Innovative Bites Limited v The Commissioners for HMRC [2026] UKFTT 500 (TC)

[2026] UKFTT 00500 (TC)Case No TC 09831
FIRST-TIER TRIBUNAL
TAX CHAMBER
Venue BirminghamHearing Heard on: 20 March 2026Date Judgment date: 31 March 2026
Appeal reference: TC/2019/06287
Value Added Tax, Zero-Rating, Food, Confectionery, Note 5 Item 2 of Group 1 Schedule 8
Value Added Tax Act 1994. Whether oversized marshmallows normally eaten with the fingers – no. Appeal allowed.
Before
TRIBUNAL JUDGE AMANDA BROWN KCTRIBUNAL JUDGE MATTHEW DONMALLMOHAMMED FAROOQ
Between
INNOVATIVE BITES LIMITEDAppellantTHE COMMISSIONERS FOR HIS MAJESTY’S REVENUE AND CUSTOMSRespondentTim Brown of counsel, instructed by the VAT Consultancy for AppellantCharlotte Brown of counsel, instructed by the General Counsel and Solicitor to HM Revenue and Customs for RespondentsDECISION

Introduction

[1]This appeal concerns whether a food product called “Mega Marshmallows” (Product) is standard-rated. In 2022, the First-tier Tribunal in decision [2022] UKFTT 00352 (TC) (FTT22) determined that the Product was not confectionery in the ordinary sense. However, the Court of Appeal [2025] EWCA Civ 293 (CA25)held that FTT22 had failed to address the specific issue of whether the Product was “sweetened prepared food which is normally eaten with the fingers” and remitted that question to a differently constituted First-tier Tribunal to determine. The parties are agreed that the Product is “sweetened prepared food”, so the narrow issue in the appeal is whether the Product is “normally eaten with the fingers” (the Issue). If it is, then it is standard-rated as HMRC contend; if it is not, then it is zero-rated and the Appellant succeeds.

Legislative background

[2]Supplies of food of a kind used for human consumption are zero-rated for VAT, pursuant to Group 1 Schedule 8 Value Added Tax Act 1994. However, zero-rating does not apply to certain “excepted items”, which are therefore standard-rated. Excepted Item No 2 (Item 2) refers to: Confectionery, not including cakes or biscuits other than biscuits wholly or partly covered with chocolate or some product similar in taste and appearance.[3]Note 5 Group 1 (Note 5)further provides as follows: … for the purposes of item 2 of the excepted items ‘confectionery’ includes chocolates, sweets and biscuits; drained, glacé or crystallised fruits; and any item of sweetened prepared food which is normally eaten with the fingers.[4]Section 96(9) of VATA 1994 provides that Schedule 8 shall be interpreted in accordance with the notes contained in that Schedule. Procedural history The assessment

Procedural history

[5]The Appellant is a wholesaler of the Product to UK retailers. HMRC decided that the Product was confectionery and ought to have been standard-rated and issued assessments totalling £472,928 to the Appellant on 14 August 2019, covering supplies of the Product in VAT periods between June 2015 and June 2019. FTT22[6]The Appellant’s appeal was heard by FTT22 on 14 July 2022. The chief operating officer of the Appellant, Stephen Foster, gave witness evidence and although he was cross-examined, FTT22 found that there was no real dispute in relation to his evidence. On 21 September 2022, FTT22 allowed the appeal, finding that the Product was not “confectionery” for the purposes of Item 2.[7]The primary findings of fact made by FTT22 were set out at [19] to [36], including as follows:19. The Product is an oversized marshmallow which the appellant originally imported from the United States. It is now produced and imported from Belgium. The marshmallows are broadly cylindrical in shape, approximately 5cm in height with a diameter of 3.5 – 4.5 cm. In comparison, regular marshmallows are also cylindrical, 2cm in height with a diameter of 2.5cm. The size of the Product means that it is more easily and more effectively roasted on a skewer over an open fire or flame. It can then be either eaten as a roasted treat, once cooled, or used to create a s’more.20. The Product is sold in various package sizes. We were provided with a 550g pack which contains approximately 27 marshmallows. The printed packaging has varied over the period we are concerned with, which is June 2015 to June 2019. The following is a description of the packaging in the period up to March 2017.21. The front of the packaging has the words “Mega Marshmallows” in large type. The Product is said to be “made to a delicious American recipe” and is said to be “perfect for roasting, s’mores or just snacking”. Below that narrative are three diagrammatical representations from left to right of a marshmallow being roasted over a fire, a s’more and a marshmallow with a bite taken out, representing snacking.22. The reverse of the packaging repeats the diagrammatical representations of a marshmallow being roasted and of a s’more, alongside “Instructions for Use” which are as follows: 1. Stick the marshmallows on a skewer. 2. Keep the stick approx. 20 cm above the heat. Do not hold the mallows in the flames, to avoid burning. 3. Keep on turning the stick, to obtain a caramelised outer skin with a liquid, molten layer underneath. 4. Let the marshmallows cool down. 5. Enjoy your snack.23. Alongside those instructions is the following warning: ATTENTION! 1. Before eating, let the marshmallows cool down. 2. Do not hold them in the fire. 3. When you use a non-electric heating system (e.g. grill), make sure there is always a bottle of water available, to avoid any danger.24. The reverse of the packaging also includes a choking warning in small print which states as follows: Eat one at a time. For children under 6, cut marshmallows into bite sized pieces.25. The reverse of the packaging also contains a description of how to make a s’more as follows: DO YOU WANT S’MORE? A s’more (“some more”) is a traditional campfire treat, very popular in the United States and Canada, consisting of a roasted marshmallow and a piece of chocolate sandwiched between two pieces of graham crackers (biscuits). Try some more! … 29. The appellant is a wholesaler, and sells the Product to UK retailers including Asda, Morrisons, Iceland and The Range. In the period covered by the assessments, the appellant also sold the Product by way of wholesale online and through cash and carry outlets. The Product is available all year round. 30. The appellant sells a large number of other mallow products which are held out for snacking and which are standard rated. Some are seasonally themed for sale at Christmas and Easter. Products which are standard rated include mini marshmallows held out for snacking. These are much smaller than regular marshmallows. The appellant also sells mini marshmallows which are held out for sale as a baking product, including use as a cake decoration, which it has zero rated. 31. We were provided with evidence as to the seasonality of the appellant’s sales of mallow products, excluding those which are seasonally themed. On the basis of that evidence, we find that sales of all types of mallows are higher in the period May to October than at other times of the year. However, sales of the Product show a greater percentage rise in this period than sales of other mallow products. In the years 2019 to 2021, 65% of sales of the Product occurred in the period May to October. In relation to other mallow products, 56% of sales occurred in that period. We infer from the evidence as a whole that the Product is more likely to be consumed in warmer months than other mallow products. This is because it is more likely to be purchased in order to be roasted over a flame. … 33. The Product is typically sold by retailers separately from confectionery and other types of marshmallows. It is generally displayed in the “world foods” section of supermarket aisles, and during the summer months it is generally also displayed in the barbecue section. 34. Mr Foster’s evidence was that roasting the product made it more palatable. That is Mr Foster’s opinion, but we do not share it. Larger marshmallows are equally palatable whether eaten as a snack or after roasting. However, roasting the marshmallows gives them a different texture and flavour. It is easier to roast a larger marshmallow than a regular size marshmallow. Roasting larger marshmallows also gives a different result in terms of the ratio of crisp outer to soft inner mallow. Regular marshmallows would not be as effective to make a s’more because there would not be sufficient soft inner mallow. 35. If a typical consumer wanted to purchase marshmallows for consumption as a sweet snack, then it is more likely that the consumer would purchase regular marshmallows. 36. Overall, we infer that consumers purchasing the Product are more likely to do so in order to roast the marshmallows over an open flame rather than consume them as a snack without roasting. We cannot say to what extent consumers might go on to use the roasted marshmallow as an ingredient in a s’more, although some consumers will do so.[8]At [40] FTT22 observed that the size of the packaging and the Product itself did not suggest that it is intended to be eaten on the go, like a packet of sweets or a smaller packet of regular marshmallows or some mini marshmallows.[9]FTT22 made no finding as to whether the Product was normally eaten with the fingers. At [42], it observed: 42. Clearly if the product is not roasted then it will be eaten with the fingers, perhaps having been cut up for children under 6. However, once roasted and cooled, the Product might be either eaten off the stick or with the fingers. In the circumstances of this product, we do not give particular weight to the means of eating.[10]The approach of FTT22 as explained at [43] was to determine whether the Product was confectionery on its “ordinary meaning”: 43. …Both parties were agreed that we should categorise the Product by reference to the viewpoint of a typical consumer and giving the term confectionery its ordinary meaning. In carrying out that exercise we consider it appropriate to give particular weight to the nature of the Product, the way in which the Product is placed in supermarket aisles, the packaging and marketing of the Product and our finding that most consumers purchasing the Product would do so in order to roast the marshmallows.[11]Having adopted that approach, at [44] FTT22 found that the Product was not confectionery on its ordinary meaning: 44. On balance we accept that the Product does not fall to be described as confectionery. The fact that it is sold and purchased as a product specifically for roasting, the marketing on the packaging of the Product which confirms that purpose, the size of the Product which makes it particularly suitable for roasting and the fact that it is positioned in supermarket aisles in the barbecue section during the summer months when most sales are made and otherwise in the world foods section, leads us to that conclusion.

The Upper Tribunal

[12]HMRC brought an appeal before the Upper Tribunal (UT) against FTT22. HMRC’s appeal was on two grounds, both of which the UT rejected in its decision of 8 April 2024, [2024] UKUT 95 (TCC) (UT24).[13]The first was that the FTT22 erred in its approach to Note 5, because any item that falls within a description in Note 5 is automatically confectionery. Even if the Product was not confectionery in the ordinary sense, it would still be excepted if it was “sweetened prepared food which is normally eaten with the fingers”. The UT disagreed. It considered at [72] that even where a product might fall within a description in Note 5 other factors might lead to a conclusion that the product is not confectionery.[14]The second ground was an Edwards v Bairstow challenge against two factual findings: a) at FTT22 [31], that the Product was more likely to be consumed in the warmer months than other mallow products because it is more likely to be purchased in order to be roasted over a flame, and b) at FTT22 [33] that the Product was generally displayed in the world foods section and in the barbecue section, separately from confectionery and other types of marshmallows. The UT rejected HMRC’s contentions, finding that the conclusions reached by FTT22 were open to it on the evidence.

The Court of Appeal

[15]HMRC appealed on the basis that the UT had erred in law in its interpretation of Note 5 in considering that a product might fall within a description in Note 5 but other factors might still lead to a conclusion that the product is not confectionery. HMRC did not appeal against the UT’s conclusion on the second, Edwards v Bairstow, ground.[16]The CA25 decision was issued on 21 March 2025 concludingthat the UT had erred in its interpretation of Note 5. CA25 considered that save in the case of an absurdity (the example given was chilli chicken bites), if the terms of Note 5 are met that is conclusive in determining that the product under consideration is standard-rated. As such, any product which is “sweetened prepared food which is normally eaten with the fingers”, it is “confectionery” for the purposes of Item 2. Further, the Court of Appeal concluded that FTT22 had not reached any conclusion on this question, at [48]: 48. The Upper Tribunal said in paragraph 91 of its decision that the FTT “seemed unable to conclude what method was more usually or more often used to eat the product”. I do not read the FTT’s decision in that way myself. The FTT explained in paragraph 42 of its decision that it did “not give particular weight to the means of eating”. It did not say that it could not reach a conclusion on how “Mega Marshmallows” are usually eaten or, in particular, whether they are “normally eaten with the fingers”.[17]HMRC had contended that if they succeeded, the matter should be remitted to a differently constituted FTT for it to decide whether the Product falls within Note 5 as “sweetened prepared food which is normally eaten with the fingers”. CA25 agreed to this course. At [49] of the principal judgment of Lord Justice Newey: 49. In the circumstances, the right course must, I think, be to remit to the FTT the question whether “Mega Marshmallows” are “sweetened prepared food which is normally eaten with the fingers” within the meaning of Note (5). To avoid any risk of being perceived as anchored to the earlier decision, the remittal should, as it seems to me, be to a differently constituted FTT. Further, the hearing should be conducted on the basis of the existing written evidence but with the potential for cross-examination and re-examination of any witness who has given evidence relating to the way in which “Mega Marshmallows” are normally eaten.[18]Males LJ and Nugee LJ agreed. For his part, Males LJ added further comment. He noted that there were two routes by which HMRC might have succeeded. The first was to persuade FTT22 that the Product was confectionery in the ordinary sense. But HMRC had failed on this first route before FTT22, and that route was no longer open to them (even though Males LJ himself clearly had reservations on the point). 54. The FTT concluded that “Mega Marshmallows” are not confectionery as that term is ordinarily understood, principally because they are generally roasted after purchase and before consumption. Mr Watkinson accepts, in the light of that finding, that the first route is not open to HMRC in this court and we have therefore heard no argument about it. Speaking for myself, however, I would not endorse the FTT’s conclusion on this issue. It is common ground that an ordinary person would regard ordinary marshmallows as confectionery, and it seems to me that such a person might well consider that it makes no difference that “Mega Marshmallows” are larger than ordinary marshmallows and that they are generally roasted before being eaten.[19]The second route was by establishing that the Product was normally eaten with the fingers. Males LJ agreed that the case would have to be remitted to the FTT to determine this. He expressed his reasoning as follows:55. So far as the second route is concerned, it is common ground that “Mega Marshmallows” are a sweetened product. Accordingly the only issue is whether they are normally eaten with the fingers. That is a question of fact on which the FTT has not made a finding.56. In some cases it will be obvious from the nature of the product whether it is normally eaten with the fingers or in some other way, but that is not the case with this particular product. No doubt “Mega Marshmallows” are sometimes eaten with the fingers and sometimes with a fork or skewer. It is not obvious what kind of evidence might be available – or may have been before the FTT – to prove on the balance of probabilities how they are ‘normally’ eaten. If there is a difficulty, however, that will be the taxpayer’s problem as the burden lies on the taxpayer to prove that the product is not normally eaten with the fingers and therefore falls to be zero-rated.57. As Lord Justice Newey has said, the case will have to be remitted so that the FTT can either make a finding one way or the other or, if appropriate, say that it is unable to do so.[20]The order duly made by the Court of Appeal provided as follows: The case is to be remitted to a freshly constituted First-tier Tribunal (Tax Chamber) tribunal, to consider the question whether “Mega Marshmallows” are “sweetened prepared food which is normally eaten with the fingers” within the meaning of Note (5) to Schedule 8 Part II, Group 1 of the Value Added Tax Act 1994. The remitted hearing is to be conducted on the basis of the existing written evidence, but with the potential for cross-examination and re-examination of any witness who has given evidence relating to the way in which “Mega Marshmallows” are normally eaten.

The hearing of the remitted appeal, evidence and submissions of the parties

[21]We heard the remitted appeal on 20 March 2026. We had a single hearing bundle of 179 pages, which the parties agreed contained all the documentary material originally before FTT22 (and CA25). Additionally after the hearing we were provided with a small further 16 page bundle which had been supplied to FTT22 following that hearing. There was no transcript of the original FTT22 hearing before us.[22]Although CA25’s direction had allowed for the cross-examination and re-examination of Mr Foster on the way in which the Product is normally eaten, Mr Foster was not brought before us to give further evidence in this way. HMRC did not wish to cross-examine him, that decision then precluded re-examination (which necessarily follows from cross examination).[23]As regards the submissions of the parties, we set out below a summary of their respective positions with the aim to do justice to them without unnecessarily lengthening the judgment. We assure the parties to the extent that an argument has not been fully set out it was nevertheless considered when reaching our conclusion. We are grateful to both Counsel for their skeleton arguments and oral submissions, including their willingness to engage with the Tribunal’s questions.[24]The Appellant’s case as developed before us by Mr Brown can be summarised as follows. We should put ourselves in the position of FTT22 applying the findings of fact it made, which were not disturbed by CA25. The nature of the Product, the way it is marketed and the way it is consumed means it is not normally eaten with the fingers. The pack size and packaging format is not aimed at the on-the-go confectionary market, and the Product is more likely to be purchased for roasting. Its placement in specialist supermarket aisles and the seasonal increase in sales during summer for barbecuing is relied upon. The Product is designed for customers to further process them by barbecuing and/or roasting. It was argued by reference to HMRC guidance that items there listed as confectionery are there eaten without any further process being applied prior to consumption. As when roasted the product is heated to a very high temperature, it is not normally eaten with the fingers, but it is normally eaten using a stick or skewer or between two biscuits as a s’more.[25]On behalf of HMRC Ms Brown made the following main points. As CA25 had held that we should not be anchored to the earlier decision, we should not start from the basis of the findings of fact made by FTT22. The burden of proof is upon the Appellant to show that the Product is not normally eaten with the fingers. If the Product is placed into a s’more, that is a means of eating that is with the fingers. There is no evidence before the Tribunal to establish that the Product is not normally eaten with the fingers; the evidence is limited to showing the ways in which the Product may be eaten. She also submitted that a sentence from paragraph 13 of Mr Foster’s witness statement is evidence which supported HMRC’s case that the Product is normally eaten with the fingers if it is roasted.

Preliminary questions

[26]As indicated, the Issue we need to determine is limited to identifying whether the Product “is normally eaten with the fingers”. Before turning to our conclusions on that Issue on the evidence, we address the following preliminary questions. i. The meaning of the provision[27]As regards the meaning of “normally eaten with the fingers”, both parties referred to the judgment of the VAT and Duties Tribunal in (2003) Unibev Limited VAT Decision 18437. That case concerned two ready-to-eat jelly products, which needed unwrapping before consumption. The Tribunal considered that the phrase sweetened prepared food which is normally eaten with the fingers “simply means such food which is not normally eaten with the assistance of a knife and fork or a spoon or some other implement such as chopsticks”, and did not depend on whether items needed to be unwrapped state. The case of Asda Stores Ltd [2009] UKFTT 264 (TC) to which the Appellant also referred is to similar effect. At [36] of that decision, the Tribunal found that the method of packaging of the product, a “Seed Stack Flapjack”, as a bar surrounded by cardboard and cellophane would be regarded by the ordinary customer as a bar which would be normally eaten with the fingers.[28]Neither of these cases involved discussion of the meaning of “normally”. We invited counsel to address us on what “normally” means in this context, where there were multiple ways of eating the product. Mr Brown for the Appellant submitted that as each way would either count as “eaten with the fingers” or not, “normally” means “more often than not”. In other words, to be “normally eaten with the fingers” the Product would need to be more frequently eaten with the fingers than not. Ms Brown for HMRC did not disagree with this approach, rather reiterated the distinct point that there was no evidence before us upon which we could reach that conclusion.[29]We agree that “normally” in “normally eaten with the fingers” requires that Product is more often eaten with the fingers (by one way or another) than not eaten with the fingers. The statutory language is concerned with a binary distinction (eaten with the fingers or not), so “normally” in this context means more often than not, i.e. over 50% of the time. This is consistent how UT23 and CA25 appear to have approached the question, albeit without their expressly addressing it, given language used in discussion at UT23 [91] (“more usually or more often”) and at CA25 [48] (“usually eaten”). ii. The proper approach to the Issue[30]We consider that the nature of the analysis that we have to undertake is in essence that discussed by the Court of Appeal in Procter & Gamble UK v HMRC [2009] EWCA Civ 407: whether the Product is “normally eaten with the fingers” is a statutory test which may require a multi-factorial assessment based on a number of primary facts [9]. We further note Jacob LJ’s observation at [14] that “This sort of question—a matter of classification—is not one calling for or justifying over-elaborate, almost mind-numbing, legal analysis. It is a short practical question calling for a short practical answer.”. Our approach is therefore to undertake a multi-factorial assessment on all the evidence, addressed to the relevant statutory test – in this case, whether the Product is “normally eaten with the fingers”. iii. Applicable principles where a party relies on the burden of proof to resolve a dispute[31]The central argument of HMRC before us was that the burden is on the Appellant to prove that the Product are not normally eaten with the fingers, and if the Appellant has not satisfied this burden, then we must dismiss its appeal.[32]As regards upon whom the burden of proof falls in this appeal, we note that in the passage at CA25 [56] cited above, Males LJ observed that “the burden lies on the taxpayer to prove that the product is not normally eaten with the fingers and therefore falls to be zero-rated”. Newey LJ at CA25 [47] indicated similarly (“The conclusions I have arrived at earlier in this judgment mean that, had the FTT decided that [the Appellant] had not discharged that burden, it should have dismissed [the Appellant’s] appeal to it.”) It was unclear to us whether these observations in CA25 were made having heard considered argument about the nature of the burden of proof in the present circumstances, and it did not appear to us to be a question free from difficulty. The corollary of the burden being on the Appellant in the way articulated is that were we to conclude, adopting that approach, that the Appellant had not proved that the Product was not normally eaten with the fingers, then the Product would be treated as standard-rated under Item 2 read with Note 5. Given the undisturbed finding of FTT22 that the Product is not confectionary in the ordinary sense, that would, in effect, mean that we had thereby necessarily concluded that the Product was “normally eaten with the fingers”, even absent our making a positive finding to that effect, and absent anything in the statutory scheme suggesting that this must be the default treatment of the Product. For that reason, we invited counsel for the parties to address us on this point during the hearing. Both counsel, including Mr Brown for the Appellant, submitted that this being the Appellant’s appeal, the burden fell on the Appellant, essentially as Males LJ indicated.[33]However, we note that the question on whom the burden of proof falls only becomes material, in terms of the resolution of a disputed issue, if a court is unable otherwise to come to a rational conclusion on that dispute either way. This is a well-established principle, discussed by the Court of Appeal on numerous occasions, including in two cases to which we drew the parties’ attention: Stevens v Cannon [2005] EWCA Civ 222 and Barnett v Medway NHS Foundation Trust [2017] EWCA Civ 235.[34]Stevens v Cannon specifically concerned the circumstances in which a court is entitled to despatch a disputed issue of fact by resort to the burden of proof. The Court of Appeal identified the following propositions, at [46]:(a) The situation in which the court finds itself before it can despatch a disputed issue by resort to the burden of proof has to be exceptional.(b) Nevertheless the issue does not have to be of any particular type. A legitimate state of agnosticism can logically arise following enquiry into any type of disputed issue. It may be more likely to arise following an enquiry into, for example, the identity of the aggressor in an unwitnessed fight; but it can arise even after an enquiry, aided by good experts, into, for example, the cause of the sinking of a ship.(c) The exceptional situation which entitles the court to resort to the burden of proof is that, notwithstanding that it has striven to do so, it cannot reasonably make a finding in relation to a disputed issue.(d) A court which resorts to the burden of proof must ensure that others can discern that it has striven to make a finding in relation to a disputed issue and can understand the reasons why it has concluded that it cannot do so. The parties must be able to discern the court’s endeavour and to understand its reasons in order to be able to perceive why they have won and lost. An appellate court must also be able to do so because otherwise it will not be able to accept that the court below was in the exceptional situation of being entitled to resort to the burden of proof.(e) In a few cases the fact of the endeavour and the reasons for the conclusion will readily be inferred from the circumstances and so there will be no need for the court to demonstrate the endeavour and to explain the reasons in any detail in its judgment. In most cases, however, a more detailed demonstration and explanation in judgment will be necessary.[35]In Barnett at [35], subsequent passages from the judgment of Auld LJ in Verlander v Devon Waste Management [2007] EWCA Civ 835 were cited, including the following:
“When this court in Stephens v Cannon used the word “exceptional” as a seeming qualification for resort by a tribunal to the burden of proof, it meant no more than that such resort is only necessary where on the available evidence, conflicting and/or uncertain and/or falling short of proof, there is nothing left but to conclude that the claimant has not proved his case. The burden of proof remains part of our law and practice -- and a respectable and useful part at that -- where a tribunal cannot on the state of the evidence before it rationally decide one way or the other.”
[36]In short, it is incumbent upon us, as the tribunal determining the Issue, which is solely a question of fact, to strive to make a rational finding in relation to that dispute. It is only if, having striven to do so, we cannot rationally decide the case one way or the other that we may despatch the Issue on the basis of who bears the burden of proof.[37]In the event, for the reasons that follow, we have concluded that we can rationally decide the Issue without resort to the burden of proof. For that reason, we do not discuss further any reservation that we may have about whether the burden of proof does indeed fall on the Appellant to show that the Product is not normally eaten with the fingers is immaterial, because we do not determine the Issue on that basis. iv. Our approach to the findings of fact made by

FTT22

[38]We are not persuaded by the respective submissions of either party as regards to the status of the findings of fact made by FTT22. On the one hand, we do not consider that we were necessarily obliged to adopt the findings reached by FTT22, as the Appellant’s skeleton rather suggested; had there been further evidence given before us, the situation may have arisen where we would have had to weigh that evidence as against existing evidence, with the potential of departing in some way from FTT22 factual findings. On the other hand, we also do not consider we are obliged to disregard the FTT22 factual findings. The decision of FTT22 that the Product is not confectionery in the ordinary sense remains good and is predicated on the findings of fact that FTT22 reached. Further, none of those factual findings of FTT22 were impugned by UT23 or CA25, and the reason that this appeal was remitted by CA25 was because there was an absence of any finding by FTT22 on the Issue, whether the Product is normally eaten with the fingers, rather than any problem with any of the facts that the FTT22 had found.[39]In short, CA25 observed that we were not to be anchored by FTT22, and we proceed on that basis. We do not consider ourselves bound by the findings of fact reached by FTT22. However, as it transpired, neither party sought to introduce new evidence before us, by way of cross-examination or re-examination or otherwise, and so we see no reason why we should depart from any of the FTT22’s findings of fact actually made and undisturbed on appeal.

The Issue: Discussion

[40]Both parties accepted that there were four ways of eating the Product on the evidence, consistent with the findings at FTT22 [36] and [42].(1) Roasted on a skewer / stick (collectively skewer) and eaten from the skewer (Way A).(2) Roasted on a skewer, taken off the skewer after it has sufficiently cooled and eaten with the fingers (Way B).(3) Roasted on a skewer, inserted in the middle of two biscuits with a piece of chocolate and eaten as a s’more (Way C).(4) Eaten straight from the pack with the fingers (Way D).[41]It was further agreed that Way A, eating from the skewer, was not eating with the fingers, whereas Ways B and D did constitute eating with the fingers. There was a dispute as to Way C which we deal with next: as HMRC contended that consuming the Product after putting it in a s’more constituted “eating with the fingers”, whereas the Appellant submitted the opposite[42]In our judgment, putting the roasted Product with a piece of chocolate and sandwiching it between two biscuits as a s’more, and then holding the s’more and consuming it, does not constitute eating the Product “with the fingers”. We reach this conclusion on either of two analyses, each of which would be sufficient in itself.(1) First, it is not the Product itself which is being held and eaten with the fingers, but the biscuits. The evidence presented to FTT22 and thereby to us demonstrates that after roasting, the Product has a caramelised outer skin and is liquid and molten underneath. We agree with FTT22 at [34] that the larger size of the Product over regular marshmallows is better for making s’mores because there is more of the soft inner mallow, which is to say that there is more residual heat from the hot molten liquid inner marshmallow to melt the chocolate in the biscuit sandwich. When eating a s’more, a person does not eat the (largely molten) marshmallow with the fingers; the biscuits are in effect implements by which it is eaten.(2) An alternative analysis, arriving at the same conclusion, is that when the roasted Product is placed into the chocolate and biscuit sandwich to form a s’more, it is being used as an ingredient in the s’more, and thereafter it is not the Product that it is being eaten, but the s’more. In the same way, take the example of a burger, where tomato ketchup has been put between the beef patty and the bun. The burger is then held by way of the burger bun as it is eaten. We do not consider that this would constitute eating the tomato ketchup per se “with the fingers”, rather, it would be eating the burger as a whole with the fingers, which is a different thing.[43]As a result, we conclude that two of the four ways of eating, Ways A and C, do not constitute eating “with the fingers”, whereas Ways B and D do constitute eating with the fingers. The Issue is therefore a dispute as to whether the Product is more often eaten by Ways A and C or by Ways B and D. That can be expressed mathematically as the question: is (A + C) > (B + D).[44]HMRC’s fundamental contention is that there is no evidence on the relative frequency of the use of each of the four ways of eating of the Product, and so there is no way to resolve this question and the appeal must fail as the Appellant cannot discharge the burden of proof upon it.[45]We do not agree. While the evidence before us is not as extensive or helpful as it might ideally have been, per Cannon v Stevens it is incumbent upon us to strive to resolve the dispute on the facts rather than resort to the burden of proof, and there is sufficient material before us to decide the appeal rationally on the evidence.[46]First, we find that when roasted on a skewer, the Product is more likely to be eaten directly from the skewer than taken off the skewer and eaten with the fingers. In other words, Way A is more frequent than Way B. This is for the following reasons:(1) When roasted, the Product becomes molten liquid underneath the caramelised outer skin.(2) Consequently, after roasting the Product is not like it is when in its unroasted state, where it can be easily held in the fingers. We infer that when roasted, it would lose sufficient structural rigidity to be easily held in the fingers.(3) Conversely, there is no such difficulty in eating from the skewer, because the skewer holds the Product in its altered caramelised/molten state, and so this in our judgment is the more likely way to be used.(4) While it may be the case that after roasting, provided the consumer waited for long enough, the roasted Product might cool sufficiently to gain a greater degree of internal coherence, we think it is less likely that a consumer would wait longer than otherwise necessary to consume the roasted Product in this way, compared to eating directly from the skewer.[47]On this point, we were unpersuaded by the argument that HMRC raised with regards to paragraph 13 of Mr Foster’s evidence. That paragraph, so far as is material, stated:
“Further guidance is provided on the back of the product as to the manner in which to eat the roasted marshmallow and in particular… The need to ensure the product has cooled down before eating as roasted marshmallows are incredibly hot and will burn fingers if eaten straight from roasting”
(emphasis added). The italicised text was not itself from the packaging, which twice states “let the marshmallows cool down”, but is by way of explanation of that instruction. On the face of it, the italicised text is no more than a statement of (obvious) fact: on roasting the marshmallows are so hot that they would burn fingers if eaten straight from roasting. However, counsel for HMRC invited us to draw an inference from the fact that Mr Foster stated “will burn fingers if eaten straight from roasting” rather than “will burn mouth if eaten straight from roasting",as evidence that the roasted Product is more often eaten with the fingers rather than direct from a skewer. We do not consider that it is safe to draw that inference. HMRC have chosen not to question Mr Foster before us as to what he meant by those words, and while we do not know what was covered by way of cross-examination before FTT22, it was not suggested to us that there had been any material questioning or response on this point. We consider that it is also possible that Mr Foster was meaning to make the simple point that the roasted Product was so hot that it would burn fingers, as a way of giving some context to “incredibly hot”, and did not intend to imply that he thought that the roasted Product was more frequently taken eaten with fingers than eaten direct from the skewer. In any case, even if we were to draw the inference as invited (which we do not), Mr Foster’s view on the question of the relative frequency of that mode of eating would a piece of evidence for us to consider as part of the multi factorial exercise, rather than being in any way determinative.[48]Second, we find that the Product is more likely to be roasted and eaten as part of a s’more, than eaten directly from the bag, unroasted. In other words, Way C is more frequent than Way D. Our reasons are as follows:(1) As the FTT22 found at [35], if a typical consumer wanted to purchase marshmallows for consumption as a sweet snack, then it is more likely that the consumer would purchase regular marshmallows rather than the Product.(2) The Product is typically sold by retailers separately from confectionary, in “world foods” section of supermarket aisles and in summer months in the barbecue section, FTT22 [33].(3) The Product is over double the height of regular marshmallows, FTT22 [19], which is to say that it is much bigger than the size of marshmallows generally sold to consumers in confectionery aisles, and we infer from this that the size of them is therefore much greater than the general preference when eating direct from the bag.(4) Conversely, the larger size of the Product is precisely what makes it preferable for making s’mores, FTT22 [34].(5) The packaging of the Product has, in the order of text and in the order of images, read left to right, roasting, s’mores and then, last, “just snacking” / the graphic image of the Product with a bite taken out of it. This order is consistent with the other evidence that eating the Product direct from the packet is the last, or least likely, way in which it would be eaten.[49]For the above reasons, we find that 1) Way A, eating direct from the skewer, is more frequently used than Way B, taking from skewer and eating with fingers; 2) Way C, s’more, is more frequently used than Way D, unroasted from the packet. It therefore follows that in aggregate, we find as a fact that that the Product is more frequently eaten by one of the non-finger ways than by one of the with-the-fingers ways. To put it in mathematical formulation, because A > B and C > D, (A + C) > (B + D).[50]For these reasons, we find that the Product is not normally eaten with the fingers. The Product does not fall within Note 5 and so does not fall within Item 2 for that reason. As FTT22 has already found that the Product does not fall within Item 2 as confectionery in the ordinary sense, the Product does not fall within that exception to the zero-rating of supplies of food of a kind used for human consumption. We therefore allow the Appellant’s appeal.

Right to apply for permission to appeal

[51]This document contains full findings of fact and reasons for the decision. Any party dissatisfied with this decision has a right to apply for permission to appeal against it pursuant to Rule 39 of the Tribunal Procedure (First-tier Tribunal) (Tax Chamber) Rules 2009. The application must be received by this Tribunal not later than 56 days after this decision is sent to that party. The parties are referred to “Guidance to accompany a Decision from the First-tier Tribunal (Tax Chamber)” which accompanies and forms part of this decision notice. Release date: 31 March 2026