The negative averages matter because they stop you from romanticizing the signal. A director sale in a small cap is not automatically bearish, but the bucket itself has not been a rich hunting ground on average. The 38.6% win rate says the hit rate is below coin-flip territory. The -3.28% 90-day average return says the cohort has, on average, lost money over that horizon. The -11.87% 365-day average return says the weakness is not just a short-term artifact. If you are trying to extract a directional edge from this filing alone, the historical bucket does not help you much.
That is exactly why the cluster and the company context matter. A weak cohort does not erase a live filing. It just keeps you from overpaying for the signal. In a name like Sidetrade, the better use of the data is to ask whether the current pattern is meaningfully different from the historical bucket. Here, the answer is mixed. The cluster is real. The role is relevant. The price context is ordinary. The cohort is weak. That is a balanced read, and balanced reads are usually the ones that survive contact with the tape.
The fundamental screen is decent, not heroic
InsiderTrades fundamental data gives Sidetrade a score of 56, with a rank of 9047 out of 21371. The value pillar is 43 and the quality pillar is 68. Growth is not provided in the dossier, so there is no reason to pretend otherwise. The screen is not screaming cheap, and it is not screaming broken. It looks like a middle-of-the-pack software name with a better quality profile than value profile.
That matters because insider sales in a company with weak fundamentals can be easier to dismiss as part of a broader deterioration. Here, the picture is more nuanced. A quality score of 68 suggests the business is not being treated as a weak operator by the screen. A value score of 43 says the market is not exactly offering a deep discount either. So the filing is not arriving against a backdrop of obvious distress or obvious bargain pricing. It is arriving in a name that looks operationally respectable, but not cheap enough to make every sale look like a gift to buyers.
This is where the fundamental screen earns its keep. It does not tell you whether Dupont-Aldiolan’s sale was wise. It tells you the company is not sitting in an extreme fundamental bucket that would make the insider tape easy to interpret. If the screen were much stronger, you might be more comfortable treating the sale as routine trimming. If it were much weaker, you might lean harder into caution. Instead, the screen is middling. That leaves the insider data to do the heavy lifting, and the insider data is mixed. That is the honest state of play.
The cluster picture is the real story here
InsiderTrades data marks this as a cluster event. There are four distinct insiders in the recent picture and five recent declarations. The list includes Frédéric Dupont-Aldiolan, Robert Harvey, Olivier Novasque, and Mark Sheldon. The recent filings show Dupont-Aldiolan selling on June 23, Robert Harvey selling on June 12, Olivier Novasque buying on June 11, Mark Sheldon buying on May 18, and Robert Harvey selling again on April 7. That is not a one-way tape.
The presence of both buys and sells is the point. Clusters are useful when they show alignment. They are also useful when they show disagreement, because disagreement tells you the market is still digesting the story internally. Here, the recent sequence is mixed. One director sells, another director sells earlier, the CEO buys, and another director buys. That does not give you a clean bullish or bearish cluster. It gives you a live one. The market is seeing activity from multiple insiders over a short window, and the direction is not uniform.
That mixed cluster is more informative than the single filing alone. If you only looked at Dupont-Aldiolan’s sale, you might overstate the bearishness. If you only looked at Novasque’s buy, you might overstate the confidence. Put them together and you get a company where insiders are active, but not synchronized. That is often what a transition phase looks like in practice. It is also what makes the tape worth watching rather than declaring done.
What the filing can and cannot tell you
The filing tells you that a CHRO sold 1,000 shares at EUR 156.44, that the transaction was worth about EUR 156,440, and that it sits inside a cluster. It does not tell you why the sale happened. It does not tell you whether the insider was rebalancing, funding something personal, or simply taking money off the table after a run. It does not tell you whether the company’s AI deployment story is stronger or weaker than the market thinks. It does not tell you whether the next filing will be a buy or a sell.
That is the boundary line. You can read the signal. You cannot turn it into a confession. The temptation with insider data is always to over-interpret motive because motive makes for a neat story. But the better trade is usually the less dramatic one. Here, the less dramatic read is that Sidetrade is a small-cap software name with a mixed insider cluster, a weak historical cohort bucket, and a decent but not exceptional fundamental screen. That is enough to justify attention. It is not enough to justify certainty.
If you are weighing the name, the part to sit with is the combination of cluster activity and the weak cohort backdrop. The cluster says insiders are active. The cohort says the role-and-size bucket has not rewarded that kind of activity on average. The company context says the business is still in an execution phase, with AI deployment and organizational changes in the news. Those three facts do not point in the same direction. That is why the filing is interesting.
What to watch next
The next thing to watch is whether the cluster extends or fades. One more director sale would make the bearish side of the tape harder to ignore. Another buy from the CEO or another operating insider would keep the picture mixed and probably keep the signal in the “watch, do not chase” category. In other words, the next filing matters more than the last one, because the last one is already embedded in a cluster.
Also watch whether Sidetrade’s company news shifts from deployment headlines to financial commentary. The brief gives you product and organization updates, not a fresh financial reset. If the company starts talking more explicitly about margins, bookings, or demand, the insider tape will be easier to place. If it stays on the AI deployment track, then the market is left to infer the operating trajectory from filings and scattered announcements. That is not ideal, but it is common enough.
For now, the cleanest conclusion is modest. Frédéric Dupont-Aldiolan sold a meaningful but not outsized block of Sidetrade shares on June 23. InsiderTrades data flags the trade because it is a director filing, part of a cluster, and lodged in a small-cap name where insider activity has historically mattered more than it does in larger, more liquid stocks. The historical cohort is weak, the fundamentals are middling, and the cluster is mixed. That is a real signal, just not a loud one.
SIDETRADE story">
SIDETRADE insider-trading story">