How SAVNG calculates every number
Methodology, data sources, and — just as important — the limitations. Last updated September 8, 2026.
Who builds this
SAVNG was built by Pouyan Golshani, MD — a physician and the founder of Gighz — for busy professionals who want to understand what they own without reading 200-page filings. That origin shapes the product: plain-English explanations, every assumption visible, and tools that let you change the assumptions yourself.
Data sources
- Financial statements: SEC EDGAR company filings (10-K / 10-Q XBRL), normalized into 5-year series. Refreshed continuously; each page shows its own freshness stamps.
- Prices: exchange end-of-day quotes, refreshed daily and re-checked when you open a page.
- Insider activity: SEC Form 4 — open-market buys and sells only. Grants, option exercises and tax-withholding are excluded; sells are further split into discretionary vs pre-scheduled 10b5-1 plan sales (since 2023, Form 4 carries a mandatory 10b5-1 checkbox and adoption date).
- Trading-plan adoptions/terminations: Reg S-K Item 408 disclosures in each 10-Q/10-K — a leading indicator, disclosed about 3 months before plan selling can begin.
- Short positioning: FINRA consolidated short interest (all U.S. exchanges, bi-monthly, ~5 years of history) and FINRA Reg SHO daily short volume.
The valuation models — routed by business type
A single formula can't value every business, so each company is routed to the model that fits how it actually makes money:
| Business type | Model | Why |
|---|---|---|
| Profitable operating company | Free-cash-flow DCF (3 scenarios, blended) | Cash the business generates, discounted to today. |
| Cyclical / commodity producer | Normalized (mid-cycle) FCF | Peak/trough cash flow misleads; we use the cycle median and say so. |
| Bank / insurer | Residual income (book value + excess returns) | Banks are valued on what they earn on their capital, not FCF. |
| REIT / utility / MLP | Dividend models + sector lenses (AFFO, P/B, coverage) | Payout-driven businesses are judged on distributions and asset value. |
| Pre-profit growth | Revenue-based growth DCF + reverse DCF | No profits to discount — so we show what the price assumes instead. |
| Crypto funds & Bitcoin-treasury companies | No fair-value verdict | The price tracks the coin; a cash-flow number would be theater. |
Key inputs are shown on every page: the discount rate comes from each stock's own risk profile (CAPM: risk-free rate + beta × equity-risk premium, clamped to 6.5–16%), growth comes from the company's delivered history blended with sector norms (capped at 25%), and terminal growth is clamped below the discount rate. The reverse DCF runs the same math backwards: what growth would justify today's price?
Where our models are weak — read this part
- Foreign companies (ADRs) that report in another currency: we suppress the dollar fair-value rather than mix currencies. A converted valuation is on the roadmap.
- Alternative asset managers (private-equity firms): GAAP cash flow understates their economics, so our DCF reads conservative — the page says so.
- Sum-of-parts businesses (conglomerates, holding companies) and companies with broken multi-class share counts: per-share math is suppressed and replaced with the appropriate lens.
- Any model output can be wrong. A DCF is a disciplined estimate, not a prophecy. That's why every page shows the assumptions, a calculator to change them, and the cross-checks (peers, football field, financial health).
When a number fails our internal sanity checks, the page suppresses it and explains why instead of printing it. We'd rather show you an honest "this model doesn't fit" than a precise-looking wrong answer.
Freshness & corrections
Each stock page carries three freshness stamps (price · SEC financials · insider/short data). Prices refresh daily and on page view; filings refresh as companies file; insider and short data refresh on a rolling basis. If you find an error, email [email protected] — verified errors are corrected on the page and noted.