Three screens report $35.9m of Boersma selling. He sold $18.0m. The error is structural Travis Boersma · Dutch Bros · Form 4, 30 October
- Who
- Travis Boersma, co-founder and Executive Chairman of Dutch Bros, holding indirectly through DM Trust Aggregator, LLC and DM Individual Aggregator, LLC, of which he is manager.
- What he did
- Sold 408,023 Class A shares on 29 October at a weighted $44.0266 ($17,965,252.69), then 57,478 on 30 October at $44.09 ($2,534,113). Combined for the two days: 465,501 shares, about $20.50m. Class A held after: 6,314,242.
- The document
- Form 4 filed 30 October 2026, 16:07 and 18:40 ET, reporting both LLCs on one form. The 22 October form covering the 20 October tranche cites a Rule 10b5-1 plan adopted 19 February 2026.
| Date | DM Trust | DM Individual | Total shares | Wtd price | ≈ Value |
|---|---|---|---|---|---|
| 20 Oct 2026 | 94,978 | 50,709 | 145,687 | $44.0080 | $6.41m |
| 29 Oct 2026 | 266,000 | 142,023 | 408,023 | $44.0266 | $17.97m |
| 30 Oct 2026 | 37,470 | 20,008 | 57,478 | $44.09 | $2.53m |
| October total | 398,448 | 212,740 | 611,188 | — | ≈$26.9m |
| Held after 30 Oct | 4,116,182 | 2,198,060 | 6,314,242 | — | — |
Consensus read
The co-founder is liquidating heavily into a broken chart. Aggregators showed Boersma at $17.96m, DM Trust Aggregator at $11.71m and DM Individual Aggregator at $6.25m in the same weekly table — read as roughly $36m of chairman selling in two days, days before earnings, with the stock 42% off its high. A reasonable person reads that as a founder getting out.
Contrarian read
Two of those three rows are the same shares. 266,000 + 142,023 = 408,023, exactly the “Boersma” row. He sold $17.97m on the 29th, not $35.9m, because a combined Form 4 reports the manager and each LLC and a flat screener counts all three.
The real finding sits one layer down and is the opposite of a dump: tranche size tripled while the price stood still. That is the fact worth having, and it is invisible if you are busy being alarmed by a number that is double-counted.
Evidence that discriminates
- The holdings columns reconcile exactly. DM Trust 4,419,652 → 4,153,652 is −266,000. DM Individual 2,360,091 → 2,218,068 is −142,023, then → 2,198,060 is −20,008. Those are the whole of the 29–30 October selling. There is no third pocket for a separate $17.96m to have come out of. This settles the double-count on arithmetic rather than on inference.
- Price is constant across the three tranches — $44.0080, $44.0266, $44.09, a spread of about 0.19%. A price-triggered plan tranche cannot grow 2.80× on a 1.9-cent move. The variable that changed is not price.
- It is not an earnings-window effect either. Dutch Bros reports Q3 on 4 November; the window ahead of a print is the closed one, not the open one. So the usual mundane explanation — “the blackout lifted” — is unavailable here, which is what makes the size jump worth flagging rather than filing away.
Inference, labelled: the remaining live explanations are a calendar-tranched plan with deliberately uneven sizes, or a plan adopted or modified after 19 February 2026. The filings reachable today do not distinguish them, and I am not going to pretend otherwise.
What would falsify me
The 30 October Form 4 footnote citing the same 19 February 2026 plan with a disclosed schedule that already called for a larger October tranche. That would make the size jump pre-ordained and the flag wrong. Equally: a Form 4/A showing additional shares sold by a third entity on 29 October, which would make the $35.9m figure right and me wrong on the double-count.
Pattern vs his own historyOctober’s disclosed selling, 611,188 shares for roughly $26.9m, is about 8.8% of the Class A he held at the start of the month. The holdings here are Class A only; Dutch Bros also has a high-vote Class B, and nothing in these filings touches voting control. That asymmetry — monetise the economics, keep the votes — is the structure doing its job, and it is the part that generalises.