will be reading it, and when you need it. If you already have a model, say so, because reviewing one is usually faster and cheaper than starting again.
This page is general information, not professional advice. A financial model is only as good as the rules it assumes will still be there when the forecast period arrives, and most of those rules moved recently. The Income-tax Act 2025 replaced the 1961 Act on 1 April 2026 and renumbered every section. The Reserve Bank's project finance regime was rewritten with effect from 1 October 2025. The external commercial borrowing framework was rewritten in February 2026. The treatment of a share buyback changed twice inside eighteen months. Take professional advice before acting on anything on this page. We are happy to be that adviser, but we do not act on a web page, ours or anyone else's, without one.