Why Volatility Is the Hardest Input
Of all the inputs to the Black-Scholes model, expected volatility (σ) has the single largest impact on fair value — and for private companies, it's the one input you can't directly observe.
Public companies can use their own historical stock price volatility. Private companies don't have a traded stock price, so ASC 718 allows you to estimate volatility using comparable public companies — often called "comps" or "guideline public companies."
Step 1: Select Comparable Companies
Choose 3–10 public companies that are similar to yours in terms of:
- Industry — Same sector or sub-sector (e.g., B2B SaaS, fintech, biotech)
- Stage / size — Revenue range, market cap, and growth trajectory
- Business model — Recurring revenue vs. transactional, hardware vs. software
- Risk profile — Pre-revenue or profitable, regulatory risk, customer concentration
Auditor tip: Document why each comp was selected and why others were excluded. A well-reasoned comp set with 5 companies is stronger than a generic list of 15.
Step 2: Choose a Lookback Period
The lookback period is the historical window over which you measure each comp's stock price volatility. ASC 718 guidance states that it should be commensurate with the expected term of the option.
| Expected Term | Recommended Lookback |
|---|---|
| 5 years | 5 years |
| 6.5 years | 6.5 years |
| 7 years | 7 years |
The logic: if you're estimating volatility for an option with a 7-year expected life, you want volatility measured over a similar horizon. Using a shorter window (e.g., 3 years) may not capture full market cycles.
Step 3: Pick a Frequency
You can compute historical volatility from daily, weekly, or monthly return data. Each has trade-offs:
Daily
Most data points. Can be noisy for thinly traded stocks. Most common choice.
Weekly
Reduces microstructure noise. Good for small-cap comps with low liquidity.
Monthly
Least data points. Can underestimate true volatility. Use only for very long lookbacks.
Whichever frequency you choose, be consistent across all comps. Don't mix daily for one company and weekly for another.
Step 4: Data Hygiene
Raw price data often needs cleaning before computing volatility:
- Minimum observations — Exclude comps with insufficient trading history (e.g., fewer than 250 daily observations for a 1-year window).
- Winsorization — Trim extreme returns (e.g., top/bottom 0.5 %) to reduce the impact of one-off events like M&A announcements or flash crashes.
- Adjusted close prices — Always use split-adjusted and dividend-adjusted closing prices.
- IPO proximity — Stock prices in the first 3–6 months after IPO can be unusually volatile. Some practitioners exclude this period.
Step 5: Calculate and Select the Central Estimate
After computing the annualized historical volatility for each comp, you need a single number for the Black-Scholes model. Common approaches:
What Ranges Are Normal?
| Company Type | Typical σ Range | Notes |
|---|---|---|
| Large-cap tech | 25–40 % | MSFT, ORCL, GOOG |
| Mid-cap SaaS | 40–60 % | DDOG, NET, HUBS |
| Early-stage / biotech | 55–80 % | Pre-revenue, clinical stage |
| Hardware / semis | 35–55 % | Cyclical, capital intensive |
If your estimated volatility falls outside these ranges, it's not necessarily wrong — but you should document why. An unusually low or high number will likely prompt auditor questions.
Putting It Together
Example: Series B SaaS startup
- Expected term: 6.5 years (simplified method)
- Comps: DDOG, NET, ZS, HUBS, MDB (5 mid-cap SaaS companies)
- Lookback: 6.5 years, daily frequency
- Hygiene: Winsorize at 0.5 %, minimum 250 daily observations
Let ValPack do the volatility analysis for you
Add your comps, pick a lookback period, and get a documented volatility estimate with sensitivity analysis — ready for your auditor.
Open calculator