Our Methodology
How our valuation engine actually works — the inputs, distribution assumptions, simulation procedure, Reverse-DCF logic, and honest limitations.
User Inputs & Adjustable Variables
Every valuation starts from a handful of drivers you can inspect and adjust in the analysis report. Revenue growth is based on recent results and the five-year average. Operating margin uses trailing twelve-month (TTM) data. ROIC measures how much return the company earns on its invested capital. WACC is our discount rate, and capital intensity captures how much invested capital is needed per dollar of revenue. Each of these variables is exposed as a slider in the analysis report, so you can move a single assumption and instantly see how sensitive the resulting valuation is to it. Our WACC framework follows the discount-rate methodology taught in Aswath Damodaran's discounted cash flow valuation materials.
Distribution Assumptions (Concept Level)
Rather than assuming a company will grow linearly forever, we model uncertainty around plausible ranges. Revenue growth, operating margin, and ROIC are assumed to follow mean-reverting paths in log space, with normally-distributed variation layered on top. Companies with higher quality are modeled as holding their current levels for longer before reverting toward the long-run mean. WACC is sampled from a normal distribution and then clipped within a plausible upper and lower bound. The terminal growth rate is sampled uniformly within a range anchored to long-term GDP-level growth. These are conservative statistical conventions used to generate a range of outcomes, not deterministic forecasts. Distribution and scenario sampling follow the Monte Carlo procedure — defining a distribution per input, sampling, and repeated runs — described in NIST's Monte Carlo Tool (ASTM E1369).
Simulation Procedure
We run thousands of simulated scenarios per valuation. Each scenario generates a 10-year DCF path that yields a per-share value, and the collection of scenarios forms a distribution of possible share prices. From that distribution we report P10, P25, P50, and P75 percentiles, the probability of a positive return, and a histogram of outcomes. Where the simulated distribution diverges substantially from the current market price, we blend the result conservatively toward the market price — a Bayesian-anchoring concept that keeps extreme model outputs from dominating the answer.
Reverse-DCF Logic
A standard DCF asks what a stock is worth. Reverse-DCF asks the opposite: what growth rate is the market already pricing in? We numerically solve — using a bracketed root-finding method solved against the current enterprise value — for the growth rate that justifies today's price. The search range spans from extremely low to very high growth scenarios. The implied growth rate (g_implied) is then compared against the company's actual five-year growth rate (g_actual). The gap between the two is used to classify the stock into one of four zones: Strong Buy, Buy, Fair, or Expensive. Reverse-DCF framing is informed by equity valuation standards discussed in CFA Institute insights.
Data Sources
We pull financial statements and market data primarily from Finnhub, with SEC EDGAR XBRL as a fallback. Data is refreshed daily, and calculation results are cached for 24 hours for performance. For the full breakdown, including provider details and update frequency, please see our data sources page. Primary filings are retrieved through SEC EDGAR's official XBRL data access.
Model Limitations
We try to be transparent about what this model cannot capture. Future cash flows are inherently uncertain, and distribution assumptions are subjective choices rather than objective facts. The model does not directly reflect macro shocks, mergers and acquisitions, changes in capital structure, or industry-specific regulation. Small-cap and newly listed companies may have thinner data coverage, which makes their results noisier. We also apply a single sector-neutral baseline for comparability, which can miss company-specific tail risks. Finally, every output is a probability estimate drawn from a model — not a prediction or a recommendation.
Disclaimer
Nothing on this page is investment advice, nor should any valuation or classification be treated as a recommendation to buy or sell any security. Always do your own research and consider your own risk tolerance. By using this site you agree to our terms of use.
Last updated: 2026-08-26
Ready to analyze a stock?
Try It Now