What the court decided
The case is Valcrum, LLC v. Dexter Axle Company, LLC, a trademark and trade dress fight over hubcap designs. On July 15, 2026, U.S. Magistrate Judge Scott J. Frankel of the Northern District of Indiana issued an opinion and order on a motion to compel. He ordered Dexter to produce all Teams messages sent three days before and after the Bates-labeled messages in Valcrum's amended Third Requests for Production Nos. 112 and 113, by each custodian and in response to each custodian.
Dexter must also give Valcrum unredacted copies of Exhibits 13, 16 and 17. It must do the same for any other threaded Teams conversations containing the identified messages or produced in volume 11 of its production. All other relief Valcrum requested in the motion to compel was denied.
The Law.com commentary on the ruling, by Daniel J. Toal, Geoffrey R. Chepiga and Ross M. Gotler, summarizes it this way. Teams messages need enough surrounding context to be understood, and discovery methods should be tailored to chat data.
How the dispute unfolded
The underlying case is Valcrum's claim that Dexter copied its red, hexagonal hubcap design. The discovery fight was over Dexter employees' Teams chats. The court's recitation of the production history shows how the problem built up:
- Dexter produced individual Teams messages on a rolling basis from mid-January to early December 2025. They were scattered through volumes 1–10.
- On January 10, 2026, volume 11 arrived in a new format. Previously disparate messages were combined into chronological conversation threads.
- Valcrum said the new volume contained correspondence it had not seen before, and that some of the threads were redacted.
- Valcrum then served Third Requests for Production on February 6, 2026. It asked for the full conversations behind earlier messages and for unredacted threads.
Valcrum argued that seeing only context-free messages hurt its depositions of three witnesses. It also argued that Dexter's search methodology missed messages that should have matched Dexter's own criteria. Dexter said its search was proper. It also said it produced messages as kept in the ordinary course of business, and that the threaded renderings were a courtesy, not an admission.
Why keyword hits alone didn't cut it
The court did not accept that search-term hits alone were enough for chat data. It quoted a 2026 Central District of California decision, Connie Kim v. Cushman & Wakefield. That decision says keyword searches alone don't suffice for Teams because conversations are shorter, more informal and less likely to include full names than email.
Anyone who has skimmed a Teams channel will recognize this. A message like "yep, go with that one" matches no search term and means nothing without the lines around it. The three-day window is the court's compromise: enough neighbouring chat to make the hits readable, without turning discovery into a dragnet.
The court also rejected Valcrum's broader ask. Producing all Teams correspondence was too broad, the judge found. The window was meant to give context while keeping the burden manageable.
What the court did not require
This is not a blanket rule, and the limits matter for anyone drawing lessons from it:
- No forced thread reconstruction. The court found Dexter had met its burden by explaining that Microsoft stores individual Teams messages in the inbox of each participating custodian, not as separate threads. It held it was inappropriate to compel Dexter to construct and produce entire threads.
- No single mandated format. The court worked within Federal Rule of Civil Procedure 34(b)(2)(E). That rule says ESI must be produced in the form ordinarily maintained or in a reasonably usable form if no form is specified.
- No fee award. The court declined to award Valcrum its expenses. It found Dexter's objections substantially justified. It also cited other circumstances, including Valcrum's failure to follow good-faith certification requirements.
- No Rule 30(b)(6) order. The request to compel Dexter's corporate-representative deposition was denied.
A DLA Piper write-up and a "Meet and Confer" podcast episode from Kelly Twigger both cover the case. The podcast frames the lesson as agreeing early on short message context. It also stresses the procedural pitfalls: meet-and-confer certification and a real paper trail. That is practitioner commentary, not part of the holding.
Redactions and boilerplate objections
Dexter had redacted parts of the threads it built itself. It objected that the requests were overly broad, unduly burdensome, vague, ambiguous and disproportionate. It did not elaborate. The court treated those objections as waived. It cited a Northern District of Indiana decision criticizing the "reflexive invocation" of a stock list of objections. The practical point is that a generic list of objections, with no explanation, did not justify the redactions.
Timing arguments
Dexter argued the motion came too late and that Valcrum had waited to issue its follow-up requests. The court disagreed on two grounds:
- Valcrum served its requests in February 2026, well before fact discovery was set to close.
- A responding party has an automatic duty to supplement under Rule 26(e). The court noted that Dexter kept producing documents until March 23, 2026.
The court also adjusted the schedule. Fact discovery now closes August 28, 2026, expert reports are due October 1, 2026, and dispositive motions are due December 29, 2026.
What this means for Microsoft 365 and IT teams
The ruling does not change how Teams works. It does shape what lawyers will ask IT to do. The following points are my own analysis, not requirements from the court:
- Know where chat lives. Dexter's explanation of per-custodian storage carried weight. Admins who can explain clearly how their tenant stores and exports Teams data are better placed to answer a motion like this.
- Expect context requests. If a keyword hit pulls out one message, opposing counsel may ask for its neighbours. Being able to export bounded windows around a hit by custodian and conversation is more useful than all-or-nothing exports.
- Document the collection method. The court noted that a party's description of its search carried weight. Written, repeatable collection steps make that description credible.
- Agree on format early. The case grew out of a production format that changed midstream. Settling on thread or message format with the other side at the start can head off a motion.
- Back up objections with specifics. Compliance and legal teams should supply concrete burden detail. Naming categories of objection is not enough.
The caveats
This is a single magistrate judge's discretionary order in a trademark case. It is not binding on other courts, and its three-day window is tied to this dispute, not a universal formula. Other courts may choose different windows or different approaches. Still, as an early example of a court dealing with the shortcomings of keyword search on chat data, it is a useful reference for enterprise IT and legal teams working with Teams.