"Fashion changes," says Coco Chanel, "but style endures."1 In this securities fraud case, a class of investors alleges, among other things, that Pier 1 Imports, Inc. is a "trend-based fashion retailer" whose inventory carried a significant markdown risk that the company's executives failed to disclose. Because we conclude that Pier 1 operates largely in the sturdier business of style and that the investors failed to adequately plead scienter, we AFFIRM.
I.
Pier 1 Imports, Inc., is a Fort Worth-based retailer that sells home furnishings at more than 1,000 stores in the United States and Canada and through its website. In 2007, in the wake of seven consecutive quarterly losses and total losses of $227 million for that fiscal year, Pier 1 tapped defendant Alex Smith as its new CEO. Smith, working with defendant Charles Turner, Pier 1's CFO, had significantly *428improved the company's financial position by 2009. At that point, however, Pier 1 faced new pressure to respond to consumer demand for on-line shopping. To enter this market, in August 2012, Smith and Turner launched an initiative called "1 Pier 1," which would allow customers to shop on-line and have their purchases either shipped to their homes or picked up without shipping charges at Pier 1's U.S. stores.
The 1 Pier 1 initiative did not go as planned. Pier 1's stock plummeted from $18.57 on April 28, 2014, to $4.75 on December 17, 2015-a drop of almost 75 percent. Disappointed, a class of investors who had purchased Pier 1's stock between April 10, 2014, and December 17, 2015 (the "Class Period"), brought this lawsuit, claiming that Pier 1, Smith, and Turner (collectively, "the company") illegally hid and misrepresented information that, once disclosed, caused this stock-price tailspin.
Specifically, the complaint alleges that, while carrying out the 1 Pier 1 initiative, the company failed to tell investors about significant "markdown risk"-the risk that Pier 1 had so much inventory that it could get rid of it only by lowering prices dramatically. According to the investors, this markdown risk was exacerbated by the fact that Pier 1's "seasonal" and "specialty fashion" products are "subject to changing consumer tastes." The investors allege that Pier 1's distribution channels were so severely "flooded with excess merchandise" that the company had to employ outside labor and third parties to manage it. To keep up, the investors note, Pier 1 made nondiscretionary, capital-improvement expenditures that were almost seven times higher than the yearly average for the ten previous years.
According to the investors' amended complaint, Pier 1 did not start to tell investors about the existence and magnitude of its markdown risk until the company made a series of "partial corrective disclosures" in 2015. On February 10, 2015, Pier 1 announced that it had higher costs because of "unplanned supply chain expenses" and announced the departure of Turner as CFO. In response to these disclosures, the price of Pier 1's stock fell 25 percent overnight-from $16.97 per share on February 10, 2015, to $12.84 per share on February 11, 2015.
The investors allege that, during the months that followed, Pier 1 made additional misrepresentations and omissions, telling investors that Pier 1's inventory complexion was "clean," "healthy," and did "not pose a significant immediate markdown risk." On September 24, 2015, however, Pier 1 announced that its inventory had caused "issues" within its supply chain, that there were "inventory related inefficiencies within the Company's distribution center network," and that it needed to turn to clearance activity to sell off the extra inventory. The investors allege that, in response to these announcements, Pier 1's stock price fell by more than 12 percent-from $8.67 per share on September 24, 2015, to $7.61 per share on September 25, 2015.
The investors further allege that on December 16, 2015-the penultimate day of the Class Period-Pier 1 announced that it would take at least eighteen months before inventory levels would be in line with actual demand. Pier 1's interim CFO, Laura Coffey, also disclosed that only four of the company's six distribution centers were operating with "acceptable levels of efficiencies." The investors allege that, in response to these disclosures, Pier 1's shares again plummeted by 20 percent in one day-from $5.95 per share on December 16, 2015, to $4.75 on December 17, 2015.
The investors filed their complaint in the Northern District of Texas, asserting that *429Pier 1 and its executives violated § 10(b) of the Securities Exchange Act and SEC Rule 10b-5 by failing to disclose Pier 1's significant markdown risk. See 15 U.S.C. § 78j(b) ; 17 C.F.R. § 240.10b-5. Pier 1 moved to dismiss under Federal Rule of Civil Procedure 12(b)(6). The district court granted the motion, finding that the investors failed to plead facts that would support a strong inference of scienter, the required mental state for these claims. The court allowed the investors to amend their complaint to fix this defect.
After the investors filed an amended complaint, Pier 1 again moved to dismiss. This time, the district court dismissed the case with prejudice, concluding that "[a]lthough [the investors] ha[d] made substantial changes to [their] pleading in an attempt to cure the deficiencies identified in [the previous] Order, ... the Amended Complaint still fail[ed] to plead the requisite strong inference of scienter." The investors appealed.
II.
We review de novo a district court's dismissal of a civil complaint. Barrie v. Intervoice-Brite, Inc. , 397 F.3d 249, 254 (5th Cir. 2005). To survive a Rule 12(b)(6) motion to dismiss, a plaintiff's complaint must comply with the familiar Twombly /Iqbal standard, which requires the complaint to contain sufficient factual matter to state a claim for relief that is plausible on its face. See Emps.' Ret. Sys. v. Whole Foods Mkt., Inc. , 905 F.3d 892, 899 (5th Cir. 2018). Because the investors allege securities fraud, their amended complaint must also satisfy Federal Rule of Civil Procedure 9(b), which requires that a plaintiff "state with particularity the circumstances constituting fraud or mistake." Under Rule 9(b), "a plaintiff must 'identify the time, place, and contents of the false representations, as well as the identity of the person making the misrepresentation and what that person obtained thereby.' " Whole Foods , 905 F.3d at 899 (alteration omitted) (quoting Owens v. Jastrow , 789 F.3d 529, 535 (5th Cir. 2015) ).
To state a claim under § 10(b) of the Securities Exchange Act and SEC Rule 10b-5, a plaintiff must allege: (1) a material misrepresentation or omission; (2) scienter (a "wrongful state of mind"); (3) a connection with the purchase or sale of a security; (4) reliance; (5) economic loss; and (6) a "causal connection between the material misrepresentation and the loss." Dura Pharm., Inc. v. Broudo , 544 U.S. 336, 341-42, 125 S.Ct. 1627, 161 L.Ed.2d 577 (2005). To impute liability to Smith and Turner-the alleged "control persons" of Pier 1 under § 20(a) of the Securities Exchange Act-the investors had to show a "primary violation" under § 10(b): If the § 10(b) claim is inadequate, then so is the § 20(a) claim. Southland Sec. Corp. v. INSpire Ins. Sols., Inc. , 365 F.3d 353, 383 (5th Cir. 2004). Allegations under § 10(b) and Rule 10b-5 must comply with the Private Securities Litigation Reform Act (PSLRA), which requires plaintiffs to "specify each statement alleged to have been misleading, the reason or reasons why the statement is misleading, and, if an allegation regarding the statement or omission is made on information and belief, ... state with particularity all facts on which that belief is formed." 15 U.S.C. § 78u-4(b)(1). In pleading scienter, plaintiffs must "state with particularity facts giving rise to a strong inference that the defendant acted with the required state of mind." Id. § 78u-4(b)(2)(A).
III.
We agree with the district court that the investors have failed to plead a "strong inference" of scienter. Scienter is a "mental state embracing intent to deceive, *430manipulate, or defraud." Tellabs, Inc. v. Makor Issues & Rights, Ltd. , 551 U.S. 308, 319, 127 S.Ct. 2499, 168 L.Ed.2d 179 (2007) (quoting Ernst & Ernst v. Hochfelder , 425 U.S. 185, 193 n.12, 96 S.Ct. 1375, 47 L.Ed.2d 668 (1976) ). Both intent and "severe recklessness" are sufficient. Nathenson v. Zonagen, Inc. , 267 F.3d 400, 408-09 (5th Cir. 2001). Plaintiffs properly plead scienter when they allege that a company knowingly or recklessly made statements to the market while aware of facts that, if not disclosed, would render those statements misleading. See Plotkin v. IP Axess Inc. , 407 F.3d 690, 696-97 (5th Cir. 2005).
The investors raise two challenges to the district court's dismissal of their original and amended complaints: (1) that in dismissing their original complaint, the district court improperly analyzed their scienter allegations; and (2) that the scienter allegations in their amended complaint were sufficient to state a securities fraud claim. Neither argument persuades us that the district court erred.
A.
We have explained that a district court "may best make sense of scienter allegations by first looking to the contribution of each individual allegation to a strong inference of scienter, especially in a complicated case[.]" Owens , 789 F.3d at 537. If any "single allegation, standing alone, create[s] a strong inference of scienter," then the court may stop there. Id. If it does not, then the district court must take "a holistic look at all the scienter allegations." Id.
The investors contend that, in evaluating their original complaint, the district court failed to follow this analysis and instead analyzed all of their allegations collectively under Diodes , which holds that, in four "special circumstances," a defendant's corporate title coupled with a severe problem within the company sufficiently alleges scienter. Local 731 I.B. of T. Excavators & Pavers Pension Tr. Fund v. Diodes, Inc. , 810 F.3d 951, 959 (5th Cir. 2016). This was error, the investors argue, because Diodes does not apply here. They also complain that the district court failed to correct this error when it dismissed their amended complaint. We disagree.
The district court's order dismissing the original complaint first concluded that the investors failed to allege scienter because, inter alia , (1) Smith and Turner lacked a motive to commit fraud; (2) Pier 1 "repeatedly disclosed ... [its] increasing inventory, disappointing sales, and the risk of markdowns"; and (3) the investors' proffered "general accounts from confidential witnesses" did not support their scienter allegations. After undertaking this evaluation of the investors' allegations, the district court then assessed whether Diodes applied and found that it did not because none of the four "special circumstances" were present here. This analysis did not run afoul of our directives in Owens.
B.
We turn next to the investors' amended complaint. The investors' scienter theory relies on three categories of allegations: (1) allegations of motive; (2) allegations that Smith and Turner knew Pier 1 had high inventory; and (3) allegations that Smith and Turner knew Pier 1 had significant markdown risk. Each set of allegations falls short of creating the "strong inference" of scienter required under the PSLRA.
1.
"To demonstrate motive, plaintiffs must show 'concrete benefits that could be realized by one or more of the false statements and wrongful nondisclosures *431alleged.' " Ind. Elec. Workers' Pension Tr. Fund IBEW v. Shaw Grp., Inc. , 537 F.3d 527, 543 (5th Cir. 2008) (quoting Phillips v. LCI Int'l, Inc. , 190 F.3d 609, 621 (4th Cir. 1999) ). Motive is a critical-though not essential-aspect of a successful claim for securities fraud: "[A]llegations of motive ... may enhance an inference of scienter[.]" Abrams v. Baker Hughes Inc. , 292 F.3d 424, 434 (5th Cir. 2002). A failure to show motive means that "the strength of the circumstantial evidence of scienter must be correspondingly greater." Neiman v. Bulmahn , 854 F.3d 741, 748 (5th Cir. 2017) (alteration omitted) (quoting R2 Invs. LDC v. Phillips , 401 F.3d 638, 644 (5th Cir. 2005) ).
The investors allege that Smith and Turner had two motives for concealing Pier 1's markdown risk: (1) they "staked their careers" on 1 Pier 1, which drove them to overstate the success of that initiative; and (2) their employment contracts promised them cash bonuses based on Pier 1's earnings before interest, tax, depreciation, and amortization. Neither motive finds support in our precedent.
First, in Abrams , we held that a desire "to protect [one's job in an] executive position[ ]" was not "the type[ ] of motive that support[s] a strong inference of scienter." 292 F.3d at 434 (citing Melder v. URCARCO , 27 F.3d 1097, 1102 (5th Cir. 1994) ). In the absence of "an allegation that the defendants profited from" the alleged fraud, an allegation of motive based on career prospects is insufficient. Id. Therefore, the investors' reliance on Smith's and Turner's instincts for career survival fails to overcome the pleading hurdle.
As for the second motive allegation, we have held that "incentive compensation 'can hardly be the basis on which an allegation of fraud is predicated' " because "the vast majority of corporate executives" receive this type of compensation. Ind. Elec. Workers' , 537 F.3d at 544 (quoting Tuchman v. DSC Commc'ns Corp. , 14 F.3d 1061, 1068 (5th Cir. 1994) ). However, in a limited set of circumstances-when the potential bonus is extremely high and other allegations support an inference of scienter-performance-based compensation can establish motive. See Barrie , 397 F.3d at 261. In Barrie , the defendant received a performance-based bonus that was 175 percent of his base salary, and we held that his compensation package contributed to a strong inference of scienter. Id.
But Barrie is not this case. Here, even the lowest possible performance-based bonuses that Smith and Turner could receive-which were only 11.5 percent and 8 percent of their respective base salaries-proved to be well out of reach: Pier 1's Fiscal Year 2015 earnings before interest, tax, depreciation, and amortization were $60 million lower than the lowest target number in their employment contracts. Although these contracts also included potential bonuses as high as 288 percent and 200 percent of Smith's and Turner's base salaries, Pier 1's earnings were $125 million shy of the target number for those bonuses. Accordingly, because the likelihood that Smith and Turner would actually receive these high-level bonuses was quite small, Barrie does not apply here. We reject the investors' motive allegations as creating any inference of scienter, much less a strong one.
2.
The investors make the following allegations related to Smith's and Turner's knowledge that Pier 1's inventory was high:
(1) Smith made comments at an employee "town hall" meeting in *432March 2014 admitting that the company overbought inventory;
(2) Smith knew of the company's failure to meet its sales goals and the resultant decision not to pay bonuses;
(3) Smith knew about the company's "inventory problems" and, in 2013, directed the company to end the use of temporary storage at stores;
(4) Smith had a conversation with another executive regarding a 1,000-container backlog at Pier 1's Baltimore distribution center;
(5) The company received regular contemporaneous reports on sales figures, inventory, and purchases;
(6) Pier 1 had large inventory backlogs at its stores and distribution centers; and
(7) Smith and Turner made Sarbanes-Oxley certifications of SEC filings that stated that the company made "conservative inventory purchases."
We note at the outset that, even if we were to assume that all of these allegations satisfy the requisite pleading standards, they would not, standing alone, create a strong inference of scienter. After all, the investors do not allege that Smith and Turner misrepresented Pier 1's inventory. To the contrary, the above allegations include several public disclosures of Pier 1's high-inventory problem. Instead, the investors' theory of the case is that Smith and Turner misled the public about Pier 1's ability to offload that excessive inventory without significant markdown risk : Their amended complaint alleges that "Pier 1 repeatedly assured investors that ... its increasing inventory was 'clean' and did not present 'immediate,' 'significant,' or 'substantial' markdown risk."
Knowledge of high inventory does not necessarily equate to knowledge of significant markdown risk-an equally plausible inference is that Smith and Turner reasonably believed they could fix the excessive inventory problem without resorting to markdowns. See Abrams , 292 F.3d at 433 ("[I]nventory write downs ... can easily arise from negligence, oversight or simple mismanagement, none of which rise to the standard necessary to support a securities fraud action."). Thus, even if they have adequately alleged that Smith and Turner knew about Pier 1's high inventory, the investors must still allege facts demonstrating an "intent to deceive" or at least "severe recklessness" relating to Smith's and Turner's failure to disclose Pier 1's markdown risk. See Tellabs , 551 U.S. at 319, 127 S.Ct. 2499 ; Nathenson , 267 F.3d at 409.
The "high inventory" allegations also suffer from other defects that cast doubt on their sufficiency to create a strong inference of scienter under the PSLRA. We briefly address each allegation in turn.
First , the investors point to Smith's comments at an employee town-hall meeting in March 2014-before the Class Period. At this meeting, Smith showed a PowerPoint slide with the phrase, "[w]e became victims of our own ambition." The investors cite meeting attendees who said that this slide referred to overbuying inventory and that Smith admitted that this was a mistake and that it was his "fault." The investors rely on In re Dynegy, Inc. Securities Litigation , where a district court found a strong inference of scienter when the defendant allegedly "held an all-hands meeting" and falsely told employees that trades at issue were made as stress tests of a new energy trading platform. 339 F. Supp. 2d 804, 901 (S.D. Tex. 2004).
This allegation pertains to backward-looking statements about events that took place outside of the Class Period, which *433cannot establish scienter. See Southland , 365 F.3d at 383 ("[F]raud cannot be proved by hindsight."). In addition, as Pier 1 notes, the information Smith shared was apparently distributed to-not concealed from-the investors before and after the meeting. Moreover, In re Dynegy is distinguishable because the investors do not allege that Smith's statements at the meeting were false. Cf. 339 F. Supp. 2d at 901.
Second , the investors allege that Smith saw that the company failed to meet sales goals and, as a result, did not authorize any employee bonuses at any point during the Class Period. According to the investors, Smith himself signed at least one letter informing employees that no bonuses would be paid. However, this bonus information was repeatedly disclosed to the public and therefore does not reveal any "secret" information that Smith and Turner were trying to hide.
Third, the investors allege that a confidential witness reported that "[d]uring at least one inventory review meeting in mid-2014, Smith acknowledged that inventory problems were pervasive." The investors allege that Smith knew that the problem was pervasive because he knew that Pier 1 was resorting to temporary storage units at hundreds of stores. Smith knew about these temporary units, the investors allege, because Smith himself ordered Pier 1 to stop using the temporary storage units in late 2013.
We note first that "courts must discount allegations from confidential sources." Ind. Elec. Workers' , 537 F.3d at 535. Also, Smith's 2013 directive to stop using temporary storage units was made well before the Class Period, which began in April 2014. This allegation is also vague: it shows only that there were amorphous "inventory problems" and does not explain what those problems were.
Fourth , the investors allege that Pier 1's former Director of Distribution and Transportation reported that Smith asked him about almost 1,000 trailers full of inventory parked outside Pier 1's Baltimore distribution center. The investors assert that these containers housed old inventory not available for sale and Pier 1 failed to process the inventory into its distribution network. The investors analogize to City of Pontiac General Employees' Retirement System v. Dell Inc. , 2016 WL 6075540 (W.D. Tex. Sept. 16, 2016), in which the court found that scienter was adequately pleaded because the plaintiffs alleged that the defendant's company had "ballooning inventories and operations deficiencies within its sales divisions ... resulting in stagnant growth, stalling shipments, and stockpiling inventories[.]" Id. at *4. Just as in that case, the investors argue, their complaint alleged that Smith was aware of excessive inventory that was not getting sold quickly enough.
We disagree that City of Pontiac applies here. In that case, the plaintiffs alleged that the defendants "were aware of undisclosed facts that would undermine the accuracy of their forward-looking statements." Id. at *3 (emphasis added). But mere knowledge of an inventory backlog would have been known to anyone who looked at Pier 1's earning statements. As we have explained, the investors do not allege that Pier 1 misstated the amount of its inventory, its costs, or its sales figures. Moreover, this Baltimore-specific allegation says nothing about inventory problems across Pier 1's other five distribution centers and thousands of stores.
Fifth , the investors allege that the company received "numerous daily, weekly, and monthly reports on sales figures, inventory figures, and purchases that would increase inventory." The investors cite employees who either drafted, sent, or saw *434these reports. The reports allegedly "showed the current sales plan and the actual results every quarter, compared inventory to sales," and demonstrated that "extremely aggressive sales goals could not be met" in 2014.
Internal corporate reports can support a strong inference of scienter only when they meet two requirements: (1) the complaint has corroborating details of the reports' contents, authors, and recipients; and (2) the reports are connected to the speaking executive in a persuasive way. Neiman , 854 F.3d at 748. As the district court concluded, the first element is not satisfied in this case because the investors do not allege that any of these reports revealed the information that is relevant here: the existence of significant markdown risk (as opposed to merely high inventory).
Sixth , the investors argue broadly that Pier 1 had a lot of inventory backlogs at stores and distribution centers. For example, the investors allege that Pier 1 "distribution centers were bursting at the seams with inventory," that (before the Class Period) "stores began telling headquarters that they could not take more inventory," and that "2014 Christmas merchandise could not reach stores in time because of inventory backlogs." However, these statements are all from confidential witnesses who do not relate any interaction with Smith or Turner, so we must discount them. See Ind. Elec. Workers' , 537 F.3d at 535. In addition, the investors "fail to tie these statements" to the alleged fraud: like in Indiana Electric Workers' , the fact that Pier 1 had high inventory "does not necessarily lead to the conclusion that" the company intentionally concealed that fact or otherwise failed to disclose significant markdown risk. Id. at 537.
Seventh , the investors allege that Smith's and Turner's Sarbanes-Oxley certifications of SEC filings support a strong inference of scienter. Sarbanes-Oxley certifications support scienter only if there are "facts establishing that the officer who signed the certification had a 'reason to know, or should have suspected, due to the presence of glaring accounting irregularities or other "red flags," that the financial statements contained material misstatements or omissions.' " Id. at 545 (quoting Garfield v. NDC Health Corp. , 466 F.3d 1255, 1266 (11th Cir. 2006) ). As the district court observed, the investors have not alleged these essential facts.
Having evaluated the investors' "high inventory" allegations de novo , we conclude that they fall far short of establishing a strong inference of scienter, either individually or taken together.
3.
We move on to consider the final category of allegations the investors offer: allegations that Smith and Turner knew that Pier 1 faced significant markdown risk. The company argues that the investors have not created a strong inference of scienter because there is an equally plausible inference to be drawn from Smith's and Turner's knowledge of Pier 1's high inventory: that Smith and Turner genuinely and reasonably believed that they could get rid of that excessive inventory by buying less new inventory and selling the existing inventory gradually at market prices. The company suggests that this alternative explanation is more compelling because Pier 1 kept ordering inventory, which would be reasonable if Smith and Turner believed they could sell it fairly quickly, but unreasonable if they knew they could not.
As we explained, to establish scienter, the investors must allege that Smith and Turner knew of significant markdown risk *435and either intentionally or severely recklessly failed to disclose it to investors. See Plotkin , 407 F.3d at 696-97. Accordingly, this final category of allegations attempts to disprove the company's alternative explanation by demonstrating that Smith and Turner affirmatively knew that they could not sell Pier 1's high inventory without marking it down significantly. Specifically, the investors allege the following:
(1) Pier 1's products are all subject to immediate markdown risk because of their seasonal nature;
(2) Pier 1 held clearance sales events beginning in May 2015;
(3) Various other "red flags" put Smith and Turner on notice of Pier 1's markdown risk; and
(4) Item 303 of Regulation S-K required Smith and Turner to disclose "reasonably expect[ed]" markdown risk.
We will address each in turn.
First, the investors allege that "Pier 1's products are particularly subject to markdown risk because Pier 1 is a trend-based fashion retailer" that is subject to the whims of "consumer trends." Once the fashion changes, the investors' theory goes, any inventory that Pier 1 still has can no longer be sold without significantly marking down the price. This logic finds support in the Second Circuit's decision in Novak v. Kasaks , which found scienter when the plaintiffs alleged that women's apparel retailer Ann Taylor sat on years of out-of-style clothing without marking down its inventory value. 216 F.3d 300, 311-12 (2d Cir. 2000). The investors cite the company's own statements in support of this allegation. For example, they emphasize that in December 2016, Smith said that Pier 1's "products reflect current fashion trends." They also highlight the following statement in the company's SEC filings: "The success of the Company's specialty retail business depends largely upon its ability to predict trends in home furnishings consistently and to provide merchandise that satisfies customer demand in a timely manner."2
Pier 1 responds by criticizing this characterization of its business. Specifically, Pier 1 contends that it never describes itself as a "trend-based fashion retailer" subject to markdown risk. Instead, while some of the products are designed to be predictive of trends in home décor, a large percentage of its inventory is, as the district court noted, comprised of "long-standing collections" of "products that do well for [Pier 1] day in and day out." For example, Pier 1 emphasized in its brief and at oral argument that approximately 50 percent of its inventory during the Class Period was "rebuy" goods, such as the company's well-known papasan chair. Pier 1 points out that the investors even acknowledge in their complaint that Pier 1's inventory includes "durable or 'regular rebuy' " inventory.
We are persuaded by Pier 1's arguments: The investors' allegations do not create a "strong inference" that all (or even most) of Pier 1's inventory is so trend driven that it could not be sold without significant markdowns. First, the investors' general allegation that Pier 1 is a "trend-based fashion retailer" is conclusory. Even at the pleading stage, we need not take such statements as true. See *436Cent. Laborers' Pension Fund v. Integrated Elec. Servs., Inc. , 497 F.3d 546, 550 (5th Cir. 2007) ("[W]e do not accept as true conclusory allegations, unwarranted factual inferences, or legal conclusions."). Turning to the specific, non-conclusory facts that the investors offer, Smith's 2016 statement was made outside the Class Period, and the statement in the SEC filings is too vague-it does not say whether all inventory must be sold timely or precisely how timely the inventory needs to be sold. Ultimately, the investors' argument is self-defeating: they must allege both that Pier 1 said that all of its inventory is subject to markdown risk if not sold quickly (to characterize Pier 1 as a trend-driven business) and, in the same breath, that Pier 1 did not say that (to make an allegation of fraud). The investors cannot have it both ways. And the investors have not explained why Pier 1 executives kept ordering more inventory when they supposedly knew deep down that they would not be able to sell it.
Second , the investors allege that Smith and Turner must have known about looming markdown risk because the company published ads announcing "extraordinary markdowns (up to 50% and 70%)" in clearance sales starting in May 2015. However, the investors first referenced these ads in their response to the motion to dismiss their amended complaint-they did not plead these facts. See Lohr v. Gilman , 248 F. Supp. 3d 796, 810 (N.D. Tex. 2017) ("A plaintiff may not amend [its] complaint in [its] response to a motion to dismiss."). In addition, as Pier 1 points out, the investors' reliance on these sales is merely a temporal-proximity argument: the allegation asserts only that Pier 1 said in April 2015 that there was not much markdown risk and then had a sale in May 2015. Temporal proximity is weak circumstantial evidence of fraud. See Coates v. Heartland Wireless Commc'ns, Inc. , 55 F. Supp. 2d 628, 641 n.18 (N.D. Tex. 1999).
Third , the investors point to an array of "red flags" that they allege should have put Smith and Turner on notice of looming markdown risk. Because these "red flags" were set out in an expert report that the district court struck, we cannot consider them. See Munoz , 200 F.3d at 303 (declining to consider evidence that was struck and therefore "not before the district court").
Fourth , the investors allege that Pier 1 had a duty to disclose markdown risk under Item 303 of Regulation S-K, which directs companies, when filing with the SEC, to "[d]escribe any known trends or uncertainties ... that the registrant reasonably expects will have a material ... unfavorable impact on ... revenues." 17 C.F.R. § 229.303(a)(3)(ii). We have never held that Item 303 creates a duty to disclose under the Securities Exchange Act, and other circuits are split. Compare Stratte-McClure v. Morgan Stanley , 776 F.3d 94, 102 (2d Cir. 2015) ("Item 303 imposes the type of duty to speak that can, in appropriate cases, give rise to liability under Section 10(b)."), with In re NVIDIA Corp. Sec. Litig. , 768 F.3d 1046, 1056 (9th Cir. 2014) ("Item 303 does not create a duty to disclose for purposes of Section 10(b) and Rule 10b-5.").
In any event, we need not address this issue because the investors' argument assumes its conclusion: that Smith and Turner "reasonably expect[ed]" that the high inventory ran the risk of significant markdowns. Moreover, as the district court noted, Pier 1 did disclose these facts when it announced in December 2015-at the end of the Class Period-that it would take eighteen months to clear Pier 1's inventory. The investors need other, independent allegations showing that Smith or Turner reasonably expected that excess inventory levels would have a material unfavorable *437impact on revenues at some point prior to the December announcement.
In summary, we conclude that the investors' allegations that Smith and Turner knew that Pier 1 had significant markdown risk fail to create a strong inference of scienter. The investors' scienter theory suffers from defects similar to the appellants' theory in Diodes . As in that case, rather than concealing its inventory problem, Pier 1 "repeatedly alerted investors" that the problem "would affect the company's output" throughout the Class Period. See Diodes , 810 F.3d at 960. Moreover, like in Diodes , Smith's and Turner's conduct belied any attempt to conceal the impact of that problem: were Pier 1 "attempting to conceal" significant markdown risk, continuing to order inventory "would be counterproductive." See id. at 960. Thus, because none of the investors' scienter allegations satisfy the pleading standards of Rule 9(b) and the PSLRA, we hold that they have not adequately alleged a securities fraud claim.3
IV.
For the reasons described, we AFFIRM the district court's judgment of dismissal.