OPINION OF THE COURT

BECKER, Circuit Judge.

1. IntRoduction

When a debtor makes a payment to an ordinary unsecured creditor within 90 days before declaring bankruptcy, assuming that a number of other statutory elements are met, the payment will be stigmatized as an avoidable “preference.” See 11 U.S.C.A. §§ 547(b), (f) (1993). If so, the debtor’s estate will be able to recoup the payment, forcing the erstwhile preferred creditor to fend for the remnants of the estate as an equal alongside the rest of the unsecured creditors. A creditor may sidestep that unhappy contingency and bring itself within one of the unavoidable exceptions to the preference statute, however, if it can show that the transfer was (A) incurred in the ordinary course of both the debtor’s and the creditor’s business; (B) made and received in the ordinary course of their respective businesses; and (C) “made according to ordinary business terms.” 11 U.S.C.A. § 547(c)(2) (1993).

In this appeal the parties ask us to decide the meaning of the last of these three elements — that the transfer must have been “made according to ordinary business terms.” Almost all courts of appeals to have explored its meaning have read the phrase broadly to refer to what is ordinary in the trade or business in which the creditor and debtor operate; just one court of appeals has read the term more narrowly, giving the term a construction that refers exclusively to what is “ordinary” between the involved creditor and debtor. 1 The question is, for us, one of first impression. 2 It is also one of great importance, as the requirement at issue strikes a balance between two conflicting policies, each central to the effectuation of the purposes underpinning the preference rule.

On the one hand the preference rule aims to ensure that creditors are treated equitably, both by deterring the failing debtor from treating preferentially its most obstreperous or demanding creditors in an effort to stave off a hard ride into bankruptcy, and by discouraging the creditors from racing to dismember the debtor. On the other hand, the ordinary course exception to the preference rule is formulated to induce creditors to continue dealing with a distressed debtor so as to kindle its chances of survival without a costly detour through, or a humbling ending in, the sticky web of bankruptcy. See H.R.Rep. No. 595, 95th Cong., 1st Sess. 177-78 (1977), reprinted in 1978 U.S.C.C.A.N. 5787, 5963, 6138; Union Bank v. Wolas, — *220 U.S. -, -, 112 S.Ct. 527, 533, 116 L.Ed.2d 514 (1991).

We believe that the Court of Appeals for the Seventh Circuit delivered the best rendering of the text of § 547(c)(2)(C) when it held that ‘“ordinary business terms’ refers to the range of terms that encompasses the practices in which firms similar in some general way to the creditor in question engage, and that only dealings so idiosyncratic as to fall outside that broad range should be deemed extraordinary and therefore outside the scope of subsection C.” In re Tolona Pizza Prods. Corp., 3 F.3d 1029, 1033 (7th Cir.1993) (emphasis in original). We will embellish the Seventh Circuit test, however, with a rule that subsection C countenances a greater departure from that range of terms the longer the pre-insolvency relationship between the debtor and creditor was solidified.

In this case the bankruptcy court rejected creditor/appellant Fiber Lite Corporation’s (“Fiber Lite” or “the creditor”) claim that it fit within the § 547(c)(2) exception and instead entered judgment in favor of the debt- or/appellee Molded Acoustical Products, Inc. (“the debtor”), in the amount of the preference adjusted downward by the amount of “new value” Fiber Lite provided the debtor during the preference period, and the district court affirmed the bankruptcy court’s order in all respects. But neither court applied the interpretation of subsection C which we announce today. We will nonetheless affirm the district court’s order, as, upon review of the record developed in the bankruptcy court, it is clear beyond cavil that Fiber Lite did not sustain its burden of showing that the transfers at stake were made according to ordinary business terms, even under our moderated interpretation of that phrase. That is to say, we are satisfied that even had the bankruptcy court applied the proper standard in its determination of whether the third prong of § 547(c)(2) was met, it would have had to reach the same result as it did using the standard it in fact applied.

2. FACTS AND PROCEDURAL HISTORY

a. Background

Fiber Lite sells, among other things, uncured fiberglass, a product that the debtor, itself a supplier of molded fiberglass products to the automotive industry, purchased for use in its fiberglass molding processes. Fiber Lite and the debtor had been conducting business for about twenty-one months prior to May 26, 1989, the day when the debtor voluntarily filed a petition for reorganization under Chapter 11 of the Bankruptcy Code (the “Code”), 11 U.S.C.A. §§ 101-1330 (1993). 3 Six months into the bankruptcy, the debtor instituted an adversary proceeding against Fiber Lite pursuant to sections 547(b) and 550(a)(1) of the Code, 11 U.S.C.A. §§ 547(b), 550(a)(1) (1993), in an effort to avoid and recover allegedly preferential transfers made by it during the 90-day pre-petition preference period.

The complaint alleged that during that time period the debtor had transferred to Fiber Lite payments in the amount of $552,-389.44, and further alleged that all of that amount constituted an avoidable preference. Throughout the 90-day period preceding the debtor’s filing of its bankruptcy petition, however, Fiber Lite had continued to supply the debtor with fiberglass products on open account. But the debtor had caught up on its overdue payments during this critical period: the parties stipulated (at trial) that the debtor had transferred $451,224.64 to Fiber Lite during the pre-petition preference period, whereas Fiber Lite had provided the debtor with only $269,328.04 in new goods. See In re Molded Acoustical Prods., No. 89-20868T, Order at 3 (Bankr.E.D.Pa.1992) (Adv. No. 89-1077).

Following a bench trial, the bankruptcy court entered judgment in favor of the debt- or and against Fiber Lite, finding inter alia that the debtor was insolvent at the time of the transfers 4 and that Fiber Lite had failed *221 to bring itself within the safe harbor § 647(c)(2) affords. Although the bankruptcy court found that Fiber Lite had met, according to § 547(g), 5 its burden of proving that the transfers in question were made in the ordinary course of each party’s business and in the ordinary course of dealings between the two parties, see § 547(c)(2)(A), (B), the court held that they were not made according to the ordinary terms in the industry generally, see § 547(c)(2)(C). In reaching this conclusion, the court stressed that Fiber Lite’s “industry terms” evidence, which consisted solely of its separate dealings with one other company and that company’s wholly-owned subsidiary, constituted evidence of its dealings with but a single entity other than the debtor, and, consequently, was as a matter of law insufficient evidence to establish an industry standard. See In re Molded Acoustical Prods., No. 89-20868T, Order at 3. 6

On Fiber Lite’s appeal, the district court affirmed the bankruptcy court’s order in a published opinion. See Fiber Lite v. Molded Acoustical Prods., Inc. (In re Molded Acoustical Prods., Inc.), 150 B.R. 608 (E.D.Pa. 1993). The district court agreed with the bankruptcy court that § 547(c)(2)(C) requires the creditor to prove by a preponderance of the evidence that the transfer was in accord with an objectively determinable industry standard. See id. at 615. It reasoned that, in light of Fiber Lite’s failure to introduce any evidence as to industry standards save for its president’s testimony as to its own collection practices with another delinquent customer and its subsidiary, it had failed to establish an industry standard. See id. at 616. Like the bankruptcy court, the district court declined to refine the contours of the “industry standard” provision of the Code, satisfying itself with applying a ease-by-case approach to the problem.

b. Evidence of The Parties’ Relationship

The decisive factual question we must ultimately confront relates to the pattern of the debtor’s payments on the Fiber Lite account. Fiber Lite maintains that:

[t]he business relationship during these ninety (90) days was substantially similar to the relationship as had existed between the parties for the prior twenty-one (21) months. (The total length of the debt- or/Fiber Lite relationship). Fiber Lite at all times during the preference period continued to conduct business as usual with the debtor. Fiber Lite extended “new value” and post-petition transfers in excess of Two Hundred Seventy-seven Thousand Dollars ($277,000.00) to the debtor. The conduct of Fiber Lite was to continue to extend credit as it had previously done, which directly resulted in slowing down the debtor’s slide into bankruptcy. Fiber Lite did not increase, or alter its normal collection activities during this period, and certainly did not request or receive treatment that was “preferential” or favorable in any way. There was no change in the manner of payment, nor did Fiber Lite attempt to gain additional security, or initiate any legal proceedings to collect past due amounts.

*222 Br. of Appellant at 5. Not surprisingly, the debtor paints a less sanguine picture. It portrays the image of a zealous, hounding creditor and characterizes the relationship between the parties as “not congenial[, but] an unduly burdensome relationship where there was enormous creditor pressure placed by Fiber Lite upon the debtor for accelerated payment.” Br. of Appellee at 7.

The bankruptcy court did not enter detailed findings of fact, which might have proved helpful in resolving the somewhat conflicting evidence concerning what the credit terms of the parties were throughout the duration of their relationship. Since we will apply a different legal standard than the bankruptcy court applied, we will look at the evidence contained in the record in the light most favorable to the appellant Fiber Lite and determine whether a reasonable factfinder applying the proper legal standard could have reached a verdict in its favor. 7

The debtor presented the testimony of its president John D’Amico, Sr., while the creditor Fiber Lite presented the testimony of its vice-president of finance Douglas Edward Blasiman. D’Amico testified, and Blasiman corroborated, that the contractual repayment terms throughout the life of the parties’ relationship was 40 or 45 days net (albeit in mid-1988 Fiber Lite sought unsuccessfully to reduce the terms for some of its goods to 30 days). But in fact the debtor never complied with those written terms, and D’Amico conceded on cross-examination that in 1988 the payments on invoices averaged 53 or 58 days net. George Miller, a partner in the accounting firm Miller, Tate & Co., which the bankruptcy court had appointed as the debtor’s accountant, confirmed that number to be 58. 8

During the preference period, Miller testified, the average payment was made 99 days after invoice, although the debtor’s exhibit showed it to be 89 days. Blasiman testified that during the preference period the debtor paid its invoices within a range of 71 to 113 days. Fiber Lite’s exhibit displaying the aging of the debtor’s accounts payable during the preference period clearly demonstrates that the debtor was further in arrears with Fiber Lite than with its other creditors. According to both Blasiman and Miller, moreover, the same month the debtor filed its Chapter 11 petition, a restless Fiber Lite tired of its forbearance and attempted to implement a payment plan to catch the debt- or up on its late payments, though the debtor never complied, or even attempted to comply, with the plan.

In terms of the industry standard, the evidence is sketchy. The debtor proffered some vague but unrefuted evidence that it used the same 40-45 day net terms with another creditor supplying it with similar goods. Fiber Lite essentially exhausted its evidence with a comparison of the debtor’s payment history against that of two other firms to whom it sold similar product portfol *223 ios, Renaissance and YanDresser, one of which was the subsidiary of the other, both of which were delinquent in their payments to Fiber Lite, and both of which eventually also threw in the towel and exercised the bankruptcy option. The bulk of their payments, Fiber Lite sought to prove, also ran from 71 to 113 days. According to the testimony of Blasiman, Fiber Lite’s practice of collecting from Renaissance and VanDresser, which together had purchased 60% of Fiber Lite’s fiberglass product during the preceding two years whence the debtor had purchased only 10%, mirrored the one it employed with the debtor.

As already noted, the only issue we confront in this case is the interpretation to be given § 547(c)(2)(C) of the Code, the debtor not having appealed the rulings that the debt had been incurred in the ordinary course of business of the debtor and the creditor and that the payment had been made and received in the ordinary course of their respective business affairs.

3. “ORDINARY Business Terms”

Neither a perfunctory survey of the bare language of § 547(c)(2), nor a careful, resolute stare, would lead the average reader to an appreciable understanding of what subsection C adds to subsections A and B in § 547(c)(2). Having gleaned little from the text, we next consult J.P. Fyfe, Inc. of Florida v. Bradco Supply Corp., 891 F.2d 66 (3d Cir.1989), where we confronted § 547(c)(2) and held that its three subparts must be read in the conjunctive. Unfortunately, that case does not get us much nearer our elusive destination: while we used that occasion to interpret § 547(c)(2)(B), we did not have the opportunity to construe § 547(c)(2)(C). See supra at 219 n. 2. The sparse legislative history to § 547(e)(2), which tersely reveals that “‘[t]he purpose of the exception is to leave undisturbed normal financing relations, because it does not detract from the general policy of the preference section to discourage unusual action by either the debtor or his creditors during the debtor’s slide into bankruptcy,’ ” J.P. Fyfe, 891 F.2d at 70 (quoting S.Rep. No. 989, 95th Cong., 2d Sess. 88 (1978), reprinted in 1978 U.S.C.C.A.N. 5787, 5874); accord H.R.Rep. No. 595, 95th Cong., 1st Sess. 373 (1977), reprinted in 1978 U.S.C.C.A.N. 5963, 6329, bestows us with precious little assistance as a starting point either.

We have before us, then, a relatively clean slate, and as we observed at the outset, we are greatly swayed by the incisive analysis in Tolona Pizza. Focusing on one of the purposes of the preference statute — the prevention of favoritism toward particular creditors — Judge Posner, writing for the Seventh Circuit, observed that absent some mutual deterrent, the mass of creditors of a shaky entity will fear that one or a few of their number will break ranks and devour the entity’s meager assets. See Tolona Pizza, 3 F.3d at 1032; accord Clark v. Balcor Real Estate Fin., Inc. (In re Meredith Hoffman Partners), 12 F.3d 1549, 1553 (10th Cir.1993). Mutual distrust, the observation continued, will spur a common response, a run on the entity’s resources, thereby undesirably escalating or precipitating the entity’s tumble into bankruptcy. The preference statute, by allowing the estate to recover in full — for the benefit of all the unsecured creditors — assets which a pushy unsecured creditor unilaterally plucked for itself or which a fawning debt- or used in an irregular transaction to advantage a favored creditor, seeks to foreclose these tragedy-of-the-common type responses to a struggling business.

The court of appeals reasoned that “[f]rom this standpoint, ... the most important thing is not that the dealings between the debtor and the allegedly favored creditor conform to some industry norm[,] but that they conform to the norm established by the debtor and the creditor in the period before, preferably well before, the preference period.” 3 F.3d at 1032. But the court was troubled that if this is all that the third subsection of § 547(c)(2) requires, it might seem to add nothing to the first two subsections, which in toto require both that the debt be incurred and the payment be made within the ordinary course of business of both the debtor and the creditor. We agree, and having held in J.P. Fyfe that subsection C must be satisfied separately, recognize that our attention *224 must focus beyond solely what is normal between the debtor and the creditor.

But this does not imply that the creditor must prove the existence of some single, uniform set of industry-wide credit terms, a formidable if not insurmountable obstacle given the great variances in billing practices likely to exist within the set of markets or submarkets which one could plausibly argue comprise the relevant industry. The Seventh Circuit, conscious of this difficulty, eschewed a bright line approach, concluding that:

“ordinary business terms” refers to the range of terms that encompasses the practices in which firms similar in some general way to the creditor in question engage, and that only dealings so idiosyncratic as to fall outside that broad range should be deemed extraordinary and therefore outside the scope of subsection C.

3 F.3d at 1033 (emphasis in original), quoted in Jones v. United Sav. & Loan Ass’n (In re U.S.A. Inns of Eureka Springs, Ark.), 9 F.3d 680, 685 (8th Cir.1993). Preferring to stay true to what scarce legislative history there is, we substitute the word “unusual” for “idiosyncratic” but otherwise adopt Tolona Pizza’s definition.

The Seventh Circuit’s analysis has much to commend it. It affords meaning to subsection C, and yet does not render it so exacting that a creditor is forced — both when soliciting new customers and when embroiled in a preference proceeding — to depend upon information about its competitor’s trade practices, information that the competitors oft will be reluctant to yield and that frequently the creditor will find difficult to obtain. Indeed, the specter of antitrust liability, however remote, would often prove a resolute deterrent to such sharing of information between or among competitors (as well as conformity to one’s competitors’ practices even if such information were shared). We note, too, that the problems of proof can quickly become unmanageable where there are many firms in the industry, for evidence of the practices of a few firms might not satisfactorily establish the industry standards, especially if the debtor puts on contrary evidence (or so the creditor may reasonably anticipate, forcing it to prepare excessively for trial or causing it to settle unfavorably).

There is yet another attraction to the Seventh Circuit’s approach. Bringing to bear a lesson hard learned from antitrust law, we can predict that the identification of the industry whose standards should obtain would many times be problematic. The Seventh Circuit’s approach, and the approach we will employ, though still requiring that the creditors make some showing of an industry standard, is quite accommodating about what the proper industry is, ameliorating this sometimes intractable problem.

But we do not think simply relaxing the breadth of the term “industry” is enough to temper the aforementioned problems, as the parties may simply end up bickering over which relaxed definition of “industry” is the least relaxed, thereby doing too little to inject certainty and predictability into the law. 9 Rather, we think that the duration of the parties’ relationship is logically pertinent to the touchstone of the statutory policies undergirding § 547(c)(2), policies which we must always keep within our sights especially when trying to make sense out of such terse language as is found in § 547(c)(2). Furthermore, we are persuaded that resort to the length of the parties’ relationship will remedy many of the defects otherwise apparent in that section.

The preference provisions are designed not to disturb normal debtor-creditor relationships, but to derail unusual ones which threaten to heighten the likelihood of the debtor filing for bankruptcy at all and, should that contingency materialize, to then disrupt the paramount bankruptcy policy of the equitable treatment of creditors. That much the structure of the statute and its legislative history import. Therefore, when *225 the relationship in question has been cemented long before the onset of insolvency — up through and including the preference period — we should pause and consider carefully before further impairing a creditor whose confident, consistent, ordinary extension of trade credit has given the straitened debtor a fighting chance of sidestepping bankruptcy and continuing in business. Bankruptcy policy, as evidenced by the very existence of § 547(c)(2), is to promote such continuing relationships on level terms, relationships which if encouraged will often help businesses fend off an unwelcome voyage into the labyrinths of a bankruptcy. See O’Neill v. Nestle Libbys P.R., Inc., 729 F.2d 35, 37 (1st Cir.1984) (noting the exception encourages creditors to continue conducting business with struggling creditors); cf. In re Hoffman, 12 F.3d at 1553. On the other hand, where the relationship is of recent origin, a significant departure from credit terms normal to the trade bears the earmarks of favoritism and/or exploitation, and to countenance such behavior could be unfair (or could appear unfair) to the remaining creditors who exhibit the virtue of patience.

With all that said, we adopt the following rule of construction as an aid to resolving these problems: the more cemented (as measured by its duration) the pre-insolvency relationship between the debtor and the creditor, the more the creditor will be allowed to vary its credit terms from the industry norm yet remain within the safe harbor of § 547(c)(2). The likelihood of unfair overreaching by a creditor (to the disadvantage of other creditors) is reduced if the parties sustained the same relationship for a substantial time frame prior to the debtor’s insolvency. After all, if at the starting point of the relationship insolvency was a distant prospect, a trade creditor does not unfairly overreach, impel insolvency, or inequitably advantage itself at other creditors’ expense by tolerating more generous or commanding more stringent repayment schedules than its competitors.

Variations in credit terms within an industry demonstrate that the market is accommodating variances in buyers’ and sellers’ preferences or needs, at least when the parties deal at arms length and brandish equivalent bargaining power, and we do not think that Congress intended to hamper competition in this area more than necessary to accomplish its stated goal of curbing unusual behavior which impels bankruptcy and/or treats creditors inequitably. 10 Rather, it is the potential manipulation of the credit schedules, the threat or initiation of legal action, or other unusual behavior designed to improve the lot of one creditor at the expense of the others at a time when bankruptcy looms on the horizon of an infirm debtor-to-be that invokes the need to subdue a creditor’s predilection toward self-aggrandizing behavior. In short, we think that a trade debt payment made according to longstanding practice between two solvent parties most often does not “prefer” that creditor to the disadvantage of the debtor or other creditors. 11

Although our approach in some respects may resemble the analyses under §§ 547(c)(2)(A) and (B), we think it imports substantial independent significance to § 547(c)(2)(C). When the relationship between the parties is of recent origin, or formed only after or shortly before the debt- or sailed into financially troubled seas, the credit terms will have to endure a rigorous comparison to credit terms used generally in a relevant industry. That is because in that class of eases we lack something better to look at to verify that the creditor is not *226 exploiting the debtor’s precarious position at the brink of bankruptcy so that it may advantage itself to the detriment of other creditors who continue to extend credit within the letter and spirit of the Code, or at the very least to verify that the creditor is refraining from “unusual” action to collect ordinary debts. In other words, in those situations there is no baseline against which to compare the pre-petition transfers at issue to confirm the parties would have reached the same terms absent the looming bankruptcy. See Tolona Pizza, 3 F.3d at 1032 (arguing that in this set of cases the evidentiary value of compliance with industry norms lends great support to the rule).

We find that subsection C under our reading serves a second important function. Even when the debtor/creditor relationship has been well-settled prior to the debtor’s insolvency, should the creditor be unable to fit its terms within the sliding-scale window surrounding the established industry’s norm, the preferential transfer will not be deemed unavoidable by virtue of § 547(c)(2), although the terms of §§ 547(c)(2)(A) & (B) are fulfilled. That is to say, the parties’ longstanding credit terms, although consistent as between them, may depart so grossly from what has been established as the pertinent industry’s norms that they cannot be seriously considered usual and equitable with respect to the other creditors. An example which comes to mind is where the range of time transpiring between invoices and payments has always varied greatly between the parties, because then one cannot feel confident that “late” payments during the insolvency period were unmotivated by unusual creditor pressure or dictated by the debtor’s financial woes (the baseline problem resurfaced). Alternatively, behavior like filing lawsuits or attaching assets may be ordinary between the parties but be unusual in the industry, and so even an attachment or lawsuit “routine” between the parties may fail to fall within the purview of subsection C.

Our interpretation fully comports with the purpose of, and is facially consistent with the text of, the statute. Indeed, our reading is tailored to fulfill the prime statutory directive of avoiding “unusual” arrangements— as the legislative history denotes them— which cast doubt on the integrity of the party’s actions and the equities vis-a-vis other creditors. Moreover, because the Code does not define “ordinary business terms,” nothing prevents the phrase from incorporating as an intelligible, guiding consideration the parties’ longstanding practice and, when unavailable, industry practice in lieu thereof.

It might be said that this rule favors longstanding creditors over new ones, and that as a result a business teetering on insolvency would experience difficulty obtaining trade credit from new sources. But we find it unlikely that a creditor aware of a debtor’s financial instability/insolvency would agree to extend credit in the first place. Should the creditor wish to do so anyway, although we do not wish to discourage the practice, we note that our interpretation only mandates that the creditor ensures its credit terms comport with some reasonably relevant industry’s norms. Were it otherwise, it might appear that the creditor is simply trying to accommodate the debtor’s precarious financial condition or to strike itself a superior deal at the expense of a frail debtor, causing the transfer to lose its flavor of normalcy. That such an appearance of impropriety alone will suffice to invalidate a transfer we think is a decision Congress reached, and one that we will not question.

In sum, we read subsection C as establishing a requirement that a creditor prove that the debtor made its pre-petition preferential transfers in harmony with the range of terms prevailing as some relevant industry’s norms. That is, subsection C allows the creditor considerable latitude in defining what the relevant industry is, and even departures from that relevant industry’s norms which are not so flagrant as to be “unusual” remain ■within subsection C’s protection. In addition, when the parties have had an enduring, steady relationship, one whose terms have not significantly changed during the pre-petition insolvency period, the creditor will be able to depart substantially from the range of terms established under the objective industry standard inquiry and still find a haven in subsection C.

*227 4. Applying the Standahd

We turn now to the application of these principles to the case at bar. They require us to look first at the range of terms on which firms comparable to Fiber Lite on some level provide credit to firms comparable to the debtor on some level. Unfortunately, the record does not help us to define the industry or tell us what the norms are in it. We do have evidence that Fiber Lite dealt with two other firms in the industry much like it did with the debtor. However, we will largely disregard that evidence for two reasons. First, since one of them was the other’s subsidiary, we cannot conclude that the bankruptcy court’s treatment of them as one combined firm for terms of credit policy or that its discounting the evidence accordingly was erroneous. Just as one swallow does not a spring make, one firm does not an industry make (at least not ordinarily; an exceptionally large firm may be an industry unto itself). 12

Second, Fiber Lite admitted that both Renaissance and VanDresser were delinquent in their payments during the critical period when Fiber Lite compared their credit terms to the debtor’s, and additionally that both those firms eventually filed for bankruptcy. These facts standing alone undermine any claim that Fiber Lite’s credit terms with Renaissance and VanDresser were “ordinary.” We think ordinary terms are those which prevail in healthy, not moribund, creditor-debtor relationships. See In re Hoffman, 12 F.3d at 1553 (“Ordinary business terms ... are those used in ‘normal financing relations’: the kinds of terms that creditors and debtors use in ordinary circumstances, when the debtors are healthy.”). But see Jones, 9 F.3d at 685 (holding evidence of a regular practice in the savings and loan industry of working with delinquent customers on special terms sufficient to satisfy subsection C).

Thus, we are faced with the fact that Fiber Lite put on no meaningful evidence on this issue, despite its bearing the burden of proof. We take this as confirming our intuition that 70 to 113 days — the time that Fiber Lite asserts it regularly allowed the debtor and its “two” other customers to retire their accounts receivable — is quite a departure from what we might describe as a “benchmark” industry standard. In fact, Fiber Lite conceded that the written terms of its agreement with the debtor ranged from 40 to 45 days, and given the fact that the evidence showing the debtor to have had equivalent terms with a similar creditor stands uncon-troverted, we are willing to accept this as sufficient evidence of an industry standard, weak though it may be.

Proceeding to step two under the assumption that the evidence established 45 days as the outer limit of an appropriate industry norm, we think it conceivable that the 58 days average obtaining in 1988 between the parties could pass muster. As explained supra at 224, in the second step we look to the length of the parties’ relationship predating the debtor’s insolvency to estimate the size of the customized window surrounding the industry norm which was established in the first step. Here Fiber Lite and the debtor’s relationship, which commenced somewhat over eighteen months before the debtor’s insolvency, was not extremely lengthy, but was of a sufficiently long duration that the relationship is entitled to some leeway, meaning we might approve a not insubstantial departure from the established 45-day industry norm. Under the circumstances the 58-day average might well be within that window, but we need not and do not actually decide that debatable point for Fiber Lite quite clearly fails to mount the third step.

As the third step in our analysis, we inquire whether the relationship remained relatively stable leading into and throughout the insolvency period, the clear answer to which in this case is negative. Payments during the preference period, made in the neighborhood of an average 89 days (with peaks of up to 113 days), greatly exceeded what they had been through the greater part of the Fiber *228 Lite-debtor relationship, when they had averaged 58 days. Additionally, Fiber Lite had tried to place the debtor on a “payment plan,” an act which does not suggest business as usual within the industry but, quite to the contrary, resembles a calculated response to a deteriorating creditor-debtor relationship. 13 Finally, there was evidence that Fiber Lite had become ever more insistent that the debtor make larger payments as a precondition to further shipments of goods; indeed, the debtor despite its insolvency had made substantial inroads on its indebtedness to Fiber Lite during that time frame. 14 Cf. Xtra, Inc. v. Seawinds Ltd. (In re Seawinds Ltd.), 888 F.2d 640, 641 (9th Cir.1989) (holding that a creditor’s use of economic pressure took the debtor’s payment out of § 547(e)(2) protection).

We have without reward scoured the record for evidence of a range of terms and practices in any industry, or even for a pre-insolvency-period established practice between the parties, close to the ones which prevailed during the preference period. The scant evidence of industry terms we found had much shorter payment dates (45 days) than what transpired between these parties for most of the period of their relationship (58 days). But what is clearly the dispositive factor in this case, which allows us to conclude as a matter of law that the payments at issue here were not “made according to ordinary business terms,” is the evidence that the terms dominating throughout the pre-insolvency relationship between the parties (58 days) were far shorter than the preference- (insolvency-) period payment terms (89 days). 15 Moreover, we have before us evidence that Fiber Lite altered its collection practices within the preference period in several ways besides its extension of the debt- or’s repayment period: it attempted to institute a “payment plan” and successfully mounted its pressure on the debtor to increase its payments. This evidence indicates to us that Fiber Lite itself did not deem its relationship with the debtor to be normal at that time.

5. Conclusion

The generalized objective industry standard test that both the bankruptcy and the district courts applied was not equivalent to the one we have just announced. But although we view the evidence in the record of the parties’ relationship through a different lens, we see a similar image. We are satisfied that Fiber Lite has as a matter of law not placed sufficient evidence in this record to support a finding that the trade practices between Fiber Lite and the debtor comported with “ordinary business terms.” See 11 U.S.C.A. §§ 547(g), (c)(2). Hence whatever portion of the payments received by Fiber Lite during the 90-day period prior to the debtor filing bankruptcy not offset by Fiber Lite’s provision of new value, as provided by § 547(c)(4), constitutes a preference. See supra at 221 n. 6. The order of the district court affirming the order of the bankruptcy court will therefore be affirmed.

. Compare In re Hoffman, 12 F.3d 1549, 1553— 1554 (10th Cir.1993) (reading subsection C to require the debtor and creditor to use terms which are normal between debtors and creditors in the industry when the debtor is healthy, but not specifying who makes up the industry), In re Tolona Pizza Prods. Corp., 3 F.3d 1029, 1033 (7th Cir.1993) (adopting the construction requiring proof of compliance with the standard of a relevant, loosely-defined industry), Logan v. Basic Distrib. Corp. (In re Fred Hawes Org.), 957 F.2d 239, 243-44, 245-46 (6th Cir.1992) (same), Jones v. United Sav. & Loan Ass’n (In re U.S.A. Inns of Eureka Springs, Ark.), 9 F.3d 680, 684-85 (8th Cir.1993) (adopting the definition of industry set forth by In re Tolona Pizza, supra) and Mordy v. Chemcarb, Inc: (In re Food Catering & Housing, Inc.), 971 F.2d 396, 398 (9th Cir.1992) (requiring proof of "prevailing business standards”) with Marathon Oil Co. v. Flatau (In re Craig Oil Co.), 785 F.2d 1563, 1565 (11th Cir.1986) (per cu-riam) (applying the same subjective analysis for purposes of both subsections B and C and thus not requiring proof of industry terms). See also WJM, Inc. v. Massachusetts Dep't of Public Welfare, 840 F.2d 996, 1011 (1st Cir.1988) (reading subsection C to impose an independent requirement but not defining what that might be).

. J.P. Fyfe, Inc. of Florida v. Bradco Supply Corp., 891 F.2d 66 (3d Cir.1989) briefly referenced subsection C in footnote 5, see id. at 71 n. 5, but that discussion was pure dicta, for as we explicitly stated, that case dealt solely with subsection B. Only the creditor had appealed, and the creditor had prevailed on the subsection C issue.

. Fiber Lite no doubt thinks it critical to squeeze inside the ordinary course preference exception, as the debtor will most likely have no assets whatsoever to distribute to its unsecured creditors.

. Section 547(f) creates a rebuttable presumption that the debtor was insolvent during the preference period. Congress enacted that section in 1978 to spare debtors the difficult burden of proving insolvency which existed before its en *221 actment. See, e.g., Constructora Maza, Inc. v. Banco de Ponce, 616 F.2d 573, 577 n. 3 (1st Cir.1980).

. Section 547(g) provides:

For the purposes of this section, the trustee has the burden of proving the avoidability of a transfer under subsection (b) of this section, and the creditor or party in interest against whom recovery or avoidance is sought has the burden of proving the nonavoidability of a transfer under subsection (c) of this section.

11 U.S.C.A. § 547(g) (1993).

. Fiber Lite was not entirely unsuccessful in its defense, however: the bankruptcy court, having determined that Fiber Lite had found cover under 11 U.S.C.A. § 547(c)(4), another exception to the preference rule, entered judgment in the lesser amount of $204,612.33.. Id. at 1. Specifically, Fiber Lite was entitled to set off against the stipulated amount of pre-petition transfers ($451,224.64) the stipulated amount of "new value" invoiced to the debtor within the same period prior to the petition filing date which also met the remaining statutory elements ($269,328.04). Fiber Lite was also held entitled to set off the stipulated value of its post-petition transfer of goods from Fiber Lite to the debtor ($18,955.20) from the total avoidable amount of the unauthorized post-petition transfer of cash' — which transfer unintentionally occurred post-petition due to the delay normally attendant to the negotiation of a check — from the debtor to Fiber Lite ($41,-670.93). Id. at 3-4 & nn. 1, 2.

. Although the bankruptcy court’s failure to apply the correct legal standard should normally result in a remand, we need not remand if the record permits of only one resolution of the issue. See A.J. Canfield Co. v. Honickman, 808 F.2d 291, 307 (3d Cir.1986) (quoting Pullman-Standard v. Swint, 456 U.S. 273, 291-92, 102 S.Ct. 1781, 1792, 77 L.Ed.2d 66 (1982)); cf. Tait v. Armor Elevator Co., 958 F.2d 563, 569 (3d Cir.1992) (standard for a directed verdict).

. Blasiman testified, based entirely on data tabulated in a defense exhibit, that 70% of the debt- or's payments in 1988 were late, that in the first half of 1988 the debtor made its average (mean) payment 71 days after invoice, and that in the second half of 1988 the debtor made its average payment 68 days after invoice, numbers which clearly contradict the debtor's evidence. Further relying on the same exhibit, he also tried to corral the range of time in which late payments were made that year and to indicate the distribution of late payments in the time spectrum. But because cross-examination revealed that the exhibit was based on substantially incomplete information (Miller later testified without contradiction that at least 40% of the 1988 invoices were excluded from consideration by Fiber Lite when it constructed the exhibit), a problem apparently brought about by Fiber Lite's technically imprecise wording of a request for the production of documents, the bankruptcy court did not admit the exhibit into evidence, describing it as “unreliable." As the act of incantation does not interject the data from the missing documents into the informational summaries the exhibit contained, we think the testimony derived directly from the faulty exhibit is equally unreliable, and to give one weight but not the other would be an abuse of discretion. Hence we discard that portion of Blasiman's testimony derived from the unreliable exhibit.

. Curiously, at oral argument, both parties, to whose attention we had called Tolona Pizza, urged us to adopt it (although Fiber Lite advocated it only as a fallback to its primary position that the relationship between the parties itself is dispositive). Each obviously thinks that on the same record it will prevail under the Tolona Pizza standard, a signal perhaps that further refinement would be beneficial.

. See, e.g., Deel Rent-a-Car, Inc. v. Levine, 721 F.2d 750, 754 & n. 13 (11th Cir.1983) (recognizing the two central policies of the preference statute to be “preventing a race to the courthouse by creditors" and “ensuring the equal distribution of assets among the creditors”); In re Smith, 966 F.2d 1527, 1535 (7th Cir.) (same), cert. dismissed, - U.S. -, 113 S.Ct. 683, 121 L.Ed.2d 604 (1992); cf. Grover v. Gulino (In re Gulino), 779 F.2d 546, 548-49 (9th Cir.1985) (stating that the two purposes of the preference statute are to treat creditors equivalently and to discourage “secret liens” upon the debtor’s collateral).

. Cf. Tolona Pizza, 3 F.3d at 1033 (noting the absence of evidence that "the dealings were designed to put [the creditor] ahead of other creditors of [the debtor] or that other creditors of [the debtor] would have been surprised to learn that [the creditor] had been so forbearing in its dealings with [the debtor]”).

. The question of the relevant industry is one of fact subject to clearly erroneous review. Cf. Tunis Bros. Co. v. Ford Motor Co., 952 F.2d 715, 722 (3d Cir.1991) (finding of relevant market in antitrust case subject to clearly erroneous review); Fishman v. Wirtz, 807 F.2d 520, 531 (7th Cir.1986) (same). On this note, we think that looking to precepts of antitrust law will often prove helpful to the bankruptcy court in defining the relevant industry when disputed.

. Apparently, in 1988 Fiber Lite requested the debtor to accelerate its payments so as to shorten the time period within which the debtor retired its old invoices, but since the creditor made this request almost a full year before the petition date and since it never affected the parties’ actual relationship, we may safely ignore it.

. Despite this large change in terms, the bankruptcy court (without making particularized findings of fact to aid our understanding) somehow concluded that the preference period payments were made and received in the ordinary course of the parties’ respective businesses. But cf. J.P. Fyfe, 891 F.2d at 71 (holding that the creditor’s institution of a lien procedure it had not utilized before was not in the ordinary course of the parties' business for purposes of subsection B). The debtor not having appealed this issue, we need not comment on the correctness of that court’s conclusion with respect to subsection B.

.Even had the 89-day payment period prevailed throughout the parties’ relationship, in light of their 18-month pre-insolvency relationship, we think it quite probably too radical a departure from the 45-day industry norm to pass muster under step two of our subsection C analysis. But we need not reach that issue here.