Two years of data. Over 2,400 STRONG and VERY_STRONG insider purchase signals. Every hold period from 30 days to one year, benchmarked against the S&P 500. Here is what the numbers actually say about when insider signals work, how long to hold, and which months delivered the strongest edge.
The Baseline: How Both Years Performed
Before comparing years, the aggregate picture across 2024 and 2025 is striking. Following STRONG and VERY_STRONG insider purchase signals — the top two tiers of the Verity Signals model — produced positive returns across every single hold period studied.
65%
Average 3-month win rate across STRONG/VERY_STRONG purchase signals in 2024 and 2025 combined
That consistency — two out of three signals profitable at 90 days — is not what you'd expect from random stock selection, where win rates hover around 50% over comparable periods. The gap is the signal.
2024 vs 2025: The Year-by-Year Breakdown
The two years look similar on the surface but tell different stories when you dig into alpha — the return above what the S&P 500 delivered over the same period.
| Hold Period | 2024 Avg Return | 2024 Win Rate | 2024 Alpha vs S&P | 2025 Avg Return | 2025 Win Rate | 2025 Alpha vs S&P |
|---|---|---|---|---|---|---|
| 30 days | +6.7% | 64.4% | +4.5% | +6.8% | 65.8% | +5.6% |
| 90 days | +9.7% | 65.0% | +5.1% | +12.9% | 64.8% | +7.6% |
| 180 days | +13.4% | 63.3% | +5.3% | +24.0% | 65.4% | +10.4% |
| 1 year | +15.6% | 60.1% | −0.6% | +32.2% | 65.3% | +14.4% |
The most important number in that table is the 2024 one-year alpha: −0.6%. Insider signals in 2024 returned +15.6% over a full year — impressive in absolute terms — but the S&P 500 delivered 16.2% over the same windows. Holding a diversified index for 12 months marginally outperformed following insiders, once you account for the market's direction.
This is not a failure of insider signal analysis. It is a reflection of how a strong bull market compresses alpha: when the tide raises every boat, outperformance concentrates at shorter hold periods where the insider's information edge is still fresh. At 12 months, macro factors dominate.
In a strong bull market, insider alpha concentrates at 30–90 days. After a market correction, patience pays — the 2025 one-year alpha was +14.4%.
The Best Hold Period: 90 Days
If you could only choose one hold period across both years, the data points to 90 days. Here is why:
- Consistent win rate: 65% in both 2024 and 2025. The 1-year win rate dropped to 60.1% in 2024, suggesting that holding too long introduces noise that degrades the signal.
- Meaningful alpha in both regimes: +5.1% alpha in the 2024 bull market, +7.6% in the more volatile 2025. Unlike the 1-year hold, 90 days generated positive alpha regardless of market conditions.
- Capital efficiency: A 90-day rotation allows four full cycles per year. Compounding four 9–13% returns is structurally more efficient than one 15% annual hold, assuming you have signals to redeploy into.
The 180-day hold is compelling for 2025 signals specifically — insiders who bought during the early-2025 selloff and were held for six months saw average returns of +24%. But that result is context-dependent (discussed below). As a default hold period, 90 days is the most robust choice across market environments.
Which Months Generated the Best Signals
The monthly breakdown reveals a pattern that experienced investors will recognize immediately: insiders buy dips, and those dips are often the best entry points. The full 24-month picture makes this impossible to miss.
| Month | Signals | Avg 3m Return | Win Rate |
|---|---|---|---|
| 2024 | |||
| Jan 2024 | 69 | +13.7% | 76.8% |
| Feb 2024 | 96 | +5.0% | 55.2% |
| Mar 2024 | 106 | +10.8% | 65.1% |
| Apr 2024 | 68 | +10.5% | 79.4% |
| May 2024 | 150 | +5.9% | 63.3% |
| Jun 2024 | 104 | +16.4% | 65.4% |
| Jul 2024 | 44 | +10.6% | 61.4% |
| Aug 2024 | 120 | +15.5% | 77.5% |
| Sep 2024 | 90 | +16.8% | 73.3% |
| Oct 2024 | 47 | +16.1% | 59.6% |
| Nov 2024 | 108 | +6.9% | 61.1% |
| Dec 2024 | 119 | −3.3% | 47.9% |
| 2025 | |||
| Jan 2025 | 55 | +3.0% | 34.5% |
| Feb 2025 | 95 | −4.3% | 33.7% |
| Mar 2025 | 165 | +9.6% | 55.2% |
| Apr 2025 | 82 | +21.4% | 85.4% |
| May 2025 | 189 | +21.5% | 76.2% |
| Jun 2025 | 99 | +21.9% | 80.8% |
| Jul 2025 | 47 | +14.2% | 63.8% |
| Aug 2025 | 143 | +15.6% | 64.3% |
| Sep 2025 | 73 | +13.8% | 58.9% |
| Oct 2025 | 62 | +5.4% | 69.4% |
| Nov 2025 | 181 | +12.0% | 72.4% |
| Dec 2025 | 73 | +7.5% | 60.3% |
STRONG and VERY_STRONG open-market purchase signals only. 3-month return measured from trade date. Win rate = % of signals with positive 3m return.
The Tariff Crash Signal: Q1–Q2 2025
The most important data point in the table is the cluster of December 2024, January 2025, and February 2025: three consecutive months with win rates of 48%, 35%, and 34% respectively. These are the worst insider signal months in our two-year dataset. Insiders who bought into the late-2024 highs and the early-2025 tariff-driven selloff saw their positions immediately underwater.
But read the next row carefully: April 2025 had the highest win rate in the entire dataset — 85.4% — with an average 90-day return of +21.4%.
What happened between February and April? The tariff selloff deepened and then reversed sharply. Insiders who bought in February bought into falling prices. Insiders who bought in April bought near the bottom, after the capitulation. The signal is the same — a Form 4 filing, a STRONG or VERY_STRONG score — but the market context determines whether you're catching a falling knife or buying the dip.
April 2025: 85% of insider signals profitable at 90 days. February 2025: 34%. Same signal type. Different market context. The difference between catching a falling knife and buying the bottom.
This is the core challenge in following insider data: insiders are not market timers. A CEO who buys $2M of stock in February because they believe in the long-term value of their company is not wrong about the fundamentals — they may simply be early. The market can stay irrational longer than their holding period assumption.
High Signal Volume Months vs High Quality Months
An underappreciated tension in this data: the months with the most signals are not always the best-performing months.
May 2025 had the highest signal count of any month — 189 STRONG or VERY_STRONG purchase signals — and delivered +21.5% average 90-day return. That convergence is rare. More commonly, high-volume months reflect broad market stress where many insiders are buying simultaneously (which is itself a useful signal about market regime), but the quality distribution widens.
November 2025 had the second-highest signal count (181) with a solid 72.4% win rate and +12.0% average 3m return. March 2025 (165 signals) had a moderate 55.2% win rate — below average for the dataset — because signals were generated during active market uncertainty where outcomes diverged sharply.
The practical implication: signal volume is a secondary indicator. The quality tier (STRONG vs VERY_STRONG) and the market context at the time of purchase matter more than raw count.
Best Individual Signals: 2024
Several 2024 signals stood out across the full dataset by absolute one-year return:
- UMAC (Unusual Machines): CEO Allan Evans purchased $250K in October 2024. One-year return: +807%. A small-cap defense/drone company at the right time, with the CEO deploying personal capital right before a sector re-rating.
- SMMT (Summit Therapeutics): CEO Mahkam Zanganeh — VERY_STRONG signal, $790K purchase in March 2024. One-year return: +432%. A CEO-level conviction buy in a clinical-stage biotech ahead of trial results.
- NANX (Nanophase Technologies): President/CEO Jess Jankowski purchased $1.6M in February 2024. One-year return: +455%. One of the largest CEO purchases as a percentage of company market cap in the dataset.
What these share: C-suite level insiders, significant personal capital at risk, and companies where insider knowledge of product/clinical/contract pipelines was material. None were diversified large-caps where CEO purchases are routine governance theater.
Best Individual Signals: 2025
Because 2025 signals don't yet have full one-year data, three-month returns are the benchmark:
- AEVA (Aeva Technologies): A $57.8M institutional VERY_STRONG purchase in March 2025 — right at the tariff selloff bottom. Three-month return: +476%. The size alone ($57.8M) made this one of the largest signals in the dataset by dollar value.
- RGLS (Regulus Therapeutics): CEO Joseph Hagan purchased $80K in January 2025 during the selloff. Three-month return: +698%. A CEO buying during maximum market pessimism.
- BW (Babcock & Wilcox): CEO Kenneth Young — VERY_STRONG — $80K purchase in August 2025. Three-month return: +324%. Industrial/energy sector, CEO doubling down in a cyclical company during sector weakness.
What This Means for 2026
The two-year data points to a framework for interpreting insider signals going forward:
- Default hold period: 90 days. Consistent alpha in both bull and volatile markets. Beyond 90 days, alpha varies dramatically by macro regime.
- Market context matters: Insider signals filed during market stress (post-selloff, elevated VIX) historically outperform signals filed at market highs. April 2025's 85% win rate vs February 2025's 34% win rate illustrates this in the same year.
- Role and size: CEO/CFO purchases at meaningful personal cost (relative to their compensation) have historically shown stronger predictive value than director purchases or token amounts. The SMMT, NANX, and AEVA signals above all cleared this bar.
- Cluster buying amplifies the signal: When multiple insiders from the same company or sector buy in the same window, outcomes improve. The April–May 2025 cluster (dozens of insiders buying the tariff dip simultaneously) produced the dataset's best monthly performance.
+7.6%
Average alpha above S&P 500 at 90 days for STRONG/VERY_STRONG insider purchase signals in 2025 — the strongest 90-day edge in the two-year dataset
Insider signals are not a standalone strategy. They are an informational input — one with a documented edge when filtered for quality (role, size, context) and held at the right duration. The 2024–2025 data confirms that the edge is real, consistent at 90 days, and strongest when insiders are buying into market weakness rather than extending market highs.