When one executive buys their own company's stock, it might mean a lot or nothing at all. Maybe they think the shares are cheap. Maybe they just got a bonus and their adviser told them to show some confidence.
When three or more insiders buy the same stock within a couple of weeks, it's harder to explain away. We call that a cluster buy, and it's the main thing this site tracks. The obvious question is whether it actually means anything for the stock price.
So we checked.
How we measured it
The rules were kept simple, and we set them before looking at any results:
- A cluster buy is 3 or more different insiders (officers, directors or 10% owners) buying the stock on the open market within 14 days. Stock grants, option exercises and tax-related trades don't count.
- The clock starts at the close on the day the last insider's Form 4 was filed with the SEC. That's the first day an ordinary investor could have known about all of it. Starting from the insiders' own trade dates would make the results look better than anything a reader could actually get.
- We look at the price change over the next 90 calendar days and compare it with the S&P 500 over the exact same days.
- Every cluster counts. We don't drop the ones that went badly.
That gives us 436 cluster buys with a full 90 days of results, filed between Oct 1, 2025 and roughly three months ago.
The headline numbers
| 90-day result | |
|---|---|
| Average return | +5.4% |
| Median return | +3.7% |
| S&P 500, same days (average) | +4.9% |
| Average difference vs. the S&P 500 | +0.5% |
| Share that beat the S&P 500 | 46% |
| Share that went up at all | 58% |
On average, cluster buys landed almost exactly in line with the market. The signal, at least over 90 days in this sample, hasn't been a shortcut to beating the S&P 500.
The "share that beat the S&P 500" line matters most to us. Even a strong average comes with plenty of losers along the way. Here fewer than half of the clusters beat the market, so on any single stock the odds were not in your favor.
Does it matter if the CEO or CFO is buying?
A common belief is that buying by the CEO or CFO means more than buying by board members, because they see the numbers first.
| Clusters | Average 90-day return | vs. S&P 500 | Beat the S&P 500 | |
|---|---|---|---|---|
| Included the CEO or CFO | 276 | +5.6% | +0.6% | 48% |
| Directors and others only | 160 | +5.1% | +0.2% | 44% |
In our data so far, it hasn't made much difference either way.
Do bigger clusters mean more?
| Clusters | Average 90-day return | vs. S&P 500 | Beat the S&P 500 | |
|---|---|---|---|---|
| 4 or more insiders | 224 | +6.5% | +1.7% | 50% |
| Exactly 3 insiders | 212 | +4.2% | -0.8% | 43% |
More insiders buying has gone with better results so far, which makes intuitive sense: it's harder for four or five people to all be wrong in the same way.
The best and the worst
Here are the three best and three worst clusters, measured against the market. We always show both, because a list of winners on its own tells you nothing.
Best 3
| Stock | Insiders | Filed | 90 days | S&P 500 |
|---|---|---|---|---|
| ALMS | 5 | Dec 1, 2025 | +304.4% | +1.2% |
| AP | 4 | Dec 2, 2025 | +243.1% | +1.0% |
| MANE | 5 | Feb 27, 2026 | +124.4% | +10.3% |
Worst 3
| Stock | Insiders | Filed | 90 days | S&P 500 |
|---|---|---|---|---|
| FBLG | 3 | Mar 5, 2026 | -87.2% | +11.0% |
| HCWB | 3 | May 27, 2026 | -77.7% | +2.3% |
| PSNY | 7 | Jul 6, 2026 | -70.0% | +3.4% |
The spread between those two lists is the real lesson. A cluster buy tells you that people inside the company were willing to put their own money in. It doesn't tell you they're right.
What we take from this
- Cluster buys are a starting point, not a buy signal. They're a good way to find stocks worth a closer look. They aren't a reason to buy on their own.
- Look at the whole spread, not the average. Even when the average is good, a lot of individual clusters lose money.
- Check why the stock is down. Insiders often buy after a big drop. Sometimes that's a bargain, and sometimes the drop was deserved. A company's recent 8-K filings and news usually tell you which. Both appear on every stock page here.
- Give it time. Insiders are usually thinking in years, not weeks. Ninety days is a short window, and we'll extend this analysis as our history grows.
Caveats
This is a young dataset, built from about a year of filings, and short-term results swing around a lot. The returns are price changes only, before trading costs, taxes and dividends. Past results don't predict future ones.
These numbers update every time the site rebuilds, so they may be different by the time you read this. The full list, including every cluster and not just the top and bottom three, is on the scorecard. You can see the current cluster buys here.