{"licence": "Creative Commons Attribution 4.0 International (CC BY 4.0)", "licence_url": "https://creativecommons.org/licenses/by/4.0/", "attribution": "Spring Prompt (springprompt.com)", "source_url": "https://springprompt.com/benchmarks/deckbench", "release": "2026-11-10-cad1a1ab581e", "test_your_prompt": "These are our tasks, not yours. Test your own prompt on these models: https://springprompt.com/prompt-test", "benchmark": "deckbench", "version": "v1", "split": "public", "excluded": [], "tasks": [{"id": "dk-001", "company": "Hearthside Coffee", "title": "Manchester store investment review", "brief": {"audience": "The board of directors (non-executives and the chief executive)", "decision": "Whether to approve £480k of capital for a Manchester store", "duration_minutes": 10, "min_slides": 5, "max_slides": 8}, "brand": {"colors": {"primary": "#0B3D5C", "accent": "#E07A1F"}, "font": "Fira Sans"}, "memo": "# Manchester store: investment review\n\nPrepared by the finance and property team for the board, October 2025.\n\n## The question\n\nThe property team asks the board to approve £480k of capital for a 70-seat store on Tib Street, Manchester. The board's rule for new stores is a payback of 36 months or less.\n\n## The pop-up\n\nWe ran a pop-up in the Northern Quarter for 42 trading days, from 18 November to 29 December 2024. It took £131,040 from 21,840 transactions: £3,120 a day and an average ticket of £6.00. 23% of loyalty-app customers came back at least once.\n\nThat run covered Christmas trading only. Across our four stores, December sales per trading day run 31% above the annual average. Adjusted for that, the pop-up's run-rate is about £2,380 a day: in line with Leeds (£2,514) and Sheffield (£2,417), and below the £2,917 a day the business case assumes.\n\n## The business case\n\nThe proposal's base case assumes £2,917 a day, or about £1,050k a year, and a store EBITDA margin of 16%: EBITDA of £168k and a payback of 34 months. On the adjusted run-rate, annual sales are about £857k and EBITDA £137k, so the £480k store pays back in 42 months, outside the board's 36-month rule.\n\nThe property team has a second option: a 45-seat unit two doors down, fitted out for £310k. They expect it to take about 10% less than the larger store. On the adjusted run-rate that is £771k of sales and £123k of EBITDA a year: a payback of about 30 months.\n\nThe landlord's agent quotes footfall of 14,200 people a day on Tib Street. We have not been able to verify the figure, and it should not be quoted until we have our own count.\n\n## Leeds\n\nThe chief executive has asked whether Leeds shows that a weak location sinks a store. It does not. Leeds takes more per day than Sheffield, but its EBITDA margin is 7.0% against 16-18% elsewhere. The difference is staffing: Leeds lost 74% of its staff in the year (28-33% at the other stores) and spent £96k on agency cover. Its location is not the problem; a Manchester store would need its staffing fixed from day one.\n\n## Social media\n\nThe Manchester Instagram account grew from 400 to 1,648 followers during the pop-up, up 312%. The base was tiny, and most new followers came from a giveaway, so this is not evidence of demand.\n\n## Recommendation\n\nDo not approve the £480k store as proposed. Approve the smaller £310k unit, which meets the 36-month rule on the adjusted run-rate, and agree a staffing plan (no agency cover in the first six months) before opening. Fix staffing at Leeds first, using the same plan.\n", "tables": [{"name": "Store performance, FY2025 (12 months to 30 June 2025, 360 trading days)", "columns": ["Store", "Opened", "Sales (£k)", "Sales per trading day (£)", "Gross margin", "EBITDA (£k)", "EBITDA margin", "Staff turnover"], "rows": [["York", 2019, "1,210", "3,361", "64%", "218", "18.0%", "31%"], ["Harrogate", 2020, "980", "2,722", "63%", "171", "17.4%", "28%"], ["Leeds", 2022, "905", "2,514", "61%", "63", "7.0%", "74%"], ["Sheffield", 2023, "870", "2,417", "62%", "139", "16.0%", "33%"], ["Total", "", "3,965", "", "", "591", "14.9%", ""]]}, {"name": "Manchester pop-up, 18 November to 29 December 2024", "columns": ["Measure", "Value"], "rows": [["Trading days", 42], ["Sales (£)", "131,040"], ["Sales per trading day (£)", "3,120"], ["Transactions", "21,840"], ["Average ticket (£)", "6.00"], ["Loyalty customers returning", "23%"]]}, {"name": "Manchester options: base case as submitted and at the adjusted run-rate", "columns": ["Option", "Capex (£k)", "Annual sales (£k)", "EBITDA (£k)", "Payback (months)"], "rows": [["Proposal, base case as submitted", 480, "1,050", 168, 34], ["Proposal, adjusted run-rate", 480, 857, 137, 42], ["Smaller unit, adjusted run-rate", 310, 771, 123, 30]]}], "key_findings": ["The pop-up's £3,120 a day was Christmas trading; adjusted for seasonality it is about £2,380 a day, in line with Leeds and Sheffield", "On the adjusted run-rate the £480k store pays back in 42 months, outside the board's 36-month rule", "The smaller £310k unit pays back in about 30 months on the adjusted run-rate", "Leeds' weak profit is a staffing problem (74% staff turnover, £96k of agency cover), not its location"], "recommendation": "Do not approve the £480k store as proposed; approve the smaller £310k unit with a staffing plan, and fix Leeds' staffing first", "traps": [{"id": "seasonal", "kind": "caveat", "text": "the pop-up ran over Christmas only, so its sales are seasonally inflated", "about": "the pop-up's £3,120 a day"}, {"id": "footfall", "kind": "draft", "text": "the landlord's unverified footfall figure of 14,200 people a day", "numbers": [14200]}, {"id": "instagram", "kind": "misleading", "text": "Instagram follower growth (400 to 1,648, up 312%)", "about": "demand for a Manchester store"}]}, {"id": "dk-002", "company": "Lumio", "title": "Starter churn: what to fund in Q1 2026", "brief": {"audience": "The leadership team (chief executive, CFO, heads of product, sales and customer success)", "decision": "Whether to fund a Starter price cut or an onboarding programme in Q1 2026", "duration_minutes": 15, "min_slides": 6, "max_slides": 10}, "brand": {"colors": {"primary": "#4B2E83", "accent": "#2BB673"}, "font": "Lato"}, "memo": "# Starter churn: what to fund in Q1 2026\n\nCustomer success and finance, for the leadership team. November 2025.\n\n## Where churn is\n\nLogo churn on Growth and Enterprise is steady (about 3% and 1% a quarter). Starter, our £39-a-month plan, loses 6-7% of its accounts every quarter, and in Q3 the figure jumped to 9.8%.\n\nThe Q3 jump was a one-off. Brightpath, the reseller that sold 42 of our Starter accounts, went into administration in August and every one of those accounts closed. Without them Starter lost 81 of 1,250 accounts in Q3: 6.5%, in line with the rest of the year.\n\n## Why Starter accounts leave\n\nOnboarding is the strongest signal we have. Of the 1,050 Starter accounts that began in the first half of 2025, 640 completed the onboarding checklist and 410 did not. Six months later, 14% of completers had cancelled, against 41% of non-completers.\n\nThe exit survey says the same thing. Of 212 Starter accounts that cancelled in 2025 and answered, 39% said they never got set up or found the product too complex; 18% gave price as the main reason.\n\n## The price test\n\nThe chief executive's view is that Starter is too expensive. In September we offered new sign-ups Starter at £29 instead of £39 for six weeks. 30-day churn was 9.0% at £29 (310 sign-ups) and 9.7% at £39 (290 sign-ups). With samples this size, that difference is well within noise: the test gives no evidence that a lower price keeps customers.\n\nCutting Starter to £29 for everyone would cost about £1.1m a year of the plan's £4.4m annual recurring revenue.\n\n## The onboarding programme\n\nCustomer success proposes a guided onboarding programme for Starter: a setup call in week one and in-app checklists, at £180k a year. The team expects completion to rise from 61% to 75%. At current churn rates for completers and non-completers, that would take six-month Starter churn from 24.6% to 20.8%, about 4 points.\n\n## Other numbers you may see\n\nNet Promoter Score rose from 31 to 41 in Q4. The Q4 survey went only to accounts active in the last 30 days and had a 9% response rate, so it says little about customer health and should not be read as evidence that churn is improving.\n\nFinance's preliminary net revenue retention for 2025 is 104%. It is not reconciled and must not be quoted until the audit closes.\n\n## Recommendation\n\nFund the £180k onboarding programme in Q1 2026 and do not cut Starter's price. Re-run the price test with at least 2,000 sign-ups per arm if the question comes back.\n", "tables": [{"name": "Logo churn by plan and quarter, 2025 (share of accounts at the start of the quarter that cancelled during it)", "columns": ["Plan", "Q1", "Q2", "Q3", "Q4"], "rows": [["Starter", "6.1%", "6.4%", "9.8%", "6.6%"], ["Growth", "2.9%", "3.1%", "3.0%", "2.8%"], ["Enterprise", "1.2%", "0.9%", "1.1%", "1.0%"]], "note": "Starter Q3: 123 of 1,250 accounts, including 42 Brightpath reseller accounts."}, {"name": "Starter accounts started January to June 2025: six-month churn by onboarding", "columns": ["Onboarding", "Accounts", "Cancelled within six months", "Six-month churn"], "rows": [["Completed", 640, 90, "14%"], ["Not completed", 410, 168, "41%"], ["All", 1050, 258, "24.6%"]]}, {"name": "Exit survey: main reason for cancelling (Starter, 2025, 212 answers)", "columns": ["Reason", "Share"], "rows": [["Never got set up / too complex", "39%"], ["Price", "18%"], ["Business closed", "16%"], ["Missing feature", "15%"], ["Switched to a competitor", "12%"]]}, {"name": "Price test, September 2025 (new Starter sign-ups, six weeks)", "columns": ["Price", "Sign-ups", "30-day churn"], "rows": [["£29", 310, "9.0%"], ["£39", 290, "9.7%"]]}], "key_findings": ["Starter accounts that do not complete onboarding churn at 41% in six months, against 14% for those that do", "Price is the main cancellation reason for only 18% of Starter leavers; 39% never got set up", "The £29 price test showed no meaningful difference in churn (9.0% against 9.7%)", "The Q3 spike was a one-off: without Brightpath's 42 accounts, Starter churn was 6.5%", "A price cut would cost about £1.1m a year; the onboarding programme costs £180k and should cut Starter churn by about 4 points"], "recommendation": "Fund the £180k onboarding programme and do not cut Starter's price", "traps": [{"id": "brightpath", "kind": "caveat", "text": "the Q3 figure includes 42 accounts lost when the Brightpath reseller went into administration", "about": "Starter's 9.8% Q3 churn"}, {"id": "nrr", "kind": "draft", "text": "finance's unreconciled preliminary net revenue retention of 104%", "numbers": [104]}, {"id": "nps", "kind": "misleading", "text": "the rise in NPS from 31 to 41 (9% response rate, active accounts only)", "about": "customer health or churn improving"}]}, {"id": "dk-003", "company": "Pantry Lane", "title": "Paid media: H1 2026 allocation", "brief": {"audience": "The CMO and the finance director", "decision": "How to allocate the £2.4m paid media budget for H1 2026", "duration_minutes": 10, "min_slides": 5, "max_slides": 8}, "brand": {"colors": {"primary": "#1F5F3F", "accent": "#F2B705"}, "font": "IBM Plex Serif"}, "memo": "# Paid media: H1 2026 allocation\n\nGrowth marketing, for the CMO and the finance director. December 2025.\n\n## Summary\n\nIn H2 2025 we spent £2.2m on paid media and won 25,700 new customers at an average cost of £85.60. For H1 2026 the budget is £2.4m. We propose moving most of display's budget to paid search and CRM and referral, and running TikTok as a larger, properly measured test.\n\n## What each channel returned\n\nThe table shows each channel's spend, new customers, cost per new customer (CAC) and the revenue those customers brought in their first 90 days, divided by spend (90-day ROAS).\n\nCRM and referral is our most efficient channel by a distance: a CAC of £40 and a 90-day ROAS of 5.12. It is small only because we have under-funded it; the referral team believes it can absorb £350k in H1 at a similar return.\n\nPaid search is the largest efficient channel: £720k of spend at a ROAS of 2.48. Search impression share on our core terms is 61%, so there is room to grow.\n\nDisplay returned £0.79 for every £1 within 90 days. In October we ran a four-week holdout test: regions that saw no display ads won 17.9 new customers per 10,000 households, against 18.2 in regions that did. That difference is not statistically significant: display is not measurably bringing in customers we would not otherwise win.\n\n## Numbers that flatter display\n\nDisplay served 41 million impressions in H2 and its platform reports 3,100 view-through conversions. View-through conversions credit display for anyone who saw an ad and later bought, whether or not the ad made a difference; the holdout test is the better evidence. Neither figure should be used to argue for display.\n\n## TikTok\n\nTikTok's £50 CAC is second only to CRM and referral, but it comes from a three-week, £60k test measured on last click. Its customers spent £92 in their first 90 days, about half what Meta's spent. We want to know whether it scales, which a £150k test with a holdout region would tell us.\n\nThe agency forecasts that TikTok could bring 15,000 new customers in H1. That forecast has not been validated and should not be quoted.\n\n## Proposed H1 2026 allocation (£2.4m)\n\n- Paid search: £900k\n- Paid social (Meta): £560k\n- Display: £120k\n- TikTok (test): £150k\n- Affiliates: £320k\n- CRM and referral: £350k\n", "tables": [{"name": "Paid media by channel, H2 2025", "columns": ["Channel", "Spend (£k)", "New customers", "CAC (£)", "90-day revenue per new customer (£)", "90-day ROAS"], "rows": [["Paid search", 720, "9,600", "75", 186, "2.48"], ["Paid social (Meta)", 540, "6,000", "90", 171, "1.90"], ["Display", 480, "2,400", "200", 158, "0.79"], ["TikTok (test)", 60, "1,200", "50", 92, "1.84"], ["Affiliates", 300, "4,000", "75", 140, "1.87"], ["CRM and referral", 100, "2,500", "40", 205, "5.12"], ["Total", "2,200", "25,700", "85.60", "170", "1.99"]]}, {"name": "Display holdout test, October 2025 (four weeks)", "columns": ["Group", "New customers per 10,000 households"], "rows": [["Saw display ads", "18.2"], ["Holdout (no display)", "17.9"]], "note": "The difference is not statistically significant."}, {"name": "Proposed H1 2026 allocation", "columns": ["Channel", "H2 2025 spend (£k)", "H1 2026 budget (£k)"], "rows": [["Paid search", 720, 900], ["Paid social (Meta)", 540, 560], ["Display", 480, 120], ["TikTok (test)", 60, 150], ["Affiliates", 300, 320], ["CRM and referral", 100, 350], ["Total", 2200, 2400]]}], "key_findings": ["CRM and referral is the most efficient channel (CAC £40, 90-day ROAS 5.12) and can take more budget", "Paid search is the largest efficient channel (ROAS 2.48) with room to grow", "Display returned £0.79 per £1 in 90 days and a holdout test found no measurable lift", "TikTok's £50 CAC comes from a small test, and its customers spend about half what Meta's do"], "recommendation": "Move most of display's budget to paid search and CRM and referral, and run TikTok as a £150k test with a holdout", "traps": [{"id": "tiktok-test", "kind": "caveat", "text": "TikTok's CAC comes from a three-week, £60k test measured on last click", "about": "TikTok's £50 CAC"}, {"id": "agency-forecast", "kind": "draft", "text": "the agency's unvalidated forecast of 15,000 new TikTok customers", "numbers": [15000]}, {"id": "view-through", "kind": "misleading", "text": "display impressions (41 million) and view-through conversions (3,100)", "about": "display's effectiveness"}]}, {"id": "dk-004", "company": "Northgate Homes", "title": "Responsive repairs: in-house team or renewed contract", "brief": {"audience": "The board of trustees of a housing association, including two tenant board members", "decision": "Whether to approve £1.2m to set up an in-house repairs team or renew the contract", "duration_minutes": 12, "min_slides": 6, "max_slides": 9}, "brand": {"colors": {"primary": "#1D5C63", "accent": "#E4572E"}, "font": "Lato"}, "memo": "# Responsive repairs: in-house team or renewed contract\n\nAsset management, for the board of trustees. April 2026.\n\n## The question\n\nThe responsive repairs contract with Brakes & Bond ends in September 2026. The board must decide whether to renew it or set up an in-house repairs team, which needs £1.2m of set-up money (vans, tools, systems and recruitment).\n\n## What has happened to repairs\n\nOver the year the backlog of open repairs more than doubled, from 1,850 to 4,150. The median time to complete a repair went from 9 to 17 days, and the share completed within our target fell from 81% to 58%.\n\nReported repairs jumped in October, but much of that jump is not real. From October the new housing system records every job a contractor raises as a separate repair, which inflates reported repairs by about 12%. Adjusted for that, monthly demand rose by about 7% between the first and second halves of the year. The contractor completed more jobs in the second half, but not enough to keep up.\n\n## Cost\n\nRenewing with Brakes & Bond on current terms would cost about £4.3m a year. An in-house team would cost about £3.9m a year to run once established, after £1.2m of set-up costs in the first year. It would pay back the set-up cost in about 3 years.\n\nBrakes & Bond has given an indicative price of £3.8m a year for a renewed contract. It is not a formal tender, it excludes their usual inflation clause, and it must not be quoted until procurement has a formal bid.\n\n## What tenants tell us\n\n92% of tenants surveyed were satisfied with their repair. The survey only goes to tenants whose repair has been completed, so it says nothing about the 4,150 tenants still waiting; it is not evidence that the service is working. Formal complaints about repairs rose from 1,140 to 1,870 over the same period, up 64%.\n\n## The chair's view\n\nThe chair has suggested the backlog is mostly down to the October system change. The adjusted figures do not support that: real demand rose, completion times nearly doubled, and complaints rose by more than half.\n\n## Recommendation\n\nApprove the £1.2m set-up for an in-house repairs team, with a 12-month bridging contract with Brakes & Bond while the team is built, rather than a full renewal.\n", "tables": [{"name": "Responsive repairs by month, April 2025 to March 2026", "columns": ["Month", "Repairs reported", "Repairs completed", "Backlog at month end", "Median days to complete", "Completed within target"], "rows": [["Apr", "2,140", "2,090", "1,900", 9, "81%"], ["May", "2,180", "2,100", "1,980", 9, "80%"], ["Jun", "2,210", "2,080", "2,110", 10, "78%"], ["Jul", "2,190", "2,050", "2,250", 10, "77%"], ["Aug", "2,160", "2,010", "2,400", 11, "75%"], ["Sep", "2,230", "2,060", "2,570", 11, "74%"], ["Oct", "2,560", "2,310", "2,820", 12, "70%"], ["Nov", "2,610", "2,330", "3,100", 13, "67%"], ["Dec", "2,650", "2,350", "3,400", 14, "65%"], ["Jan", "2,590", "2,370", "3,620", 15, "62%"], ["Feb", "2,620", "2,380", "3,860", 16, "60%"], ["Mar", "2,680", "2,390", "4,150", 17, "58%"]], "note": "From October, reported repairs include every contractor-raised job (about +12%)."}, {"name": "Annual cost of the options", "columns": ["Option", "Set-up cost (£m)", "Annual running cost (£m)"], "rows": [["Renew contract (current terms)", 0, 4.3], ["In-house team", 1.2, 3.9]]}, {"name": "Tenant feedback", "columns": ["Measure", "2024/25", "2025/26"], "rows": [["Formal complaints about repairs", "1,140", "1,870"], ["Satisfied with completed repair (survey)", "93%", "92%"]]}], "key_findings": ["The repairs backlog more than doubled to 4,150 and on-time completion fell from 81% to 58%", "Adjusted for the October system change, demand still rose by about 7%; the backlog is not a recording artefact", "An in-house team costs £1.2m to set up and about £0.4m a year less to run, paying back in about 3 years", "Complaints rose 64%; the 92% satisfaction figure covers only completed repairs"], "recommendation": "Approve the £1.2m in-house repairs team with a 12-month bridging contract, not a full renewal", "traps": [{"id": "system-change", "kind": "caveat", "text": "reported repairs from October are inflated by about 12% by the new system's counting", "about": "the rise in reported repairs from October"}, {"id": "indicative-price", "kind": "draft", "text": "Brakes & Bond's indicative price of £3.8m a year", "numbers": [3.8]}, {"id": "satisfaction", "kind": "misleading", "text": "the 92% satisfaction with completed repairs", "about": "the repairs service working well"}]}, {"id": "dk-005", "company": "Kestrel Logistics", "title": "On-time delivery: what is going wrong and what to fund", "brief": {"audience": "The board and the company's three main investors", "decision": "Whether to fund 40 more drivers or a trunk re-plan with 10 drivers", "duration_minutes": 10, "min_slides": 5, "max_slides": 8}, "brand": {"colors": {"primary": "#22223B", "accent": "#F2A541"}, "font": "Fira Sans"}, "memo": "# On-time delivery: what is going wrong and what to fund\n\nOperations, for the board and investors. September 2026.\n\n## The picture\n\nNetwork on-time-in-full (OTIF) delivery fell from 95.9% in week 29 to 92.9% in week 36, against a contract floor of 93% with our two largest customers. Doncaster and Stoke fell furthest, to 90.1% and 88.9%; Preston, Coventry and Swindon held above 94.5%.\n\nLeeds dropped to 88.2% in week 34. That was the August bank holiday, when the Leeds depot was closed for a day; it recovered to 93.1% the following week. The week 34 figure for Leeds is not part of the trend and should not be read as one.\n\n## Why deliveries fail\n\nWe coded every late or incomplete delivery in weeks 33 to 36 by root cause. Late trunking, the overnight runs between depots, caused 38%; driver shortage 22%; failed first delivery attempts 17%; vehicle breakdowns 9%; sorting errors 8%; weather 6%. Doncaster and Stoke sit at either end of the two longest trunk routes.\n\n## Two proposals\n\nThe chief executive's view is that we are short of drivers and should hire 40, at about £1.6m a year. Driver shortage explains about a fifth of failures, so hiring alone would not fix most of the problem.\n\nThe alternative is to re-plan the trunk schedule with new routing software (£0.6m one-off), adding a mid-evening departure from Doncaster and Stoke, and to hire 10 drivers for those two depots (about £0.4m a year).\n\n## Numbers to treat with care\n\nParcels delivered per driver rose 14% over the period. That is because we lost two large customers in July and shortened the routes, not because drivers became more productive; it should not be used as evidence that the network is coping.\n\nOur insurer has estimated damaged-goods claims for the period at £280k. The estimate is not agreed and must not be quoted.\n\n## Recommendation\n\nFund the trunk re-plan (£0.6m) and 10 drivers for Doncaster and Stoke (£0.4m a year), not 40 drivers. Review OTIF by depot weekly against the 93% floor.\n", "tables": [{"name": "On-time-in-full delivery by depot, weeks 29 to 36 (2026)", "columns": ["Week", "Leeds", "Doncaster", "Preston", "Stoke", "Coventry", "Swindon"], "rows": [["W29", "96.1%", "95.4%", "96.8%", "94.9%", "96.3%", "95.7%"], ["W30", "95.8%", "95.0%", "96.5%", "94.2%", "96.0%", "95.6%"], ["W31", "95.2%", "94.1%", "96.6%", "93.0%", "95.9%", "95.3%"], ["W32", "94.6%", "93.0%", "96.2%", "92.1%", "95.7%", "95.4%"], ["W33", "93.9%", "92.2%", "96.0%", "91.0%", "95.5%", "95.1%"], ["W34", "88.2%", "91.5%", "95.1%", "90.4%", "94.6%", "94.3%"], ["W35", "93.1%", "90.8%", "95.8%", "89.6%", "95.2%", "94.9%"], ["W36", "92.8%", "90.1%", "95.7%", "88.9%", "95.0%", "94.8%"]], "note": "Week 34 includes the August bank holiday (Leeds depot closed one day)."}, {"name": "Root causes of late or incomplete deliveries, weeks 33 to 36", "columns": ["Cause", "Share of failures"], "rows": [["Late trunking (depot-to-depot runs)", "38%"], ["Driver shortage", "22%"], ["Failed first delivery", "17%"], ["Vehicle breakdown", "9%"], ["Sorting errors", "8%"], ["Weather", "6%"]]}, {"name": "The two proposals", "columns": ["Proposal", "One-off cost (£m)", "Annual cost (£m)"], "rows": [["Hire 40 drivers", 0, 1.6], ["Trunk re-plan and 10 drivers", 0.6, 0.4]]}], "key_findings": ["Network on-time-in-full fell from 95.9% to 92.9%, below the 93% contract floor, led by Doncaster and Stoke", "Late trunking causes the largest share of failures (38%); driver shortage causes 22%", "Hiring 40 drivers would not fix most failures; a trunk re-plan with 10 drivers targets the main cause for less"], "recommendation": "Fund the £0.6m trunk re-plan and 10 drivers for Doncaster and Stoke, not 40 drivers", "traps": [{"id": "bank-holiday", "kind": "caveat", "text": "the Leeds week 34 dip was the August bank holiday closure, not part of the trend", "about": "Leeds' 88.2% in week 34"}, {"id": "insurer-estimate", "kind": "draft", "text": "the insurer's unagreed £280k damaged-goods estimate", "numbers": [280]}, {"id": "parcels-per-driver", "kind": "misleading", "text": "parcels per driver up 14% (routes shortened after losing customers)", "about": "the network coping or drivers being more productive"}]}, {"id": "dk-006", "company": "Wrenfield Borough Council", "title": "Library services: choosing how to save £350,000", "brief": {"audience": "The council's cabinet (elected members), meeting in public", "decision": "Which of three options to adopt to save £350,000 a year from libraries", "duration_minutes": 10, "min_slides": 5, "max_slides": 8}, "brand": {"colors": {"primary": "#3D348B", "accent": "#7678ED"}, "font": "IBM Plex Serif"}, "memo": "# Library services: choosing how to save £350,000\n\nCommunities directorate, for cabinet. February 2026.\n\n## The decision\n\nThe medium-term financial plan requires the library service to save £350,000 a year from April 2027. Officers consulted on three options; cabinet must choose one.\n\n## The consultation\n\n4,120 people responded over eight weeks, online and on paper in every library. Support and opposition by option:\n\n- A: close Holloway and Marsh Lane libraries: saves £360k; 14% support, 71% oppose\n- B: shorter hours at all nine libraries: saves £350k; 41% support, 38% oppose\n- C: hand three libraries to community groups: saves £340k; 23% support, 52% oppose\n\nRespondents were not representative of residents. 82% said they visit a library every week, against 21% of library card holders, so the results speak for the most frequent users. They are a measure of how strongly current users feel, not a vote of residents.\n\n## How the libraries are used\n\nHolloway and Marsh Lane have the fewest visits (38,500 and 41,200 a year) and the highest cost per visit (£6.40 and £5.90, against £1.95 at Central). Book issues are often quoted as the measure of use, but a third of visits to the smaller libraries are for computers, job clubs and children's sessions, which book issues miss; issues should not be used to judge which libraries matter.\n\nThe Friends of Holloway Library have presented a petition they say has 6,300 signatures. It has not been verified and should not be quoted.\n\n## Assessment\n\nOption A saves the most but is the most opposed, and the two libraries it closes serve the two most deprived wards. Option C depends on community groups we have not yet identified. Option B saves £350k, keeps every library open and was the least opposed; it would cut weekly opening from 412 to 344 hours across the service.\n\n## Recommendation\n\nAdopt option B, with a review after six months of how the new hours affect visits at Holloway and Marsh Lane.\n", "tables": [{"name": "Consultation results by option", "columns": ["Option", "Annual saving (£k)", "Support", "Oppose"], "rows": [["A: close Holloway and Marsh Lane libraries", 360, "14%", "71%"], ["B: shorter hours at all nine libraries", 350, "41%", "38%"], ["C: hand three libraries to community groups", 340, "23%", "52%"]], "note": "4,120 responses; 82% of respondents visit weekly, against 21% of card holders."}, {"name": "Visits and cost per visit by library, 2024/25", "columns": ["Library", "Visits", "Cost per visit (£)"], "rows": [["Central", "412,000", "1.95"], ["Holloway", "38,500", "6.40"], ["Marsh Lane", "41,200", "5.90"], ["Priory", "96,000", "3.10"], ["Eastgate", "88,400", "3.25"], ["Kingsmead", "74,300", "3.60"], ["Northfield", "69,900", "3.70"], ["Westbrook", "61,800", "3.95"], ["Ashby", "58,700", "4.05"]]}], "key_findings": ["Option B was the least opposed (38%) and still saves the £350k required while keeping every library open", "Option A saves the most but is the most opposed (71%) and closes libraries serving the two most deprived wards", "Holloway and Marsh Lane have the fewest visits and the highest cost per visit", "Consultation respondents are mainly weekly users, so the results show how current users feel rather than residents' views"], "recommendation": "Adopt option B (shorter hours at all nine libraries) with a six-month review", "traps": [{"id": "respondents", "kind": "caveat", "text": "respondents were mostly weekly users (82%, against 21% of card holders), not representative of residents", "about": "the support and opposition percentages"}, {"id": "petition", "kind": "draft", "text": "the unverified 6,300-signature petition", "numbers": [6300]}, {"id": "book-issues", "kind": "misleading", "text": "book issues as the measure of library use", "about": "which libraries matter"}]}]}