The Weekly Grind: AI & Tech News for Cafe Owners - 21 September 2026
Every week, we round up the most interesting AI and technology news that matters for independent cafe and coffee shop owners. No jargon, no hype - just what you need to know and why it matters for your business.
Five weeks and two days to the Budget, and this week's edition is about numbers that appear on someone else's page before they appear on yours. Late on Monday night the BBC carried, and on Tuesday the Welsh Government published, a permanent 30% business rates cut that names cafes in its own first sentence and, subject to Senedd approval, starts in April 2027, eight weeks after Westminster gave pubs 20% and cafes nothing from the same date. On Friday the RAC put diesel under 4p off its all-time record after an 11.72p rise this month, and the place it reaches a cafe is not the pump but the surcharge line on a wholesale invoice. The UK's third-biggest self-serve coffee brand is spending £12m on a new generation of machines for the 900-odd box estate it already runs, with the first new boxes going in today in Northern Ireland and an 80-drink menu with oat as standard. The shelf food inflation index stood still at 1.3% in the month the Food and Drink Federation's own report said the climb would begin, and two forecasters publishing a week apart have now put the peak in 2027. And a Scottish trade body asked around 300 pubs, bars, restaurants and hotels whether weight-loss drugs were having an impact on business, and 16% said visits were already down. In every case the useful number is the twin of the headline on your own page: your rateable value, your surcharge percentage, your cups a day, your chocolate line, your food-and-drink split. Here is where each one lives.
Wales Has Named Cafes in a Permanent 30% Business Rates Cut From April 2027 - Eight Weeks After Westminster Gave Pubs 20% and Cafes Nothing, and UKHospitality Cymru Says 16% of Ratepaying Hospitality Businesses Will Not Feel the Benefit
The BBC carried it late on Monday 14 September and the Welsh Government's own release is dated Tuesday 15 September, and its first sentence is the story: "Pubs, restaurants, cafés, bars, licensed clubs, hotels, gyms, cinemas and other high street favourites across Wales are set to see their rates bills cut by 30% as the Welsh Government backs the businesses people depend on." The mechanics, as the release sets them out: "Eligible small to medium sized properties with a rateable value below £51,000 will attract the permanent support", from 1 April 2027; the eligible list runs to food and drink hospitality venues ("pubs, restaurants, cafés, bars, food courts, licensed clubs and live music venues"), visitor accommodation and leisure venues, and retail shops "will also continue to benefit from this support, which will build on the lower multiplier in place this year"; it is paid for by "a small increase to the higher multiplier, which applies to the highest-value properties in Wales", which the Caerphilly Observer reports the Welsh Government as saying means properties with a rateable value over £100,000; and "Regulations will be brought forward in the autumn and, subject to Senedd approval, will take effect from 1 April 2027." So it is an announcement rather than law, and the release is careful about the one number every operator wants: "The precise values of all multipliers for 2027-2028 will be confirmed as part of budget preparations, following the UK government's Autumn Budget", which is on 28 October. The First Minister, Rhun ap Iorwerth, put the sector in his own words: "This 30% cut to rates for pubs, cafés, gyms, hotels and so many other local favourites is about giving those businesses the confidence to invest, grow and keep serving the communities that rely on them." The Cabinet Minister for Finance, Elin Jones, called it a "significant and permanent change" and said "We've taken a balanced approach, ensuring this support is affordable while protecting the vital contribution business rates make to our public services", and told the BBC the burden "will fall on the bigger businesses, not just in the hospitality sector", amounting to around 1p in every £1.
Now the arithmetic, because the headline flatters it. The 30% is a discount against the multiplier, and Welsh food and drink hospitality already has a 15% temporary relief this year, which the release says the permanent support "will exceed". For a cafe already on the 15%, the step is from 15 to 30, not from nothing to 30, and the two operators who have put cash on it both land in the same place. Phil Newbould, landlord of the Radyr Tap, where the minister pulled two pints for the cameras, told the BBC the cut was "not as huge as it sounds" and would save about £3,000 a year across the two pubs he runs. Becca Thomas, who owns Bacareto, a wine bar on Church Street in Cardiff, told nation.cymru: "The news of this new, additional, and especially permanent support is heartening", estimating it "will save our business around £3k a year", before adding "we do need to acknowledge that the whole sector is under enormous pressure." The one named independent cafe in the coverage is Kindred, a cafe and wine bar open five months in Cardiff city centre, whose owner Oliver Banks told the BBC: "Hopefully it can mean that we have a little bit more wiggle room with what we do with our costs on food and drink, but it also means that it's a bit lighter on people, especially in the current climate of things", and "there's only so much you can do as a business owner, and you can't put that cost all on to the guests". Carolyn Brownell of FOR Cardiff made the case that permanence is the real news, "Businesses cannot plan, recruit or invest confidently when vital support changes from one year to the next", and then the case for proportion: "The additional saving for businesses currently receiving 15% hospitality relief will be valuable, but it will not remove the considerable pressures created by wages, energy, supplies and other operating costs." The Music Venue Trust put it against the scheme everyone remembers: a 30% reduction "does not fully restore the protection previously offered by the 75% relief scheme".
Then the cliff edge, which is where the trade body's numbers come in. UKHospitality Cymru welcomed the announcement on Friday 18 September, noting it had proposed a lower hospitality multiplier in its own manifesto before the Welsh elections, and published its analysis: "84% of ratepaying hospitality businesses will now benefit from the cut", while "16% of ratepaying hospitality businesses, often very significant employers in their local communities, will not feel the benefit", and "With a surcharge being applied to properties with a rateable value over £100,000, more than 500 venues will pay even more on their business rates bills." Its executive director David Chapman: "More than 500 venues, which are significant employers, will see their rates bills rise even further, when they are straining under the same financial pressures. That just doesn't make sense." The word "surcharge" is UKHospitality's, not the government's. The line itself has names on it. Natalie Isaac, co-owner at Forty Four Group, which runs two restaurants and nine bedrooms in Cardiff city centre, told nation.cymru that "the £51k threshold creates a real cliff edge that risks leaving too many businesses on the wrong side of the line", and that one of her sites "falls just above the threshold and will receive nothing". And Cardiff's Tramshed music venue misses out by £4,000, because its rateable value has risen to £55,000, up almost 36% from £40,500 in 2023, which is the revaluation this column went through in April arriving on the wrong side of a new Welsh line. Reform's shadow finance minister in the Senedd, Cai Parry-Jones, attacked the funding rather than the cut, saying the Welsh Government "isn't paying for this rates cut, but is instead passing the bill on to bigger businesses in Wales". A Welsh Government spokesperson called the reduction "a first step".
What this means for you: In July this column ran the Westminster version of this story under the heading "Pubs Get Another 20% Off Their Rates. Cafes Get Nothing. Again." Its line was that the policy "draws a line, and the cafe is on the wrong side of it". That piece was about England and it is still right about England: the 20% cut for pubs, social clubs and live music venues also starts in April 2027 and still excludes you, so if you trade in England nothing on your bill has moved this week. What has moved is the precedent. Our VAT-cut piece in June argued that rates and National Insurance were the levers that would actually reach a small operator, and that pubs getting targeted relief "shows blanket policy isn't the only option, and that targeted relief is perfectly possible when the will is there". Wales is that sentence coming true, with cafes in the first line, and it is the counterexample to hold up when the Budget arrives on 28 October with UKHospitality's ask to push the English retail, hospitality and leisure multiplier discount "towards 20p" on the table. If you trade in Wales, the job this week is a look-up, not a celebration. Find your rateable value on the Valuation Office Agency's site and know which side of £51,000 it lands. If it is under, re-run your occupancy cost ratio on 2027 assumptions; that post already told you to re-baseline on post-April figures because the revaluation lifted rateable values and the last of the relief disappeared at the same time, without a penny of rent moving, and this is the first time that step down has been offered permanently rather than one year at a time. Keep the proportion, though. On our flat white P&L's worked example, £30,000 a year of rent and rates across 60,000 drinks is 50p a cup, and a saving of the size Becca Thomas describes, about £3,000 a year on one site, is roughly 5p of that on our arithmetic, while Phil Newbould's £3,000 across two pubs is nearer 2.5p a site. If it is over £51,000, you are the Forty Four Group site that falls just above the threshold, and Natalie Isaac's call for a tapered system rather than a hard cut-off is the one to get behind; the trade body's own ask is narrower, to exempt hospitality from the surcharge altogether. One housekeeping note: our business rates guide was written about England and did not say so, so we have added a dated Wales note to it today; read its multiplier table as the English picture.
Read the announcement on GOV.WALES ->
Read UKHospitality Cymru's response ->
Read the BBC's report, with the operator reaction ->
Diesel Was 195.32p a Litre on Friday, Under 4p Off Its All-Time Record, After Adding 11.72p So Far in September - and the Fuel Surcharge on Your Wholesale Invoice Is a Percentage of the Order, Not a Pass-Through of the Pump
On Friday 18 September the RAC's head of policy, Simon Williams, said this, as carried by Business Motoring: "Unfortunately, it's looking increasingly likely that the average price of diesel will reach a new all-time high next week as a litre now stands at 195.32p, just under 4p off the record of 199.09p set on 25 June 2022." The week he is describing is this one. As of this morning no new record had been reported, and this sentence may be out of date by the time you read it. The rest of his statement is the pace: "Diesel has risen nearly 12p (11.72p) this month, with 5p being added in the last week alone. The average price of petrol has now topped 172p (172.02p) having gone up more than 10p so far in September and 3.6p in the last week." Petrol was last this high on 17 August 2022. On the forecourt, "The cost of a full tank of petrol for a 55-litre family car is now £94.61 which is £21.55 (30%) more than at the end of February when the Iran War began", and a full diesel tank is £107.43, which Williams puts at £29, or 37%, more than on 28 February. Business Motoring's own copy, outside the quote, had diesel up 0.6p in the previous 24 hours, and its coverage earlier in the week had diesel at 192.86p, the highest since July 2022, and petrol through £1.70 for the first time since August 2022. The week before that, the same outlet had crude past $100 a barrel and then leaping another $6 overnight to almost $108.
The ONS put the same thing into its August inflation release on Wednesday 16 September, and its version is the one to keep: "The average price of petrol rose by 9.1 pence per litre between July and August 2026, compared with a rise of 0.3 pence per litre between July and August 2025", to 161.3p, the highest since November 2022, and "Diesel prices rose by 14.2 pence per litre in August 2026, compared with a rise of 0.8 pence per litre in August 2025", to 181.8p. Motor fuel prices as a whole rose 23.0% in the 12 months to August, against 15.5% the previous month, and "Transport, particularly motor fuels, made the largest upward contribution to the monthly change in both CPIH and CPI annual rates", which is how headline CPI went from 2.9% to 3.1% in a month when food, as the fourth story explains, stood still. The transport division on the CPIH measure went from 3.6% to 4.6%. None of this is a new war. Four weeks ago this column ran wholesale energy at a near four-year high and quoted Cornwall Insight on "the ongoing conflict between the US and Iran" landing on business energy bills. This is the same conflict arriving on a second bill, by a different route.
The politics landed this morning. FairFuelUK, a campaign group with a position rather than a neutral analyst, published analysis carried by Business Motoring before 9am arguing that fuel retailer margins, after removing duty and VAT, rose from 3% in 2021 to 18% in 2026 on petrol and from 3% to 20% on diesel, that Brent crude on its reading remains 20% below its 2022 peak, and that the higher margins add around £10 to a petrol fill and £15 to a diesel one against pre-2021 levels. Its founder Howard Cox: "On 28 October, the Chancellor must deliver an immediate cut in fuel duty, or at the very least guarantee a freeze for the lifetime of this Parliament." That fixes the date that matters. The 5p cut in fuel duty is due to expire at the end of December, and what replaces it is a decision for the Chancellor, John Healey, at the 28 October Budget.
What this means for you: The pump is not where this reaches you. It reaches you on the line our delivery-charges guide lists among the charges that never make the price list: "Fuel surcharges. A percentage added on top, sometimes adjusted monthly, easy to ignore because it is small per line." Note the mechanism, because it is not the one you might assume. A surcharge that is a percentage of the order is not a pass-through of pence per litre; it is a number your supplier chooses and revises, and the guide's action plan already tells you to write every supplier's fuel surcharge on one page, adding that "Most operators have never had this on one page." What is new this week is that the exercise has a deadline. An 11.72p move so far this month is in flight, and suppliers who index their surcharge will revise it over the coming weeks, so the questions to put to your two biggest suppliers before that letter arrives are these: what is the surcharge percentage now, what do you index it to, when was it last revised, and how much notice do I get when it moves. If FairFuelUK's margin analysis is even half right, the second question matters more than it sounds, because a surcharge indexed to the pump passes retailer margin into your invoice along with the crude. We are not going to model what 11.72p does to a cafe's delivered costs, for the reason the 24 August energy edition gave: nobody has published the cafe-scale version, and we are not going to invent one. What we can say is that the usual escape hatch closes at the same time. The cash-and-carry comparison costs the self-collect run in "your own time, in fuel, and in impulse buys", so dearer diesel makes fetching it yourself worse in the same week it makes delivery worse. The habit that catches all of it is the one the milk post describes: keep a delivered cost per litre with carriage and surcharge included, so that "A £1.15 milk with a £4 delivery charge on a 60-litre drop is really £1.22", and a surcharge revision shows this month rather than next spring. And when a monthly-adjusted surcharge hands you three delivered prices in six weeks, the question of which one goes in the recipe has an answer already written.
Read the RAC's statement on Business Motoring ->
Read the ONS August inflation release ->
Read FairFuelUK's analysis on Business Motoring ->
Barista Bar Is Spending £12m on New Machines for the 900-Plus Self-Serve Boxes It Already Runs, First Installs Today in Northern Ireland - Roughly £13,300 a Machine on Our Arithmetic, in an Estate Already Doing About 43 Cups a Machine a Day, From an 80-Drink Menu With Oat as Standard
On Tuesday 8 September Henderson Foodservice, the Belfast wholesaler founded in 1897 that owns the Barista Bar self-serve coffee brand, announced a £12m investment in its next generation of machines as the exclusive convenience, retail and forecourt partner in the UK for Melitta Professional's next-generation coffee system. Business Eye carried the fullest version of the release that day; a revised trade version in C-Talk reached the convenience press on 10 September and Melitta-side coverage, with a different set of quotes, followed in Foodservice Equipment Journal on 14 September. The £12m sits inside £21m over five years. The scale is the point. Barista Bar launched in 2015 and now operates 900 machines, most of them inside Henderson's own Spar, EuroSpar and Vivo stores, World Coffee Portal reports; Henderson's own release puts it at more than 14 million cups a year from more than 900 machines across Northern Ireland, Scotland and England, the revised trade version counts more than 670 locations throughout Northern Ireland and Great Britain, and Foodservice Equipment Journal prints both figures in one piece, so some sites run more than one box. World Coffee Portal, which is Allegra's own outlet, ranks it the UK's third-largest self-serve coffee brand by units, the ranking the trade releases footnote to Allegra's Project Café UK 2025 as "third-largest coffee-to-go brand". And the brand itself is not only a Northern Ireland story. Mark Stewart-Maunder, Henderson's commercial and development director: "Barista Bar has already demonstrated that it can travel successfully outside of Northern Ireland, with over 200 retail stores and forecourts carrying the brand in Scotland and England, and this investment gives us the infrastructure to accelerate that growth and explore further opportunities in Europe." The first machines go in today. The original release has them rolling out "across SPAR, EUROSPAR and ViVO branded stores in Northern Ireland from 21st September", the later trade version says "across the existing estate from 21 September onwards", and no source dates a Great Britain install or an end date.
The menu is the freshest fact. The new machine carries "more than 30 new drinks, with more than 80 drinks on the full menu", and the range "includes hot and iced beverages, oat milk as a standard dairy alternative, decaffeinated options across all espresso-based drinks and trend-led additions to cater to younger consumers such as Matcha, Chai and Iced Chocolate, alongside a range of flavoured syrups". Melitta's side adds three varieties of bean, cow's and oat milk with hot and cold foam, a specially selected decaf, and technical support and digital services. Stewart-Maunder: "This is the biggest investment we have made in the Barista Bar brand and a significant commercial milestone for the business", and, for the shopkeepers who host the boxes, "we have designed a new system to drive sales, repeat shopper visits and incremental basket spend, while giving shoppers the broader, premium menu they now expect from coffee-to-go, right on their doorsteps." In the Henderson release, Melitta's Ricardo Holland called it "a solution that raises the bar for coffee-to-go".
The arithmetic is ours, not the company's, and it needs its caveats. Nobody has said how many of the 900 machines are being replaced, and the release ties the new machines to an investment in "system technology, a refreshed brand design and an expanded drinks platform" as well as the hardware, without saying how much of the £12m is which. With that said: £12m across 900 machines works out at roughly £13,300 a machine, or roughly £17,900 a location across 670; and 14 million cups a year across 900 machines is about 43 cups a machine a day, or about 57 a location. Either denominator is defensible as long as it is named. We will use the machine, because it is the honest unit when some shops have two.
What this means for you: In June this column ran Unity Coffee, a £2m startup from the man who built Costa Express, aiming at more than 500 machines in twelve months, and its advice was "Do not try to win a price war against a box that pays no wages, rent or rates." That holds, and we will not repeat the rest of it. This week's story is the incumbent, not the startup, with 900 boxes already in the ground and £12m of new machines on the way, and it changes two things the June piece did not. The first is range. "We have more choice than a machine" is about to stop being true: eighty drinks, oat as standard, decaf on every espresso drink, matcha and chai, in a box in a Spar, starting in Northern Ireland today and reaching the rest of the estate on a timetable nobody has published. The wrong response is to bolt thirty drinks onto your own board. Our smaller-menus post opens with the trap: "More choice should mean more sales. It makes intuitive sense. But it's almost always wrong." The right response is to know which of your lines earn their place, and to cost any matcha or chai you are tempted to add before it reaches the board. Oat as standard is the sharpest detail in the release, because our oat milk post already noted that "several of the big chains quietly dropped their non-dairy surcharge a while back and made a point of saying so", and priced absorbing the difference at about £4,800 a year at one site in its worked example. A self-serve box carrying oat as a standard option, rolling out to more than 670 shops and forecourts, adds to that pressure; the post's advice stands, which is that absorbing is a fine decision as long as you have seen the number first. The second thing is volume, and our own corpus supplies the other half of the comparison. The flat white P&L lands a £4 flat white at "maybe 50-70p net per cup" and concludes that to pay yourself £35,000 "you need to sell something like 60,000-70,000 cups a year. That's 200 cups a day, every day, with no holidays." On our division, that box is doing about 43 cups a day against roughly £13,300 of kit. Same cup, same street, a fifth of the traffic, which is why these machines keep appearing where a cafe could not survive, and why the fight was never for the forecourt's customer. One update to our own guide: the April AI piece says "a robot barista costs more than your annual turnover". That is still true of the robot arm. A bean-to-cup self-serve box at this price is not, and the honest reading is that the price bar for this class of machine is now closer to a good commercial espresso machine than to a year's turnover. The technology is the quiet half of this: three bean varieties, hot and cold milk foam, a decaf on every espresso drink and the digital services Melitta is selling with the hardware are what let one box carry eighty drinks, and that capability is what your board is now being compared against.
Read the Henderson release on Business Eye ->
Read Foodservice Equipment Journal on the Melitta partnership ->
Shelf Food Inflation Held at 1.3% for a Second Month and the Sugar and Confectionery Class Fell From 11.9% to 0.6% in Ten Months - the FDF's Own Report Said the Climb Would Start in August, and Two Forecasters Now Put the Peak in 2027
On Wednesday 16 September the ONS published August's inflation figures, and the food line did not move: "The 12-month inflation rate for food and non-alcoholic beverages was 1.3% in August 2026, unchanged from July. The rate was last lower in September 2021, when it was 0.8%." Prices rose 0.4% on the month, the same as a year earlier. The number worth taking away is the contribution: food added 0.11 percentage points to the CPIH annual rate in August, "the smallest since September 2021", against 0.46 points in August 2025. Food is now adding about a tenth of a point to the headline, against nearly half a point a year ago. The headline itself went the other way, CPI from 2.9% to 3.1% and CPIH from 3.1% to 3.3%, because transport and motor fuels made the largest upward contribution, which is the diesel story above. Restaurants and hotels rose 4.1% in the year to August, up from 4.0% in July and down 0.2% on the month, identical on both measures; four weeks ago this column put the gap between the shelf and the table at 2.7 points, and it is now 2.8. Inside the basket, the class the ONS calls sugar, jam, syrups, chocolate and confectionery has a 12-month rate of 0.6%, "down from a recent high of 11.9% in October 2025", and "The easing in the rate over this longer period principally reflects changes in the price of chocolate confectionery." Two cautions on that. It is a retail class, a bar of chocolate on a shelf, not your supplier's cocoa powder. And the ONS disclosed in the same release a classification error in its scanner data for February to July 2026 with effects "of up to positive or negative 0.2 percentage points" for some items at lower levels of aggregation within food, left uncorrected in those months, so the ten-month comparison is the one to trust and the month-to-month moves at class level are not.
The forecasters are the story, because the index did not move and they did. The Food and Drink Federation's report, "UK food inflation forecast 2026-2027: Later and longer", written by its chief economist Dr Liliana Danila and published on Wednesday 9 September, says this: "Annual food inflation reached 1.3% in July 2026, the lowest rate since September 2021. We expect inflation to start accelerating from August 2026, reaching 3.9% by December and averaging 2.8% across the year. For 2027, we forecast food and non-alcoholic drink inflation to reach 5.5%, peaking at 6.4% in July." August was the month the acceleration was scheduled to begin, and it printed unchanged. The FDF's previous forecast, which this column ran in April as "at least 9%" by the end of the year and which our own 9% guide was built on, is summarised by an ITV News interview carried on STV News, whose 8 September piece quotes the FDF's chief executive Karen Betts, as 9% or 10% by Christmas: "It now thinks the rise will be much smaller and much later." The FDF's explanation for later and longer is hedging: "Following the 2022 energy shock, many manufacturers extended their hedging horizons, delaying the impact of higher energy and commodity costs for consumers", while "pressure on household budgets, intense competition and retailer resistance to price increases have limited manufacturers' ability to pass on higher costs". Betts's line to ITV News covers seven years, not one, and is worth reading in full for that reason: "Food prices are going to rise and by the middle of next year, a family food bill that might have cost them £100 a week in 2020 could cost as much as £150 a week." She named the policy she wants slowed: "A huge driver of inflation at the moment is the government's packaging and recycling reforms. If government was just to slow those down, to rationalise them, to make them coherent, they could absolutely take some of the heat out of food prices." And the FDF told ITV News that higher National Insurance and the new packaging and recycling rules have added around £2bn to the annual cost of producing food. ITV News also carries the FDF's own illustration of why an average is useless to a kitchen: iceberg lettuces and tomatoes almost twice last summer's price, spinach, raspberries and red onions cheaper.
The second forecaster published on the day of the ONS release. IGD's press release, carried by Grocery Trader on 16 September, puts average food inflation at 2.9% to 3.9% for 2026, rising to between 5.6% and 6.6% in 2027 and between 5.3% and 6.3% in 2028; the report itself sits behind a download form and we have not read it. Its chief economist James Walton is the same man this column quoted four weeks ago saying retailers appeared to be absorbing some cost increases and that plentiful supply "may have provided a temporary cushion". He has now dated his own cushion: "Food inflation's current weakness will not last. Stock buffers and hedging have delayed, not removed, the pressure building from disrupted energy markets and extreme weather." The release also escalates the forecast this column has been tracking from the commodity side since July: El Nino is now "an extremely severe event, with effects on yields, quality and availability likely to persist well into 2027 and beyond", with fruit and vegetables expected to contribute most. A third instrument, on a different window, points the same way. Worldpanel by Numerator's like-for-like grocery inflation was 2.3% in the four weeks to 6 September, up from 2.1%, as Reuters reported on 15 September, carried by Global Banking and Finance Review, with the wire naming the same three reasons the 2026 overshoot never arrived: "fierce supermarket competition, consumers' resistance to further price rises and better hedging by suppliers".
What this means for you: None of these numbers is your invoice, and the reason we keep saying so is that it is the easiest mistake in the trade. The staleness audit put it in one box: "Headline food inflation is a supermarket shelf index built for households. It is not your wholesale account, it is not your product mix, and it averages your rises away against someone else's falls. A calm headline is not evidence that your lines are calm." Its own example was matcha, up 47% between March and July in three supplier moves while the national index it cites was falling, ONS food from 2.2% in May to 1.3% in July. The July edition called a national index "a hypothesis about your kitchen, not a fact about it", and the same 24 August edition added a second clause to the price-rise method's rule about not blaming suppliers in front of customers: do not argue the index either. So the chocolate number is a prompt, not a verdict. Our cocoa post is explicit that "You pay your supplier's price, which is sticky on the way up and reluctant on the way down", so a 0.6% retail class does not mean your couverture has fallen; it means the brownie and the mocha you priced at the top of the spike are worth a look, on your own invoice, this week. What the forecast reversal actually gives you is a window. Hedging is the one explanation all three give, and what hedging does is delay. A supplier who hedged in 2025 has not sent you a letter yet, which is not the same as a supplier who never will, and both forecasters have the rise building through this autumn and the peak in 2027. The price-rise method asks for four to six weeks of preparation and a full recost before any price moves; a quiet quarter is when that is cheap to do, and July 2027, when every cafe on the street is doing it at once, is when it is not. Start with one dish and see what your own invoices have done. On the policy line, Betts's "packaging and recycling reforms" reach a cafe by the route our EPR post describes: "It doesn't land as a line called 'EPR fee.' It lands as a quietly higher price on the same cup you've been buying for years." And a housekeeping note of our own: the 9% guide now carries a dated note that the forecast it was built on has been revised down and pushed out. The method in it has not changed.
Read the IGD forecast on Grocery Trader ->
16% of Around 300 Scottish Licensed Operators Say Weight-Loss Drugs Have Already Cut Visits and 22% Say Customers Have Become More "Discerning" - the Quote to Read Is About a Shared Bowl of Soup
The Scottish Licensed Trade Association's Summer On-Trade Market Insight Report went out on the PA wire on the evening of Monday 14 September and ran in The Scotsman the following morning with an interview attached. The survey covered around 300 pubs, bars, restaurants and hotels in Scotland, and two of its findings are the reason it is here. 16% said visits had reduced because of weight-loss drugs, and a further 22% said their customers had become more "discerning" because of them. Both PA and The Scotsman print the first figure as "one in eight (16%)", which is not the fraction 16% makes, so we will stick to the percentage. The SLTA's managing director, Colin Wilkinson, told The Scotsman how the question came about, "There is a great deal of interest in weight loss drugs, so we asked our operators if they believe this is having an impact on business", and what came back: "We were quite surprised that 16 per cent did come back saying they believed that, and 22 per cent stated that customers are now more discerning." Then the quote that makes it a story: "I am hearing members noticing people having just a main course, or customers asking to order from the children's menu, or customers sharing a meal - I've even heard of one sharing a bowl of soup or sharing a glass of wine." His own hedge belongs next to it: "There has been a shift in customers' behaviour and while we can't say it is solely down to weight loss drugs, many believe this medication is having an impact on sales." The adaptation he describes is smaller, not bigger: "That could be smaller portions, or people going out to eat at a different time." And he expects it to last: "I keep hearing about how successful these drugs are for some people, so I think this is a trend that will continue."
The rest of the report is the licensed trade's usual weather. 58% of outlets said their business was in decline, even though two thirds said the World Cup had lifted trade; the report found the Commonwealth Games in Glasgow "did nothing for sales in pubs and bars"; 97% reported year-on-year cost increases and 57% said costs had risen "significantly" above inflation; and yet 71% expect to break even or be profitable, up from 63% in the SLTA's winter survey, with 61% expecting growth or stability for the rest of the year. Wilkinson to PA: "The World Cup provided a brief sales boost, but the broader economic outlook remains challenging." The second half of the coverage is the trade body's 10% VAT ask, which this column has run and which our June piece took apart, because around 45% of hospitality businesses would get nothing from it, so we will leave it there. The limits of the data are worth stating plainly. It is operator perception, not till data; the panel is Scottish and licensed, with no cafes in it; the SLTA's own domain now resolves to a for-sale parking page rather than to the report, so every figure here is as PA, The Scotsman and CLH News reported it; and only CLH News names GLP-1 injections, where the others say weight-loss or fat-loss drugs.
What this means for you: Four weeks ago the Barclays data showed restaurants, cafes and bakeries, one pooled line rather than cafes alone, taking 2.1% more in July from 0.3% fewer transactions: more cash per transaction, though that 2.1% was still a real-terms fall against the 2.8% inflation rate Barclays printed beside it. This is the other half of the picture: 16% reporting fewer visits, and a separate 22% reporting customers who have become more "discerning" about what they order, for a reason Wilkinson will not pin on the drugs alone but expects to continue. Whether or not you believe a Scottish licensed-trade survey, the shape it describes is a mix shift, and our food-and-drink GP post already explained why a mix shift is invisible on a cafe's P&L: "If you only ever look at one GP figure, you cannot tell a mix change from a margin change. They move the same number in the same direction, and only one of them is your fault." Its worked example lost four points of blended GP between two weeks in which nobody changed a recipe, a price or a supplier, because "It rained." A regular who keeps the flat white and drops the slice runs that example in reverse, and the reverse is the worse half: in that post drinks run at 78% GP and food at 58%, so dropping the food pushes the blended number up while the cash gross profit falls. That is the post's second worked case, the one it ends with "Your dashboard shows an improving business. Your kitchen just lost five points and nobody noticed." So the job is a number on your own till, not a belief about a drug. Split food and drink. Pull the sales mix, which takes about 40 minutes, and watch the dessert and large-plate lines against the drinks lines month by month. Know what a slice actually makes, because if the cabinet is where the demand goes first you want to know which of its lines you can afford to lose; the admission in that post, "We costed coffee like surgeons and costed cake like we were doing someone a favour", is mine, from seventeen years at Hunters. And reconcile this with the average-spend post rather than dropping it, because that post's lever, 40p a head across 150 covers and roughly 312 trading days coming to £18,720 a year, is still the one an independent genuinely controls. The principle holds; the target changes. A second drink, a pot of tea at a 97.8% ingredient margin, a good decaf, a small high-margin sweet thing, a bag of beans to take home, rather than a bigger plate or a meal deal. The adaptation Wilkinson describes is smaller portions, and a cafe is better placed for that than a pub. The seasonal note writes itself. The autumn switch post says that first cold Tuesday is the best day of your month, if you are ready for it, because "people order a bowl and a coffee instead of a cold drink and nothing". Soup season starts about now. If the bowl is being shared, your till will tell you before a survey does.
Read the interview in The Scotsman ->
The Brikly Take
Five stories, and not one of them was written about a cafe. A Welsh Government press release, an RAC statement, a Belfast wholesaler's investment note, an ONS table and a Scottish licensed-trade survey. What they share is that each carries a number with a twin on your own page, and the twin is the useful one. The Welsh threshold is a look-up against your rateable value. The diesel price, under 4p off its record on Friday, is a question about your surcharge percentage. The self-serve box's 14 million cups a year, about 43 a machine a day on our division, is a comparison with your own count. The sugar and confectionery class is a prompt to open one supplier's invoice. And the shared bowl of soup is a line in your food-and-drink split. In every case the national number tells you where to look, and only your own number tells you what to do.
Five weeks and two days to the Budget. The lines are still being drawn on other people's pages. Find the twin on yours.
The Weekly Grind is published every Monday by Brikly - modular intelligence tools for independent cafe and coffee shop owners. Got a story we should cover? Get in touch ->
Ed O'Brien has run Hunters Cake Company for 17 years across cafes in Witney, Burford, and a bakery in Carterton, Oxfordshire. He's building Brikly - modular tools that give independent cafe owners the same data the big chains have, without the big chain price tag.