BSH Home Appliances Corporation appeals from the district court's judgment that it pay $418,261 to Detroit Radiant Products Company for breach of contract. The dispute centers around whether the contract between the two companies, which involved custom-made stove burners, contemplated a minimum fixed number of units binding upon the purchaser or merely a non-binding estimate. Because we conclude that a minimum fixed number was indeed contemplated, we AFFIRM the judgment of the district court.

* Detroit Radiant is a manufacturer of gas-fired infrared heaters used for commercial and industrial applications. BSH is a maker of home appliances under the Bosch, Siemens, Thermador, and Gaggenau brands. BSH had planned the manufacture of a new product, the "Pro 27 Stainless Steel Range," a high-end, restaurant-quality range intended for home use. To this end, BSH supplied Detroit Radiant with detailed specifications for the Pro 27 burner, and requested a price quote based on an estimated annual usage of 30,000 units.

On July 10, 2001, Detroit Radiant quoted a price of between $27.93 and $32.40 per unit, depending on the quantity BSH ultimately ordered. BSH rejected this proposal and specifically requested a price quote for a volume of 30,000 units. Detroit Radiant responded by email with a quote of $27.45 per unit based on 30,000 units, indicating that for this volume it would be willing to absorb all tooling and research and development costs. The price was later increased to $28.25 per unit to include the cost of a "test burn" that BSH had separately requested. On August 10, 2001, BSH sent to Detroit Radiant a signed purchase order for 15,000 burner units at $28.25 per unit. Detroit Radiant began to manufacture and ship burners under this purchase order, pursuant to "release schedules" provided by BSH.1 On January 8, 2003, BSH sent to Detroit Radiant another signed purchase order, this time for 16,000 units at the same price per unit.

Between September 2001 and April 2004, Detroit Radiant shipped 12,886 burners to BSH. BSH accepted and paid for these burners and took no issue with their quality. As of April 5, 2004, BSH had also scheduled the release of approximately 6,000 additional burners for dates in the future, up to and including February 2005. On April 12, 2004, however, a revised release schedule showed future releases dwindling to zero. In other words, it appeared from this schedule that BSH no longer intended to order Pro 27 burner units from Detroit Radiant. As it turns out, by the spring of 2004, BSH had selected a company called Solaronics to be its new supplier of Pro 27 burners and had stopped ordering them from Detroit Radiant entirely, allegedly as a cost-savings measure.2

Detroit Radiant maintained that B SH was contractually bound by its two purchase orders to order a total of 31,000 burners, and that this number was consistent with its prior negotiations and price quote based on a volume of at least 30,000 units. Thus, on June 24, 2004, Detroit Radiant filed a complaint against BSH in Michigan state court, alleging breach of contract and claiming entitlement to damages for the 18,114 units that BSH never purchased. Detroit Radiant claimed $312,104 in lost profits, as well as $52,011 in unused inventory because the Pro 27 burners had been specially manufactured. BSH successfully moved to have the case removed to federal court pursuant to 28 U.S.C. §§ 1332 and 1446.3 A bench trial was held in federal district court, at which BSH argued that the 2003 purchase order superseded the 2001 purchase order, obligating it to purchase at most 16,000 units, not 31,000 units. BSH also stressed that the purchase orders represented only estimates of quantity, not obligations to purchase exact amounts. As such, BSH argued, the purchase orders were a form of requirements contract, through which BSH committed itself to buying only as many burners as needed. The primary evidence adduced by BSH for its claims was: (1) that the 2001 purchase order was specifically labeled a "Blanket Order," an industry custom indicating that it was an estimate of quantity rather than a fixed obligation; and (2) the existence of a drafted but unsigned "Supplier Agreement" containing a clause suggesting that scheduled orders would be "non-binding forecasts" rather than fixed amounts.

On March 1, 2006, the district court issued its Findings of Fact and Conclusions of Law. Applying Michigan substantive law and Michigan's version of Article 2 of the Uniform Commercial Code, the district court found that the August 10, 2001 and January 8, 2003 purchase orders "required BSH to purchase 31,000 units of the burner at $28.25 per unit." This was consistent with the court's finding of a prior contract between the two parties: namely, that Detroit Radiant had agreed to manufacture the burners at the reduced price and to absorb all tooling and research and development costs4 in exchange for BSH's agreement to purchase at least 30,000 burner units. The district court rejected BSH's argument that the purchase orders represented an estimate of burner units only; instead, the court found the two purchase orders to be unambiguous as to quantity and lacking any language that "intimates that the quantity terms are forecasts or estimates." The court also rejected BSH's claims relying on the unsigned supplier agreement, finding that this was not an enforceable contract and therefore did not "govern the terms of the agreement between the parties." Finally, the district court concluded that BSH had failed to "meet its burden in proving that the second purchase order simply replaced the first purchase order."

As to damages, the court ruled that Detroit Radiant was entitled to its lost profits for the specially manufactured burners as well as its unused inventory. Adding interest and adjusting to present value, the district court awarded to Detroit Radiant a total of $418,261. BSH now appeals, raising four issues for review.

In reviewing a bench-trial contract action, we will not set aside findings of fact by the district court unless they are clearly erroneous. Lancaster Glass Corp. v. Philips ECG, Inc. 835 F.2d 652, 658 (6th Cir.1987); see also Fed.R.Civ.P. 52(a) ("Findings of fact, whether based on oral or documentary evidence, shall not be set aside unless clearly erroneous, and due regard shall be given to the opportunity of the trial court to judge of the credibility of the witnesses."). However, the district court's interpretation and construction of a contract is a matter of law, and such matters this Court reviews de novo. Id. The parties agree that both Michigan law and Michigan's version of the Uniform Commercial Code apply to this dispute.

BSH first takes issue with the district court's conclusion of law that "BSH did not meet its burden in proving that the second purchase order simply replaced the first purchase order." D. Ct. Op. at 4. BSH claims that with this one statement, the district court "impermissibly shifted the burden of establishing the terms of the contract and the fact of the breach" from Detroit Radiant to BSH. In other words, BSH argues, because Michigan law requires that a plaintiff bear the burden of proving a breach of contract claim, "it was Detroit Radiant's burden to prove that the second purchase order was an addition to, rather than a replacement of, the first order, not BSH's burden to prove the opposite." Had the district court concluded that the 2003 purchase order (for 16,000 units) replaced rather than supplemented the 2001 purchase order (for 15,000) units, as BSH argues it should have, then a breach by BSH would have resulted in significantly reduced damages.

In support of its burden-of-proof argument, BSH contends that it had presented strong evidence that a "Schedule Agreement" had replaced the 2001 purchase order, and that this Schedule Agreement was embodied by the 2003 purchase order. On December 14, 2001, BSH sent a letter to Detroit Radiant informing it that BSH was implementing a new computer system and that as a result Detroit Radiant "will be receiving a scheduling agreement, which will replace the current purchase order." Joint App'x at 77. Under the new system, what were formerly called "Purchase Orders" would now be called "Schedule Agreements." BSH claims that it ultimately sent out this replacement Schedule Agreement in January 8, 2003, over one year after it sent out the letter. Yet this so-called "replacement" is none other than what Detroit Radiant calls the "2003 purchase order." The argument proffered by BSH simply defies credulity, especially because Detroit Radiant sent four shipments of burners to BSH (totaling 2,493 units) between the time it received the letter and the time it received the January 8, 2003 Schedule Agreement. If BSH had been serious about replacing its 2001 Purchase Order with a Schedule Agreement, it presumably would have done so immediately after issuing the December 14, 2001 letter, not — as BSH claims — more than one year later. And given this significant time lapse it is especially odd, at least under BSH's theory, that the 2003 order says nothing — either on its face or in accompanying documentation — about its being a replacement order. A plain reading of the 2003 order, then, regardless of what the parties may have called it, is that it was independent of the 2001 order. Nor does BSH present any evidence at all as to why a mere format change dictated by a new computer system — that is, from calling what used to be a "Purchase Order" under the old system a "Schedule Agreement" under the new system — would lead to a change in the number of units ordered, i.e., from 15,000 to 16,000.

BSH also claims that Detroit Radiant's Vice President of Operations, Joe Wortman, actually admitted in a June 13, 2003 email to BSH — via the words "I understand that your PO # B 6248 is no longer active" — that the 2001 purchase order had been supplanted by the 2003 Schedule Agreement. Joint App'x at 78. What BSH neglects to point out, however, is the very next sentence in Wortman's email. In full context, the email reads: "I understand that your PO # B 6248 is no longer active. I was just stating that it was never fulfilled." Id. (emphasis added). In other words, whether or not the 2001 purchase order was formally considered "active" had no bearing on the fact that BSH had promised to order 15,000 burner units and had yet to fulfill this promise.

BSH's burden-shifting arguments seem to miss a broader point. The district court found that a contract existed between Detroit Radiant and BSH for the sale of at least 30,000 Pro 27 burner units, a contract which was subsequently clarified by the two purchase orders:

1. BSH initially requested a price quote for the burners from Detroit Radiant on July 5, 2001.

2. Detroit Radiant provided its first pricing quote, based on a range of quantities, July 10, 2001.

3. Price negotiating followed for a month, and Detroit Radiant continued to lower the price based on the understanding that BSH would buy at least 30,000 units.

4. Detroit Radiant agreed to absorb the tooling cost based on the understanding that BSH would buy at least 30,000 units.

5. Detroit Radiant would not have agreed to the deal at only 15,000 total units and would not have absorbed the cost of tooling and research and development at volumes lower than 30,000.

6. While the parties adjusted the annual volume amount, the basic premise of the business deal was always based on the total purchase of at least 30,000 units.

7. The August 10, 2001 purchase order, the first purchase order, committed BSH to purchase 15,000 units.

8. The January 8, 2003 purchase order, the second purchase order, committed BSH to purchase 16,000 units.

9. The 31,000 total number of units identified by the two purchase orders is consistent with and supports the finding that BSH would buy at least 30,000 units.

D. Ct. Op. at 1-2. In finding that the two purchase orders were independent of each other, and consistent with a prior agreement, the district court implicitly rejected BSH's theory that the 2003 purchase order modified the 2001 purchase order. It was perhaps confusing for the district court to have engaged in a discussion of burdens of proof in its subsequent conclusions-of-law section, but such discussion does not alter the end result. The district court's critical finding was the existence of a binding contract between Detroit Radiant and BSH: a reduced price per burner unit and an agreement to absorb tooling and research and development costs, in exchange for a purchase of at least 30,000 units. There was nothing clearly erroneous about this finding, especially because the district court was in the best position to gauge the credibility of the actors whose words and actions gave rise to the contract. And given this finding, BSH's burden-shifting argument holds little water — that is, given the district court's finding that the parties had entered into a binding contract at the outset, it may be implied that Detroit Radiant had already met its burden to show that the 2003 purchase order did not replace the 2001 order, since the two purchase orders, only when added together, were consistent with the 30,000-unit figure.

BSH's replacement theory might stand on firmer footing had the total number of Pro 27 burner units shipped (12,886), combined with the number of units scheduled for release (approximately 6,000), been les than 16,000. But given the fact that this total number — almost 19,000 — was well over 16,000,5 it would be counterintuitive to view the January 2003 purchase order as a replacement, because such an interpretation would leave several thousand units unaccounted for. To account for this inconsistency, BSH must argue that neither of its purchase orders contemplated fixed quantities of burner units — rather, they contemplated estimates only. This argument is addressed in the ensuing two sections.

B. The Unsigned "Supplier Agreement"

Both parties admit that in August of 2001 they engaged in negotiations to draft a "Supplier Agreement."6 This agreement was intended to supplant the parties' prior agreement for at least 30,000 burner units by including numerous additional terms such as warranties, arbitration, assignability, limitations on liability, and so on. While the Supplier Agreement had been drafted by BSH, the record indicates that Detroit Radiant was particularly interested in having something in writing that limited its tort liability in the event of a consumer being injured while using the Pro 27 stove. The parties sent drafts and revisions of the agreement back and forth, but ultimately it was never signed. The final communication regarding the draft agreement was an August 9, 2001 email from Joe Wortman to Curt Sevig, a representative for BSH. Mr. Wortman's email stated: "We had the opportunity to review your [Supplier Agreement] and have revised it in some areas that we felt necessary. The file will show deletions in red, additions in blue and notations in yellow. We are prepared to sign the contract if you are in agreement with our changes." Joint App'x at 65. Neither Mr. Sevig nor BSH responded to Mr. Wortman's proposed changes. The following day, however, BSH sent a purchase order to Detroit Radiant for 15,000 burner units, and Detroit Radiant commenced production.

The significance of the Supplier Agreement, according to BSH, lies in a section entitled "PURCHASE ORDERS/RELEASES/FORECASTS," which contains the following language:

On a regular basis as agreed upon, Buyer shall issue releases/scheduling lines against such purchase orders/scheduling agreements, specifying the quantities to be delivered and delivery dates. The purchase orders/scheduling agreements and releases/ scheduling lines submitted by Buyer are constituted as acceptance, unless otherwise specified by Supplier with a written response. The first month are committed orders, and the remaining months listed on the purchase order/scheduling agreement are non-binding forecast.

Joint App'x at 67, ¶ 5 (emphasis added). BSH claims that this "non-binding forecast" language clearly indicates that its purchase orders were estimates only, and that the district court's conclusion that it was bound to purchase 31,000 burners was erroneous.

The district court found that "[t]he draft, unsigned Supplier Agreement does not cover the terms of the agreement between the parties." D. Ct. Op. at 2. The district court then held that "[t]he draft, unsigned Supplier Agreement is not an enforceable contract and therefore does not govern resolution of the dispute." Id. at 4. BSH argues that the parties' signatures were not required to render the Supplier Agreement enforceable. See M.C.L. § 440.2204(1) ("A contract for sale of goods may be made in any manner sufficient to show agreement, including conduct by both parties which recognizes the existence of such a contract."). BSH claims that because the court treated the lack of signature as dispositive of unenforceability, and nowhere did the district court expressly state that it had considered conduct of the parties with respect to the supplier agreement, therefore at minimum a remand is in order to determine the parties' intentions.

BSH is correct that a lack of signature on the supplier agreement is not dispositive on the issue of its enforceability, but this is beside the point. Rather, the key inquiry is whether the draft Supplier Agreement was understood by the parties to memorialize the terms of their contract. Yet BSH presents no evidence that such a meeting of the minds took place. In light of the fact that Mr. Wortman made Detroit Radiant's acceptance of the Supplier Agreement conditional on BSH agreeing to several suggested changes, the email from Mr. Wortman should properly be viewed as an offer (or counteroffer) that was never accepted by BSH, thus creating no enforceable contract. The mere fact that BSH issued a purchase order the following day and the parties subsequently began to make, sell, and buy the Pro 27 burners does not constitute acceptance of the Supplier Agreement. Rather, it simply indicates that the parties were beginning to perform pursuant to the less formal agreement that they had reached a few days prior. To be sure, terms of the Supplier Agreement may be used to provide evidence of the parties' course of dealing or industry custom in an effort to flesh out the terms of the less formal agreement, but the drafted and unagreed-upon Supplier Agreement is not itself binding on the parties.

Even if paragraph 5 of the Supplier Agreement is to be used, as BSH suggests, as evidence of the parties' course of dealing, it does not change the district court's result. Paragraph 5 indicates that as part of the "purchase order/scheduling agreement," the first month is a committed order, whereas subsequent months are non-binding forecasts. This provides BSH with month-to-month flexibility in its ordering, but it does not necessarily absolve BSH of its responsibility, at some point, to order the entire quantity listed on the initial purchase order which, as here, had no month-to-month specifications. At the very least, then, paragraph 5 of the Supplier Agreement is ambiguous, and thus it does not support the non-binding forecast theory as strongly as BSH would have us believe. This only begs the question, discussed below, of what was really meant by the purchase orders issued by BSH.

In its third claim, BSH takes issue with the district court's holding that "[n]o ambiguity exists as to the quantities in the two purchase orders and nothing in them states or intimates that the quantity terms are forecasts or estimates." D. Ct. Op. at 4. BSH claims that the district court improperly used what it viewed as clear language in the purchase orders to bar consideration of any supplemental evidence derived from trade usage or course of performance. See M.C.L. § 440.2202 (stating that even an integrated final agreement between parties "may be explained or supplemented . . . by course of dealing or usage of trade or by course of performance"). In particular, B SH claims that the district court ignored testimony that a "BLANKET" purchase order, the type of order used by BSH in August 2001, has a particular trade meaning: namely, a non-binding estimate of the buyer's requirements. As one court has explained:

A blanket purchase order does not oblige [the seller] to manufacture or ship any parts. That obligation arises when [the buyer] issues what is known as a shipment, production, or release order that would issue against the blanket purchase order. Blanket purchase orders can last for some time, while shipment orders are issued against them.

Urban Associates, Inc. v. Standex Electronics, Inc., No. 04-40059, 2006 WL 250020, * 1 (E.D.Mich. January 30, 2006).

BSH also argues that paragraph 5 of the draft Supplier Agreement, discussed above, reinforces the interpretation that the August 2001 purchase order was for a non-binding estimate. BSH contends that the whole point of a blanket purchase order is to allow a buyer not to have to commit to purchasing thousands of component parts for a new product, given that the buyer cannot anticipate how well the new product will sell. Because the Pro 27 range was such a new product, BSH contends, asking for flexibility via a blanket purchase order merely aligned it with industry practice.

In light of the parties' prior agreement to contract for at least 30,000 burner units, however, the August 2001 purchase order of 15,000 units and the January 2003 purchase order of 16,000 units are entirely consistent with this total amount. Thus, the fact that the 2001 purchase order contained an explicit quantity term of 15,000 militates against a finding that the order was "blanket" in nature. This is especially true in light of the fact that the 2001 Purchase Order stated: "BLANKET ORDER TERMS/CONDITIONS SEE ATTACHED TERMS AND CONDITIONS," and yet no terms and conditions were attached. Had BSH wished expressly to state that the 15,000 units listed under the purchase order's "quantity" column were an estimate only, then this would have been the place to do it. Even if;industry custom indicates that blanket purchase orders do sometimes contain a fixed quantity term,7 one would think that given the parties' prior communications, a non-binding blanket order would have listed more than 30,000 units, not fewer.

More instructive than industry custom, however, is the parties' own course of dealing which, in this case, does not appear to have a long history. If BSH and Detroit Radiant had conducted prior dealings for other kinds of radiant burners, and if purchase orders had previously been viewed by the two as non-binding estimates (regardless of whether they were "blanket" in nature or not), then BSH would have a stronger claim. But such prior dealings did not exist, or BSH surely would have mentioned them in its brief. Interestingly, the January 2003 purchase order, which, as discussed previously, was styled as a "Schedule Agreement," contained no language to the effect that it was a blanket order, and thus cannot be read to mean anything other than a contract for the full 16,000 units (thus confirming the original agreement of at least 30,000 units). And finally, it is undisputed that in 2004, when BSH began to contract with Solaronics for the same Pro 27 burners, these purchase orders specified a price, but no quantity. Why was there no quantity term? Presumably because B SH was aware that a fixed quantity term on a purchase order, even in conjunction with the words "BLANKET ORDER," might signal its agreement to be bound to pay for the full quantity of properly manufactured goods.

In light of the fixed quantity terms on the two purchase orders, as well as the prior agreement found by the district court, the district court did not err in concluding that evidence of trade usage or course of dealing did not alter BSH's ultimate obligation to purchase 31,000 burner units from Detroit Radiant.

BSH's final claim of error regards the district court's calculation of damages. The Uniform Commercial Code provides two different measures of damages when the buyer — here, BSH — breaches. The default measure of damages is "the difference between the market price at the time and place for tender and the unpaid contract price together with any incidental damages." U.C.C. § 2-708(1); M.C.L. 440.2708(1). The alternative measure of damages, "lost profits," is generally only available to the "lost-volume seller," which Detroit Radiant is not and does not claim to be. See U.C.C. § 2-708(2); M.C.L. 440.2708(2). However, lost profits damages are also available to a plaintiff who cannot adequately recoup under the default measure: "If the measure of damages provided in [M.C.L. 440.2708(1)] is inadequate to put the seller in as good a position as performance would have done then the measure of damages is the profit (including reasonable overhead) which the seller would have made from full performance by the buyer, together with any incidental damages." Id. BSH does not contest the district court's finding that the Pro 27 burners were "specially manufactured" by Detroit Radiant to BSH's specifications. D. Ct. Op. at 3-4. Nevertheless, BSH argues that just because the Pro 27 burners were specially manufactured does not necessarily mean that no reasonably accessible market existed for the burners. In other words, BSH contends that Detroit Radiant failed to prove it could not "cover" the breach, and thus it should not be entitled to lost profits damages. However, it was undisputed at trial that BSH would not allow the burners to be re-sold even if there were a market for them. BSH still considered its Pro 27 burners to be proprietary, and thus did not want Detroit Radiant to sell them to competitors, even in the event BSH breached the contract. Given the fact that it was the express intent of BSH to squelch any potential market for the Pro 27 burner (beyond the "market" of BSH itself, of course), it is disingenuous for BSH to now suggest that Detroit Radiant has failed to offer proof that there was no reasonably ascertainable or accessible market for the burner.

The best argument B SH can muster is that Detroit Radiant did not itself honor the proprietary arrangement, because Detroit Radiant included the Pro 27 burner in a November 2003 product catalog. But even if this is accurate, the record is devoid of any evidence that Detroit Radiant actually sold any of its Pro 27 burners. Detroit Radiant could, for example, have used the listing only to alert potential buyers to the kinds of specialty burners it could make, without revealing any trade secrets as to how the particular burner was made. Even if the catalog listing is to be construed as evidence of Detroit Radiant's desire to sell the Pro 27 burner, rather than simply to advertise what it was capable of manufacturing, such construction would still cut against BSH on the damages issue, because it would help prove that Detroit Radiant offered the burner to all possible buyers and yet none were interested.

Therefore, Detroit Radiant has met its burden to show that it should be granted lost profits damages under M.C.L. 440.2708(2).

The district court opinion was a short one, and we question whether a case of this complexity — if not from a legal standpoint, then at least from a factual standpoint — should have been resolved by the district court in four short pages. Nevertheless, the district court's critical findings and conclusions withstand scrutiny. It was not error for the court to have found that a bargained-for, binding contract existed between Detroit Radiant and BSH. The district judge, who heard testimony from both companies' agents as well as expert witnesses during the course of a bench trial, was in the best position to determine that Detroit Radiant would not have offered a price quote of $28.25 per Pro 27 burner unit, and would not have agreed to absorb the $60,000-$80,000 in tooling and research and development costs for the project, had it not received assurances that BSH intended to purchase at least 30,000 units. The subsequently issued purchase orders of 15,000 and 16,000 units, the only two purchase orders issued during the course of the parties' dealing, strike us — as they did the district court — as entirely consistent with this initial finding. Through these orders, BSH obligated itself to the purchase of 31,000 burner units, an obligation which it subsequently failed to honor. Detroit Radiant was left with a warehouse of burners and component parts that it could not unload, due both to the uniqueness of the Pro 27 burner and to the fact that BSH itself did not want Detroit Radiant to share any secrets as to that burner. And Detroit Radiant was further left without its anticipated profits — i.e., the benefit of the bargain that it had entered into with B SH. Michigan contract law, not to mention common sense, dictates that B SH should pay up, and thus we AFFIRM the judgment of the district court.