Session recording
Sloka BalasubramanyanDirector, Valuations, Modelling & Economics, EY
Recording and slides free until 11:59 p.m. ET on Thursday, October 29
A 16-minute edit of the session, in Sloka's own words and in the order she gave it.
The recording of Sloka's session is included with FMI Foundations and the Advanced Financial Modeler (AFM) Program, along with every other session from the conference.
But Sloka's demo workbook is free. Download below.
“I love the deck. I've been teaching modeling for 25 years, never seen these mental models articulated in quite this way, and love the way you've done that.” Ian Schnoor, Executive Director, Financial Modeling Institute, during the 24-Hour Global Financial Modeling Conference
The workbook Sloka built for the session runs one five-year operating and investment model as a rigid build, an over-flexible build and a governed build. Change the scenario on the Control Panel and watch the governed model, checks and outputs update together.
Growth, margin and discount rate typed straight into the formulas. Easy to read, but every new requirement means a formula edit.
Ten Option 1, 2 and 3 switches spread below the model, with no clear sign of which case is running.
Scenario, new initiative and launch year set in one place, with dropdowns and a model state summary.
Five checks roll up to one status, shown beside NPV, revenue, EBITDA and free cash flow for the active scenario.
“Flexibility must be tested as a system”
Sloka's argument is that a model fails in two opposite ways: too rigid to absorb change, or so configurable that only its builder can run it. The goal is governed flexibility, built for the decisions the model supports and organized so someone new can operate it and another modeler can review it.
Few assumptions, short formulas, quick to review. Then a delayed start or a new funding case arrives, and every change becomes a formula edit.
Every option looks helpful on its own. Three on/off switches create eight states to test; ten create 1,024. Often only the builder can run it.
Relevant choices, constrained inputs and a visible model state. Sophisticated inside, deliberate at the interface, and usable without calling the builder.
Switches add flexibility. The interactions between them add fragility, because every permitted combination has to be designed, tested and trusted. The question is never whether a modeler can build an option, but whether the decision needs it.
Before adding any switch, ask five questions from her deck: what decision does it support, could it change the conclusion, is the alternative plausible, how often will someone change it, and can it be handled more simply. If an expansion is a given and only its funding is in question, there is no need for an expansion on/off toggle. A feature that fails the questions gets simplified, redesigned or removed.
Keep the controls in one place, separate from the calculation engine. Restrict dropdowns to valid options, grey out invalid choices, and warn when a combination does not fit the selected scenario. Then show the model state on the outputs: active scenario, overrides in use, checks passed or failed. Nobody should have to read a formula to know which case they are looking at.
Sloka took just two switches: a scenario (base, upside, downside) and a debt-to-equity ratio (50/50, 70/30, 90/10). Each works on its own. Together they make nine combinations, and not all of them hold up. An upside case at 90/10, with only 10% equity, might not work. A downside case at 90/10, she said, definitely does not.
Her other example: push a project's COD back six months. Construction schedule, debt drawdown, interest during construction, depreciation timing, tax losses, covenants and project IRR all move. Those impacts are expected. The fragility is when each schedule has to be edited by hand, so the effort is out of proportion to the change.
Three ways to work through Sloka's argument: how quickly on/off switches multiply, what she said about each scenario and debt combination, and the governed model from her demo workbook, rebuilt to run on this page with her own assumptions and formulas.
Three on/off switches already give eight states to test. Drag to guess how many states ten on/off switches give.
Each on/off switch doubles the count. One switch gives 2 states, two give 4, three give 8, five give 32 and ten give 1,024.
Her slide puts it this way: “Not all combinations are equally likely, but every permitted combination creates a governance question.”
Sloka took a scenario switch (base, upside, downside) and a debt-to-equity ratio (50/50, 70/30, 90/10). Tap the combinations you think are valid, then see what she said about each.
She called the base case possibly okay at all three ratios, and the upside case possibly okay at 50/50 and 70/30. An upside case at 90/10, with just 10% equity, might not deliver the upside, so she defined it as an invalid combination. A downside case at 70/30 was maybe, maybe not, and at 90/10 definitely not. She did not give a verdict on the downside case at 50/50.
“Complexity is not that we have two switches in the model. The complexity is that the choices are no longer independent.”Sloka Balasubramanyan, during the session
This is the governed model from Sloka's demo workbook, rebuilt with her assumptions and formulas. Pick a scenario, switch the new initiative on, or set a launch year outside 2026 to 2031 and see what the checks do.
| 2026 | 2027 | 2028 | 2029 | 2030 | 2031 | |
|---|---|---|---|---|---|---|
| Revenue | 1,000.0 | 1,144.5 | 1,309.9 | 1,499.2 | 1,715.8 | 1,963.7 |
| EBITDA | 270.0 | 309.0 | 353.7 | 404.8 | 463.3 | 530.2 |
| Free cash flow | 160.0 | 223.1 | 255.4 | 292.3 | 334.5 | 382.8 |
Her Over-Flexible tab runs the same five-year model through ten Option 1, 2 and 3 switches. As saved, it returns an NPV of 956.7: base-case volume growth, downside price growth and upside margin at once, a combination that matches none of the three scenarios.
Rebuilt from Sloka's demo workbook: the Scenario Assumptions, Control Panel, Governed Model, Checks and Outputs sheets. The workbook does not state a currency or unit.
Each quote comes as an image for LinkedIn, dark or light, with the post already written. Pick one, download the image, and post it in three steps.
“Switches allow you to create flexibility. What do interactions do? Interactions create fragility.”
“A model must earn its complexity.”
“Wherever your model goes, your modeler goes, and that is not the sign of a great model.”
“A model is robust because it guides users towards valid combinations, and it prevents invalid ones.”
“A clean model interface is not an aesthetic or a cosmetic choice. It is a control decision.”
“Is anybody ever going to use this flexibility? And if the answer is no, then that flexibility probably doesn't deserve to be there.”
Two ways to carry the conference forward: learn how a model is put together, or prove you can build one from scratch. Both FMI programs include every conference video and the presentation materials.
Learn core financial modeling concepts at your own pace. Access all conference videos and presentation materials.
AI can generate a model in seconds, but it can't tell you if the logic holds. The AFM proves you can build trusted financial models. Access learning resources and the full conference video library.
Pick a dark or light background, then save the image. It is a 1080 by 1080 square, which LinkedIn shows in full.
The post below opens in LinkedIn already written, and goes on your clipboard too. Change it here first if you like.
In the composer, choose the photo icon and add the image you downloaded. Type @ to tag Sloka Balasubramanyan and Financial Modeling Institute, then post.
