*559 Following a jury trial, Obinna Obiora was convicted of conspiracy to possess with intent to distribute heroin, and was sentenced to 120 months' imprisonment, followed by 36 months of supervised release. On appeal, Obiora claims that a variety of alleged errors undermined the integrity of the jury's verdict and the appropriateness of his sentence. For the following reasons, we affirm.

I.

We first address Obiora's challenge to the sufficiency of the evidence against him. We describe the record relevant to such a challenge in the light most favorable to the jury verdict. See United States v. Burgos-Montes , 786 F.3d 92 , 99 (1st Cir. 2015).

Federal law enforcement officers became aware of Obiora through their investigation of a Boston heroin dealer named Antoine. Agents obtained approval to wiretap six phones associated with Antoine's activities. In several of these intercepted calls, Chukwuma Obiora -- Obinna Obiora's brother -- arranged for Obinna Obiora to supply heroin to Antoine. 1 On October 3, 2015, the day after one of these conversations, a law enforcement agent observed a car registered to Obiora arrive at Antoine's home. Pole camera footage showed a man who resembled Obiora exit the car, embrace Antoine, and then, with Antoine, disappear from view. Shortly thereafter, the man resembling Obiora returned to the car and drove off. Within about twenty minutes, Obiora called Antoine and complained, "What just happened today is not necessary ... we don't need all that." For the next several weeks, Obiora unsuccessfully tried to obtain payment from Antoine, who apparently stiffed Obiora somehow in connection with their October 3 interaction.

The federal government indicted Obiora for a single count of conspiracy to possess with intent to distribute heroin in violation of 21 U.S.C. §§ 846 and 841. Several co-conspirators were indicted for additional drug and gun crimes. At trial, the government's theory was that Obiora and Chukwuma were Antoine's heroin suppliers until Antoine took their heroin without paying on October 3. The jury found Obiora guilty of conspiracy with intent to distribute heroin, and also found him responsible for at least one kilogram of heroin.

On the first day of trial, the district judge informed the parties about "one other thing," as follows:

I read it in the most recent Harvard Law Review that the Sixth Circuit has just upheld one of my colleagues who after a trial goes back to the jury room and asks the jury individually to just write down what they think the sentence should be, and then he uses that as some advice as to how to impose a sentence .... I've been in touch with the judge who has sent me all his information and I propose to do that. You can read about it in the most recent Harvard Law Review.

Nothing more appears to have been said about the matter until Obiora's sentencing hearing, at which the district court announced that it had conducted the jury poll:

I was interested to, in a procedure developed by my colleague, Judge Gwin, in the Northern District of Ohio, where after the verdict was received, he informally *560 asked the jury privately to advise as to what sentences they would impose and then he announces an average and he takes that into account. That procedure has been expressly confirmed in United States v. Collins , 828 F.3d 386 , a Sixth Circuit case, 2016, and it's been written up with approbation in the Harvard Law Review at a note in Volume 130 at Page 793. And I've resolved to follow that procedure and I followed it in this case.

The average of the jury's suggestion is that he should be sentenced to 19.4 years. That of course is higher than constitutionally this Court could sentence him, but I announce it.

The court conducted the poll ex parte and off record. At no point did either party object to the court's administration of the poll or to its consideration of the results.

At an otherwise unremarkable sentencing hearing, the district court observed that the indictment did not charge that the one-kilogram drug amount was foreseeable to Obiora. Therefore, ruled the court, the ten-year mandatory minimum under 21 U.S.C. § 841 (b)(1)(A)(i) could not be constitutionally applied to him. Nevertheless, the court found that Obiora was responsible for one kilogram of heroin, and calculated the Guidelines range based on that amount. The court sentenced Obiora to 120 months' imprisonment, which was the maximum sentence within the guidelines range, to be followed by 36 months of supervised release.

II.

A.

Obiora appeals the denial of his motions for judgment of acquittal based on the insufficiency of the evidence. See Fed. R. Crim. P. 29(A). We review a district court's denial of a Rule 29 motion de novo, appraising the evidence in the light most favorable to the government. See United States v. Appolon , 695 F.3d 44 , 55 (1st Cir. 2012). "The verdict must stand unless the evidence is so scant that a rational factfinder could not conclude that the government proved all the essential elements of the charged crime beyond a reasonable doubt." Id. (quoting United States v. Rodríguez-Vélez , 597 F.3d 32 , 39 (1st Cir. 2010) ).

On appeal, Obiora argues that the government failed to prove that any substance he delivered to Antoine was actually heroin. But the government was under no obligation to do so. Under the federal drug conspiracy statute, "the criminal agreement itself is the actus reus ." United States v. Shabani , 513 U.S. 10 , 16, 115 S.Ct. 382 , 130 L.Ed.2d 225 (1994). The government offered ample evidence, including phone and text exchanges and witness testimony, that could persuade a rational factfinder -- and did persuade the jury -- that Obiora agreed to supply Antoine with heroin. The identity of the substance later delivered is of no consequence in gauging the record support for the conspiracy conviction. See, e.g. , United States v. Díaz-Castro , 752 F.3d 101 , 107 (1st Cir. 2014) (evidence of dealings with fake drugs was sufficient to uphold a conviction for conspiracy to possess with intent to distribute a controlled substance).

B.

Obiora next challenges several of the district court's evidentiary rulings. We generally review the district court's evidentiary decisions for abuse of discretion, see United States v. Amador-Huggins , 799 F.3d 124 , 128 (1st Cir. 2015), except to the extent they turn on an interpretation of law, which we review de novo, see Burgos-Montes , 786 F.3d at 114 . Not all erroneous *561 evidentiary rulings require reversal. "When, as now, an alleged error is not of constitutional dimension, we may affirm a conviction so long as we have 'fair assurance, after pondering all that happened without stripping the erroneous action from the whole, that the judgment was not substantially swayed by the error.' " United States v. Sabean , 885 F.3d 27 , 41 (1st Cir. 2018) (quoting United States v. Melvin , 730 F.3d 29 , 39 (1st Cir. 2013) ). In assessing such rulings and the significance of any error, we view the record "objectively." United States v. Nelson-Rodríguez , 319 F.3d 12 , 23 (1st Cir. 2003).

1.

Obiora claims that the district court erred by admitting certain lay opinion testimony of a cooperating witness named William, who interpreted for the jury several recorded phone calls between Antoine and his associates. William's interpretive testimony was based on personal knowledge he gained during several years in which he bought heroin from Antoine and assisted Antoine's drug trade by mixing heroin with other substances. Some representative examples of William's interpretive testimony follow:

Recorded statements William's interpretations Antoine: "Yo, if you could do "[I]t's clear that it's something? A quick three hundo [referring to] 300 grams of though." heroin." Antoine: "Your man Gritty was on Antoine had robbed Chukwuma of the list" his heroin. Obiora: "I'm not yet on that A brick means one kilogram of level they give me what, like a heroin. brick at a time." Obiora: "I've got one brick. I This refers to one transaction gave you the first 3, ummm, you for 300 grams of heroin, then took another 3 before this 400, another transaction for 300 you remember?" grams, then a transaction for 400 grams. Obiora: "If I can't return it to Obiora was asking Antoine for them, you know that's another payment for the drugs Antoine problem, and I can't get nothing took. else to bring you."

Obiora raises two main objections to William's testimony: (1) William improperly drew conclusions that should have been reserved for the jury; and (2) William's testimony "smuggled in" inadmissible evidence.

The parties dispute whether Obiora properly preserved these objections below. We need not decide whether Obiora's contemporaneous objection that the conversations being interpreted "were in English" and the "words were clear" was sufficient to preserve the argument, because there is no reversible error even under the abuse-of-discretion standard.

William's testimony is properly characterized as lay opinion testimony under Federal Rule of Evidence 701. See United States v. Valbrun , 877 F.3d 440 , 443 (1st Cir. 2017) (testimony of a member of a drug-trafficking ring interpreting recorded phone calls is lay opinion testimony). Rule 701 allows lay opinion testimony that is "(a) rationally based on the witness's perception; (b) helpful to clearly understanding the witness's testimony or to determining a fact in issue; and (c) not *562 based on scientific, technical, or other specialized knowledge within the scope of Rule 702." Although the district court has "considerable discretion" in deciding whether to admit lay opinion testimony, United States v. Valdivia , 680 F.3d 33 , 51 (1st Cir. 2012), "the rule requires exclusion 'where the witness is no better suited than the jury to make the judgment at issue ....' " Valbrun , 877 F.3d at 443 (quoting United States v. Vázquez-Rivera , 665 F.3d 351 , 363 (1st Cir. 2011) ).

Obiora argues that William's testimony "should have been limited to explaining the typical meaning of particular words used by members of Antoine's conspiracy rather than interpreting the overall meaning and import of the conversations between the parties to the above-described phone calls." Reviewing the interpretative testimony, for the most part we see no such neat dichotomy between individual words and overall meaning. In each instance, a peculiar usage of otherwise ordinary words ( e.g. , "do something") combined with jargon ( e.g. , "hundo") generated the meaning of a sentence, which William succinctly proffered. We do acknowledge that in some instances the jurors may well have understood the gist of a call once they knew its subject was heroin, but we see no reason to require William to parse his interpretative testimony word by word as if he were a foreign language dictionary rather than an interpreter of a conversation. After all, this kind of interpretive testimony is helpful not only because the witness can define terms that are unfamiliar to the jury, but also because the witness can "provide needed context to the events that were transpiring." Valbrun , 877 F.3d at 444 .

Obiora's alternative claim that William's interpretive testimony became a way of "smuggling in inadmissible evidence" is similarly unsuccessful. Cf. United States v. Albertelli , 687 F.3d 439 , 447 (1st Cir. 2012). A lay witness may testify based on personal knowledge to the meaning of words used in a conversation to which he was not a party. See, e.g. , United States v. Dunston , 851 F.3d 91 , 97 (1st Cir. 2017) (holding "without serious question" that a law enforcement officer with significant experience in undercover drug investigations was qualified by his personal experiences to testify to the meaning of terms used in the drug trade). And as for William's non-interpretive testimony, there is no indication that William was simply parroting what he had been told by others, rather than relying on his personal knowledge.

2.

Obiora next argues that the district court abused its discretion in admitting various out-of-court statements Antoine made about the October 3 meeting. The district court ruled that Obiora was engaged in a conspiracy with Antoine "up to and including October 3rd when the heroin was taken from him," and that any statements made before that day, and in furtherance of the conspiracy, were therefore admissible under the hearsay exemption for co-conspirator statements. See Fed. R. Evid. 801(d)(2)(E) (statements made by a party's co-conspirator during and in furtherance of a conspiracy are not hearsay). However, the district court recognized that statements made after the co-conspirators' apparent falling out on October 3 were not made during or in furtherance of the conspiracy, and were therefore not admissible under this exemption to the hearsay ban. Nevertheless, the court declined to strike from the record three sets of statements containing Antoine's description of the October 3 meeting because the court was satisfied that the statements were admissible under other hearsay exceptions.

*563 This, Obiora argues, was reversible error.

We need not determine whether the admission of such testimony was an abuse of discretion because any possible error was harmless. The testimony at issue was extraneous. All three sets of contested statements describe Antoine's failure to pay for the heroin obtained at the October 3 meeting. Whether Antoine paid or did not pay for the heroin when Obiora delivered it is irrelevant to the central question of whether Obiora agreed in the first place to distribute heroin to Antoine. Obiora argues that, "while it is true that the fact of the 'robbery' itself and whether or not Antoine paid for the heroin might be collateral, that does not undo the prejudice caused by introducing the statements identifying Chukwuma -- and by implication Obiora -- as the person who delivered heroin to Antoine." However, the jury heard copious other evidence pointing to Obiora as the person who delivered heroin to Antoine -- including recorded conversations in which Chukwuma and Antoine arranged the October 3 transaction; testimony of a detective who observed Obiora's car arrive at Antoine's place of business; surveillance footage capturing a person resembling Obiora exit the car to interact with Antoine; and phone conversations after the transaction in which Obiora demanded payment from Antoine. Given this compelling evidence that Obiora was dealing with Antoine, we are confident that "the judgment was not substantially swayed" by the admission of Antoine's out-of-court statement to that effect. United States v. Meserve , 271 F.3d 314 , 329 (1st Cir. 2001) (quoting Kotteakos v. United States , 328 U.S. 750 , 765, 66 S.Ct. 1239 , 90 L.Ed. 1557 (1946) ).

C.

Confronting Obiora's three claims of sentencing error, we review challenged factual findings for clear error, interpretations and applications of law de novo, and judgment calls for abuse of discretion. See United States v. Nieves-Mercado , 847 F.3d 37 , 42 (1st Cir. 2017). "[T]he linchpin of a reasonable sentence is a plausible sentencing rationale and a defensible result." United States v. Martin , 520 F.3d 87 , 96 (1st Cir. 2008).

1.

Obiora first challenges the district court's jury poll experiment. Our treatment of this challenge rests in large part on the standard of review triggered by counsels' silence below. The district court told counsel, before trial, what the court intended to do. The court's explanation was perhaps a bit short of detail, but nevertheless provided more than enough information to elicit reservations or inquiries. As best we can tell, both counsel decided to roll the dice, apparently gauging the odds to be favorable. The sources the district court referenced indicate that a juror poll could well be expected to produce sentence recommendations less severe than would the Guidelines. See United States v. Collins , 828 F.3d 386 , 388 (6th Cir. 2016) ("With one exception, every juror recommended a sentence less than half of the five-year mandatory minimum accompanying defendant's offenses."); Recent Case, Sixth Circuit Holds That Imposing A Significantly Below-Guidelines Sentence Informed by A Jury Poll Is Not Substantively Unreasonable. -- United States v. Collins, 828 F.3d 386 (6th Cir. 2016) , 130 Harv. L. Rev. 793 , 797 (2016). Further research would have revealed a study, conducted by the sentencing judge in Collins , suggesting that juries tend to recommend sentences significantly below the Guidelines range. See Judge James S. Gwin, *564 Juror Sentiment on Just Punishment: Do the Federal Sentencing Guidelines Reflect Community Values? , 4 Harv. L. & Pol'y Rev. 173 , 187 (2010). So defense counsel in particular had ample reason to withhold any objections that have only surfaced now that the jurors were less merciful than expected. In short, this is an instance of forfeiture, if not outright waiver.

Assuming forfeiture only, we review for plain error. See United States v. Cortés-Medina , 819 F.3d 566 , 569 (1st Cir. 2016). "Plain error review is not appellant-friendly. It 'entails four showings: (1) that an error occurred (2) which was clear or obvious and which not only (3) affected the defendant's substantial rights, but also (4) seriously impaired the fairness, integrity, or public reputation of judicial proceedings.' " Id. (quoting United States v. Duarte , 246 F.3d 56 , 60 (1st Cir. 2001) ).

The government agrees with Obiora that the district court's use of the juror poll was error. We agree and so hold. In so concluding, we do not dispute that innovation has a role in improving the courts' practices. For that reason, we have national and local bodies, like the U.S. Sentencing Commission and the Administrative Office of the U.S. Courts, tasked with considering new ideas and sometimes conducting pilot projects. With greater hands-on experience dealing with jurors and sentencing, trial judges certainly are better positioned than we are to conceive of innovations that may improve the sentencing process. But the ad hoc implementation of any significant innovation, especially off-the-record and ex parte, can leave circuit courts ill-equipped to assess the legality, fairness, and efficiency of the experimental practice. Here, for example, the docket contains no record of the polling. We do not know how the jurors were asked and answered, or even whether the average sentence recommendation was correctly calculated. The parties cannot shed light on the polling procedure, as they were excluded, albeit apparently with their silent acquiescence.

Judging from the scant information available to us, we see many reasons to doubt that any benefit can possibly be gained from considering the results of such a poll in sentencing. There is no indication that the jurors knew much of anything about Obiora's background, history, or relevant characteristics. Yet, arming the jurors with such information would likely result in a contested hearing of some sort, which might not be worth the effort, costs, and risks. Perhaps some type of jury polling might provide information relevant to the work of policymakers like the U.S. Sentencing Commission. See Gwin, supra , at 175-76 (arguing that the Sentencing Commission should sample juror sentencing opinions). But it is quite another thing to say that jurors' opinions on punishment, unaided by context, should be the object of a judge's attention in sentencing a given individual.

We therefore turn to the question of whether the error was sufficiently obvious to satisfy the second prong of plain error review. "With respect to matters of law, an error will not be clear or obvious where the challenged issue of law is unsettled." United States v. Goodhue , 486 F.3d 52 , 57 (1st Cir. 2007). Our court has never spoken to the jurors' role in sentencing in non-capital cases. The Sixth Circuit has actually rejected a challenge (albeit by the government) to consideration of the results of a jury poll in sentencing. See Collins , 828 F.3d at 388-91 . The case law, in short, provides insufficient direction -- much less holdings -- to label the error clear, at least where the poll is taken with counsel's before and after acquiescence. Hence Obiora's plain error challenge fails.

*565 2.

Obiora next contends that the district court clearly erred in determining that one kilogram of heroin was attributable to him. In drug conspiracy cases, the quantity of drugs involved largely determines the guideline sentencing range. See U.S.S.G. § 2D1.1 (sentencing table). In order to achieve procedural reasonableness, a sentencing court must calculate the Guidelines range using a reasonable approximation of the weight of the drugs that are attributable to the defendant. See United States v. Demers , 842 F.3d 8 , 12 (1st Cir. 2016). We review drug quantity calculations for clear error. See United States v. French , 904 F.3d 111 , 123 (1st Cir. 2018). Obiora offers two reasons for finding such error.

First, Obiora argues that the trial court erroneously deemed itself bound by the jury's drug-quantity finding. To be sure, the district court did note that the jury had found beyond a reasonable doubt that a kilogram of heroin was attributable to Obiora. And, on this issue, the jurors did indeed have the relevant information. But contrary to Obiora's representation, the district court recognized that "it's [the court's] responsibility to make the finding as to drug quantity."

Second, Obiora claims there was insufficient evidence to support the district court's finding by a preponderance of the evidence that a kilogram of heroin was attributable to Obiora. When reviewing a district court's drug-quantity determination, "our job is not to see whether there is any view of the evidence that might undercut the district court's finding; it is to see whether there is any evidence in the record to support the finding." UnitedStates v. Kinsella , 622 F.3d 75 , 86 (1st Cir. 2010) (internal quotation marks omitted).

Here, such evidence comes in the form of a recorded phone call, in which Obiora told Antoine, "I've got one brick. I gave you the first 3, ummm, you took another 3 before this 400, you remember?" Antoine responded, "yeah." Drawing on William's testimony that a "brick" means a kilogram of heroin, the government argues that this exchange demonstrates that Obiora and Antoine engaged in three transactions totaling a kilogram of heroin: two for 300 grams each, and one for 400 grams. Now, on appeal, Obiora reads this double reference to 300 grams as "an instance of oral repetition which referred to the same 300 grams." Perhaps, but certainly where the second reference is to "another 3," the district court need not have adopted Obiora's preferred reading, especially when the remaining evidence pointed to a transaction for a round kilogram of heroin.

3.

Obiora argues, finally, that the district court abused its discretion by imposing a harsher sentence on Obiora than it did on his co-defendants who were more culpable. But all of the others pled guilty, and thus provide inapt comparators. See United States v. Ayala-Vasquez , 751 F.3d 1 , 33-34 (1st Cir. 2014).

III.

For the foregoing reasons, we affirm Obiora's conviction and sentence.