Desktop nav

Images and videos for this page, original size: Asset library →

All frames load the live prototype, so any edit shows in all of them. Open full page →

Desktop 1440 Shop menu open on load. Hover or click Shop / Learn to try it.

Desktop 1440 - Mushroom Elixirs hovered tiles swap per category

Desktop 1280 Shop menu open

Desktop 1440 Learn menu open

Dev instructions - desktop nav (read first) 22 Sep 2026. Everything needed to build arm B. Prototype source: /top-bar (view source for markup and CSS).

#	Area	Instruction
1	Scope	Desktop only (901px and up). The mobile drawer is live and final - do not change it. Built as arm B of a desktop-only Intelligems test (arm A = current live nav). James creates the test; dev builds both arms.
2	Top bar	Left: Shop (opens mega menu, caret) \| Lab Results \| Science \| Learn (opens panel, caret). Logo centred (absolute, true centre). Right: GBP region selector, search, account, cart. No quiz button in the header. Height 68px, Morning Mist #D1E0D7, text Obsidian #161F28. Hover/open state = 2px Forest Fern underline, text stays ink (contrast).
3	Top bar links	Lab Results -> /pages/mould-and-pesticides-independent-laboratory-test-result. Science -> /pages/healthy-coffee-science. Carets only on Shop and Learn (they open panels; the others are plain links).
4	Menu behaviour	Opens on hover after 120ms, closes 200ms after the pointer leaves (no flicker crossing the bar). Also opens on click, Enter and Space. Top item is a button with aria-expanded + aria-controls. ArrowDown or Tab moves into the panel; Tab off the last link goes to the next top item; Escape closes and returns focus. One panel open at a time. A 28% Obsidian dim covers the page below; clicking it closes. prefers-reduced-motion = no fades.
5	Shop panel layout	Full-width white panel under the header, content max 1400px. Left column 240px (220px under 1380px): 4 category rows (Canela 19px, 52px rows, hairline dividers, chevron) then "Shop all" link directly under Coffee Gifts. Right: 4 equal columns.
6	Hover-to-browse	Hovering (150ms intent) or focusing a category row swaps the right-hand tiles to that category. Clicking the row opens the collection. Default and reset-on-close = Beans & Ground. All sets share one grid cell so the panel never changes height.
7	Tiles per category	Beans & Ground: Bestseller Bundle, Darkfire Energy, Stability Blend (Darkfire 2nd so it does not sit beside the Darkfire-led range image - James 22 Sep). Coffee Pods: Clean Coffee Pods (Stability) - add Darkfire pods as slot 2 when it launches (no Darkfire pods photo exists yet; designer to supply). Mushroom Elixirs: Mushroom Elixir Bundle, Lion's Mane Elixir, Cordyceps Elixir. Coffee Gifts (James 22 Sep, this order): "Pods Gift Subscription", "Coffee Gift Subscription" (/products/monthly-gift-subscription), "Mushroom Coffee Gift" (/products/lions-mane-coffee-3-month-gift-subscription). DEV: create a Pods Gift Subscription product (same setup as the Coffee Gift Subscription, with Clean Coffee Pods) - until it exists the tile links to /collections/nespresso-pods. No Gift Card tile. No Christmas imagery on any gift tile out of season.
8	Tile design	Soft card #EDF0ED, radius 8px, square pack shot shown whole (object-fit cover on the square source - no zoom), name Area 600 15px + ONE short line 13px. Build as theme section blocks (image, title, line, chip, link) so James can edit them in the customiser.
9	Chips (data-driven)	"Best seller" on Stability Blend. "Save up to 24%" on Bestseller Bundle (vs three 250g singles). "Save 24%" on Mushroom Elixir Bundle (vs four single bottles). "Sold out" must come from LIVE inventory, never hard-coded. Re-check both saving % whenever single or bundle prices change.
10	Lines on tiles	Bundle: "Three customer favourites". Stability: "Medium roast". Darkfire: "Dark roast". Pods: "Stability Blend". Elixir Bundle: "Your wellbeing toolkit". Lion's Mane: "Focus & mental clarity". Cordyceps: "Energy & endurance" (James 22 Sep: mirror the benefit tags on the live Mushroom Elixirs collection cards). Gifts: titles "Pods Gift Subscription", "Coffee Gift Subscription" (full word, SEO), "Mushroom Coffee Gift" (James 22 Sep: fits one line at 1280; "Lion's Mane Gift Subscription" wraps), lines "Nespresso® compatible", "Beans & ground", "3-month subscription". NO prices anywhere in the menu (consistency; price is shown on the product page). Every tile title and line is ONE line - the price is the LOWEST variant, pulled from the product, never typed in.
11	4th slot	ONE tile built exactly like the product tiles beside it: a square LIFESTYLE image for the category (files in nav-redesign-assets/lifestyle/: lifestyle-beans / -pods / -elixirs / -gifts.webp) and below it the CATEGORY NAME + arrow as the title ("Beans & Ground →", "Nespresso Compatible Pods →" (James 1 Oct: the left tile already says Clean Coffee Pods), "Mushroom Elixirs →", "Coffee Gifts →") with "View all" as the line, linking to that collection (James 22 Sep). One CTA per tile (James 22 Sep). No quiz card here.
11b	Quiz + Shop all	"Take the Quiz" -> /pages/quiz: plain text (14px Area 600, Obsidian), NO background, followed by a 28px round arrow button (1px Forest Fern outline, fills Fern on hover), pinned to the bottom of the left column so it lines up with the tile bottoms. "Shop all" is quieter than the category rows: 13px regular, Obsidian at 60%, hairline underline at 25%, Forest Fern on hover.
12	Learn panel	4 image tiles built with the SAME card as the Shop tiles (#EDF0ED card, 8px radius, image flush on top at 3:2, name Area 600 15px + line 13px left-aligned, 1.03 image zoom on hover), names and lines as live: How To Brew / Find your guide, Founder Story / Meet James, Sustainability / For the Planet, Read Our Blog / Get Barista Smart. NEW images supplied in nav-redesign-assets/learn/ (960x640 webp) - upload to Shopify Files.
13	Removed from desktop	Coffee, Mushroom Tinctures and Coffee Gifts as top-level items (now inside Shop). Brewing Equipment. Reviews in the top bar (still in the Trustpilot banner item and the footer). Quiz in the header.
14	Responsive	Under 1380px: narrower left column (220px) and nav gap 24px (the 4 top items must clear the centred logo). Under 1100px the tiles drop to 2 columns. At 900px and below the mobile drawer takes over.
15	Tracking (GA4)	Every link carries data-ga-item + data-ga-slot. Fire menu_open and nav_item_click (item label, section, slot, device) and select_item with item_list_name "desktop_menu". Full spec: GA4 menu tap tracking brief (James has it). Must be live on day one of the test.
16	Fonts and colours	Canela Regular for category rows, tile headings and Learn names; Area for everything else - use the theme's existing files. Colours: Morning Mist #D1E0D7, Forest Fern #516E5A, Obsidian #161F28, Arctic White #EDF0ED. Radius 8px (thumbnails 6px).
17	Do not	Do not change copy beyond what is listed here. Do not ship the review-only URL switches (?mega=, ?cat=, ?shop=list, ?rules=0). Do not launch, pause or end the test - James does that in Intelligems.

All frames load the live prototype. Mobile A and B open the menu on load. Open full page → Section lines:

Mobile A favourites first

Mobile B categories first

Dev notes (open threads)

Only the announcement bar changes colour, from black to Forest Fern #516E5A. The main nav keeps Morning Mist #D1E0D7, Obsidian Night #161F28 and the real logo.
The announcement bar scrolls continuously, with a solid white pause button.
Mobile header: burger, centred logo (35px) and cart bag only (search and account removed), 50px tall. Desktop nav is 68px: Shop (mega menu), Science, Learn (panel), Reviews on the left; Take the Quiz, region, search, account and cart on the right. Full desktop dev notes sit in the page source above the Shop mega menu.
Mobile menu: full screen, one short scroll. One image block, everything else plain text (4 images in total). Order: Coffee Bundle card, 3 customer favourites (the only image row, label renamed from “Best sellers” so it does not repeat “Bestseller” under the bundle card: Stability Blend, Darkfire Energy, Lion’s Mane Coffee), SHOP BY CATEGORY as plain Canela rows (Beans & Ground, Coffee Pods, Mushroom Elixirs, Coffee Gifts; the Lion’s Mane Coffee row was removed 15 Sep because it duplicated the best-seller tile and linked to the same product page - restore it if the 30-day review drops Lion’s Mane from best sellers), then Take the Quiz / Reviews / Science / Our Story / Chat with us (opens Gorgias chat in Shopify; the prototype falls back to the contact page) in Area on a light tinted panel (white = shop, tinted = everything else), then a bar split in two: region (flag + "GBP", same as desktop) | Account. No trust row in the menu: the announcement bar above the header already shows "Excellent 4.7 rating | 2,000+ reviews". Menu header matches the site header: Morning Mist, 50px, X left, logo centred. No cart in the open menu (James 15 Sep): the page-header cart is one tap away after closing.
The bundle card, best sellers and the category rows show without scrolling in iPhone Safari (fold is about 660px).
A/B layout: A (default) = bundle, best sellers, categories, links. B (?menu=b, or data-menu="b" in Shopify) = bundle, categories, best sellers, links. Same components, order only.
Every category row is 52px and every link row 44px. No line and no white gap where the white list meets the tinted links panel; the colour change is the separator. Space above Take the Quiz equals space below Our Story. Best sellers are Stability Blend, Darkfire Energy and Lion’s Mane Coffee. Slots 1-2 come from first-order share (17 Jun - 14 Sep 2026: 32.2%, 13.4%). Slot 3 is James’s pick (15 Sep) for range and interest, overriding the data pick Halcyon Decaf (12.1% of first orders vs Lion’s Mane 7.2%, 1.9% view-to-purchase): read its tile click-through and tile-to-purchase after 30 days before keeping it. Every menu tile gets a GA4 select_item event (item name + slot); slots reviewed quarterly.
Category rows have no thumbnails and no sublines. The sublines are kept in the dev notes as optional A/B copy.
Moved to the site footer: Subscribe & Save, Sustainability, Blog, Brew Guides, Equipment, Help Centre, Contact Us, Manage Subscription, Balance Rewards, Gifting, badges.
Copy vs live: the only new copy is the Bundle card ("Bestseller Bundle / Save up to 24% vs buying separately.", one line - the saving is against buying the same coffees separately: three 250g singles cost £48.47 vs the £36.99 bundle, a 23.7% saving; measured against current single-bag prices, not the bundle’s £45.00 compare-at; no gift claim). A/B test before treating the menu as a copy change.
Open for Ambiw: bundle card should point at the tiered bundle page once built; Balance Rewards has no page yet.
Region/currency list (desktop dropdown and mobile menu): the complete live list, 51 countries from the site’s Shopify localization form, United Kingdom (GBP £) first then A-Z, each with its flag and currency. In Shopify this is the native localization form, not custom code.
Section lines: approved by James (15 Sep) - the two headings keep a thin rule after the text. The On / Off switch above is for comparison only; the Off state (?rules=0) must not ship.
Nothing goes to Ambiw until James signs off.

Images and videos for this page, original size: Asset library →

Meta offer page: the Clean Coffee Bundle offer that every Meta advertorial CTA lands on. Separate dev task from the advertorial, dev due Fri 16 Oct. Open full page →

What changed vs live Live Meta landing today: /products/espresso-taster-pack

#	Change	Why
1	A dedicated offer page replaces the Espresso Taster Pack PDP as the page Meta traffic buys on	Every paid ad lands on an advertorial, which sends to one offer page. A PDP carries menus and choices that pull a cold visitor off the offer.
2	Three bundle tiers, 3 bags each: Discovery 3 x 250g, 3 x 500g (default), 3 x 1kg, each with price per cup and "savings + gifts"	A ladder with a default middle tier lifts order value. Per cup turns a £71.49 basket into a 71p decision.
3	Subscribe and save 15% on the same picker, with the monthly gift ladder shown	Subscriptions are the retention engine. The gift ladder gives a reason to stay past month 1.
4	Four-card monthly lab proof: chemical pesticides, mould toxins, heavy metals and microbial contaminants, plus a report action	Matches the Purity advertorial the visitor just read. The cards make the testing scope scannable; “Read the full report” keeps the proof one tap away.
5	Deep links: #bundles scrolls to the bundle picker, #lab-report opens the lab report	All 9 buy CTAs in the Purity listicle land on the picker; one lands on the report.

Offer economics after the CFO call Rebuilt 2 Oct 2026 from Neel's call (balance-coffee/marketing/meta-consultancy/calls/2026-10-02-neel-cfo-meta-economics.md). Pending Neel's per-tier model, which replaces every figure here. Profit is before ad spend. Toggle gifts on the page: bottom-left pill, or ?gifts=a / ?gifts=b.

The 3 x 500g is the tier to advertise. At the 2.2x ROAS floor it can cost £32.50 to acquire a buyer, against £16.81 for the 3 x 250g, and Meta will not deliver a buyer for £17. It also has the lowest break-even ROAS on the supplier table (1.58x). The cost basis decides whether 2.2x makes money. On the supplier table every first order clears 2.2x. On Neel's 50% and James's 43%-of-RRP bases a full-price first order roughly breaks even at 2.2x, and a subscriber's discounted first order loses £1-15. That loss is the price of the subscription: order 2 pays it back (version A order 2 profit: £13.24 / £35.94 / £52.43 on the supplier table). Neel locks one basis per tier (playbook Q10).

Decisions in force (James, 2 Oct, on Neel's structure)

Item	Rule
Prices	Meta prices match the website: £36.99 / £71.49 / £109.49. No Meta-only price (playbook BU-4).
Subscribe discount	15% off from order 1. The Ritual Mug is named as the order-2 gift on every tier (BU-5). The offer page shows this since 2 Oct.
Hero tier	3 x 500g leads the ads and is preselected on the page.
Budget-step gate	2.2x platform ROAS, checked against Shopify new-customer revenue. MER is the weekly trajectory check (ME-5).
Kill rule	Pause all spend if platform ROAS is below 1.5x once £1,500 is spent, or at £3,000 cumulative net loss. £4,500 is the CFO's hard cap on the whole test (BU-3, Q12).
Ramp	Jovana's £50 / £75 / £100 a day, two weeks each: about £3,150 over six weeks.
Downside	Plan on 20% of cold buyers subscribing. Neel's 40-50% is the warm-customer rate (Q11).

Gift ladders

Version	3 x 250g, £36.99	3 x 500g, £71.49	3 x 1kg, £109.49	Subscriber 2nd box
A (now, in stock)	Free shipping, Brewing Guides	+ Gold Blend 250g	+ Mushroom Duo samples	Ritual Mug, every tier
B (from late Nov)	Free shipping, Brewing Guides, Scoop	+ Gold Blend 250g	+ Mushroom Duo samples	Ritual Mug, every tier

First-order profit, break-even ROAS and profit at the 2.2x max CPA, version A gifts

Each cell: profit before ads / break-even ROAS / profit after ads at 2.2x.

Tier	First order	Revenue	Max CPA at 2.2x	Supplier table	Neel: 50% bundle margin	James: 43% of single-bag RRP
3 x 250g	One-time, or subscriber with 15% from order 2	£36.99	£16.81	£21.00 / 1.76x / +£4.19	£18.50 / 2.00x / +£1.68	£16.15 / 2.29x / -£0.67
3 x 250g	Subscriber, 15% from order 1	£31.44	£14.29	£15.57 / 2.02x / +£1.28	£12.95 / 2.43x / -£1.35	£10.60 / 2.97x / -£3.69
3 x 500g (ad hero)	One-time, or subscriber with 15% from order 2	£71.49	£32.50	£45.31 / 1.58x / +£12.81	£32.79 / 2.18x / +£0.30	£32.22 / 2.22x / -£0.28
3 x 500g (ad hero)	Subscriber, 15% from order 1	£60.77	£27.62	£34.82 / 1.75x / +£7.20	£22.08 / 2.75x / -£5.55	£21.50 / 2.83x / -£6.12
3 x 1kg	One-time, or subscriber with 15% from order 2	£109.49	£49.77	£63.37 / 1.73x / +£13.60	£47.79 / 2.29x / -£1.97	£43.43 / 2.52x / -£6.34
3 x 1kg	Subscriber, 15% from order 1	£93.07	£42.30	£47.31 / 1.97x / +£5.01	£31.37 / 2.97x / -£10.93	£27.01 / 3.45x / -£15.30

Same, version B gifts (Scoop added to every first order, from late Nov)

Tier	First order	Revenue	Max CPA at 2.2x	Supplier table	Neel: 50% bundle margin	James: 43% of single-bag RRP
3 x 250g	One-time, or subscriber with 15% from order 2	£36.99	£16.81	£19.69 / 1.88x / +£2.88	£17.19 / 2.15x / +£0.37	£14.84 / 2.49x / -£1.98
3 x 250g	Subscriber, 15% from order 1	£31.44	£14.29	£14.26 / 2.20x / -£0.03	£11.64 / 2.70x / -£2.66	£9.29 / 3.39x / -£5.00
3 x 500g (ad hero)	One-time, or subscriber with 15% from order 2	£71.49	£32.50	£44.00 / 1.62x / +£11.50	£31.48 / 2.27x / -£1.01	£30.91 / 2.31x / -£1.59
3 x 500g (ad hero)	Subscriber, 15% from order 1	£60.77	£27.62	£33.51 / 1.81x / +£5.89	£20.77 / 2.93x / -£6.86	£20.19 / 3.01x / -£7.43
3 x 1kg	One-time, or subscriber with 15% from order 2	£109.49	£49.77	£62.06 / 1.76x / +£12.29	£46.48 / 2.36x / -£3.28	£42.12 / 2.60x / -£7.65
3 x 1kg	Subscriber, 15% from order 1	£93.07	£42.30	£46.00 / 2.02x / +£3.70	£30.06 / 3.10x / -£12.24	£25.70 / 3.62x / -£16.61

How each basis works. Supplier table: coffee, pick and carton, Royal Mail, Shopify fees and physical gifts, itemised below. Neel: costs are 50% of the full price (his bundle gross margin), then gifts; a discount cuts revenue but not cost. James: costs are 43% of the live single-bag price of the same three bags (£48.47 / £84.47 / £137.47), then gifts. Max CPA at 2.2x is revenue / 2.2, so a discounted first order also lowers the CPA ceiling. The Cash Constraint line (55% of AOV) is looser than 2.2x on every row, so 2.2x binds.

Supplier table in full, version A: orders 1, 2 and 3+ (subscriber rows at 15% off)

Tier	Order	Revenue	Coffee	Pick, carton, overhead	Shipping	Shopify fees	Physical gifts	Gift cost	Profit	Profit %
3 x 250g	One-time order 1	£36.99	£9.16	£2.31	£3.55	£0.97	None physical	£0.00	£21.00	56.8%
3 x 250g	Subscribe order 2	£31.44	£9.16	£2.51	£4.05	£0.85	Ritual Mug	£1.63	£13.24	42.1%
3 x 250g	Subscribe order 3 onwards	£31.44	£9.16	£2.31	£3.55	£0.85	None physical	£0.00	£15.57	49.5%
3 x 500g	One-time order 1	£71.49	£15.10	£2.51	£4.05	£1.77	Gold Blend	£2.75	£45.31	63.4%
3 x 500g	Subscribe order 2	£60.77	£15.10	£2.51	£4.05	£1.54	Ritual Mug	£1.63	£35.94	59.1%
3 x 500g	Subscribe order 3 onwards	£60.77	£15.10	£2.31	£4.05	£1.54	None physical	£0.00	£37.77	62.2%
3 x 1kg	One-time order 1	£109.49	£30.20	£2.91	£4.05	£2.61	Gold Blend, Mushroom Duo	£6.35	£63.37	57.9%
3 x 1kg	Subscribe order 2	£93.07	£30.20	£2.51	£4.05	£2.25	Ritual Mug	£1.63	£52.43	56.3%
3 x 1kg	Subscribe order 3 onwards	£93.07	£30.20	£2.31	£4.05	£2.25	None physical	£0.00	£54.26	58.3%

Supplier table in full, version B

Tier	Order	Revenue	Coffee	Pick, carton, overhead	Shipping	Shopify fees	Physical gifts	Gift cost	Profit	Profit %
3 x 250g	One-time order 1	£36.99	£9.16	£2.51	£3.55	£0.97	Scoop	£1.11	£19.69	53.2%
3 x 250g	Subscribe order 2	£31.44	£9.16	£2.51	£4.05	£0.85	Ritual Mug	£1.63	£13.24	42.1%
3 x 250g	Subscribe order 3 onwards	£31.44	£9.16	£2.31	£3.55	£0.85	None physical	£0.00	£15.57	49.5%
3 x 500g	One-time order 1	£71.49	£15.10	£2.71	£4.05	£1.77	Scoop, Gold Blend	£3.86	£44.00	61.5%
3 x 500g	Subscribe order 2	£60.77	£15.10	£2.51	£4.05	£1.54	Ritual Mug	£1.63	£35.94	59.1%
3 x 500g	Subscribe order 3 onwards	£60.77	£15.10	£2.31	£4.05	£1.54	None physical	£0.00	£37.77	62.2%
3 x 1kg	One-time order 1	£109.49	£30.20	£3.11	£4.05	£2.61	Scoop, Gold Blend, Mushroom Duo	£7.46	£62.06	56.7%
3 x 1kg	Subscribe order 2	£93.07	£30.20	£2.51	£4.05	£2.25	Ritual Mug	£1.63	£52.43	56.3%
3 x 1kg	Subscribe order 3 onwards	£93.07	£30.20	£2.31	£4.05	£2.25	None physical	£0.00	£54.26	58.3%

What changes in B. The Scoop joins every first order: £1.11 landed plus £0.20 pick, so £1.31 less profit per first order on every tier. 2nd boxes are the same in both versions, because the Ritual Mug ships with order 2 on every tier. Subscriber order 1 at 15% off is in the summary tables above.

Inputs and gaps

Input	Value	Source
Coffee, 250g bags	Stability £3.04, Darkfire £2.99, Rotate £3.13	BBC price list, _reference/bc-finance-cogs-bbc.md (13 Apr 2026)
Coffee, 1kg bags	£10.06, £9.89, £10.25	Same
Coffee, 500g bags	Gap. Half the 1kg cost used (£15.10 for three)	BBC has no 500g price yet (open item in the same file)
Pick, carton, overhead	£1.10 + £0.20 per extra item, carton £0.20, overhead £0.61. Duo samples count as 2 items.	Asendia rates, same file
Shipping	Royal Mail 24hr: £3.55 up to 1kg, £4.05 for 1-2kg	Same file
Shipping over 2kg	Gap. £4.05 used for every 3 x 1kg order (about 3-3.7kg). The real cost is higher.	Rate card stops at 2kg
Weights	Net coffee weight plus gifts. The mug is assumed at 0.35kg. A 3 x 250g box with packaging may tip over 1kg (+£0.50).	Assumption, not weighed
Shopify fees	2.2% + 20p	balance-coffee/gifting-strategy/front-end-profit-model.md
Gold Blend 250g	£2.75	Racketeer sheet, finance-due-diligence/unit-economics-model-2026-06.md
Mushroom Duo	2 x 10ml samples at £1.80 = £3.60	James, 2 Oct 2026
Scoop	£1.11 landed	gifting-strategy/accessories-order-plan-2026-09-30.md (3,000 units)
Mug	£1.63 landed. Open conflict: the supplier quote adds £0.53 for the two logos. That would make it about £2.16.	Same file
Brewing Guides, free shipping	£0 extra	Digital. Postage is already counted in every row.

Needs James before Ambiw builds it

#	Decision	Detail
1	Sign off the offer page	The Purity rewrite proposal is waiting on you (task brain-wxjk). This brief stays a draft until then.
2	Gift per tier	The page now says "up to £73" (the 23 Sep draft said £91). Confirm which gift goes with which tier before dev builds the gift logic.
3	Ladder prices	£36.99 / £71.49 / £109.49 (Save 23% / 15% / 20%). Compare-at is the live single-bag total, so no Shopify price change is needed. Confirm before dev builds the variants.
4	Live URL and launch date	Dev deadline is Fri 16 Oct, with the Meta Advertorial. The URL slug is yours to set (suggested: /pages/offer-clean-coffee-bundle).

Dev brief - Clean Coffee Bundle offer page DRAFT until James signs off. Drafted 23 Sep 2026, updated 1 Oct 2026 to match the current page.

What this is. Rebuild the staging page (preview above, google-bundle.pages.dev/offer-clean-coffee-bundle) as a live Shopify page on balancecoffee.co.uk, and point every Meta advertorial CTA at it. The staging page is the spec: match its copy word for word.

1. Key copy (check against staging)

Slot	Copy
H1	Clean Coffee Bundle + Free Gifts Worth Up to £73
Sub	Your morning ritual, with the lab report to prove it. Mould toxins measured to 0.5 parts per billion, and ours came in below. 600+ pesticides screened...
Hero CTA	SHOP CLEAN BUNDLE - UP TO 20% OFF (staging still says 25%; the top tier now saves 20%)
Announcement bar / gift badge	New customer offer: up to £73 in free gifts / Up to £73 in Free Gifts
Problem heading	You've Upgraded Your Kitchen. Not Your Coffee.
Offer heading	Clean Coffee Bundle + Up to £73 in Free Gifts
Guarantee	60 Days to Change Your Mind. Full Refund, No Questions.
Final	Your First Bag Arrives. You'll Know What's in Your Cup. / CTA: Get Clean Coffee

2. Do not carry across from the old page or the PDP template

Any "no jitters", "no crash" or "clean fuel" line.
"Every batch" in any form, including the product-template FAQ.
Antioxidant, "brain fog", "gut issues", "healthier", "immune support", "Over 85% of coffee", "No pesticides used".
"Zero", "none detected" or "no detectable" for mould, toxins, metals or pesticides (29 Sep lab-claim ruling). Pesticides read "600+ screened, all under the legal limit".
The "2,847 Customers Surveyed" block and its 78% / 87% / 94% figures.
Gallery images with burned-in claims. Check every image against this list.

3. Bundles and subscription (Skio)

Tier	One-time	Compare-at	Sub order 1	Renews at (15% off)	Cadence	Per cup (15g)	Savings + gifts
Discovery 3 x 250g	£36.99	£48.47	£36.99	£31.44	30 days	£0.74	£41.47
3 x 500g (default)	£71.49	£84.47	£71.49	£60.77	45 days	£0.71	£62.96
3 x 1kg	£109.49	£137.47	£109.49	£93.07	90 days	£0.55	£119.95

The staging page shows these prices. The Ritual Mug is the free gift in every subscriber's 2nd box, both gift versions.
Skio set-up (checked against Skio help docs, 2 Oct). (1) Full price on order 1, 15% off from order 2: selling plan discount 0% on the first order, Recurring Policy display 15% (display only). Then a Journey: trigger After Order, filter Order # = 1, action Update Product Price, Percentage Off 15%. Skio's guide shows a non-zero first-order discount, so test that 0% works on a test order. (2) Ritual Mug with order 2: Surprise & Delight rule, subscription-level, gift = Ritual Mug, fires on the 2nd order. The rule fires after order 1 processes. (3) Change size: Customer Portal, Actions & workflow, enable Swap product, and turn on 'Show in edit products page for customers' for all three bundles. All three bundles need the same selling plan intervals, or the swap fails. Confirm a swapped bundle keeps 15% off.
The first subscription order is full price; 15% applies from order 2. The Subscribe toggle shows "then £X from order 2", never a discounted first-order price.
One set of gifts per first order. Rebuy Smart Cart already stacks duplicate gifts on subscription first orders (brain-yt76): test this.
Tag these subscriptions source = meta-bundle.
Place 5-10 test orders across all three tiers, one-time and subscribe. Confirm price, gifts, cadence and tag on each. 48 hours after launch, sample live orders for the same four checks.

4. Advertorial CTAs and tracking

Every buy CTA in the Meta Advertorial (Purity listicle, task 6) links to this page's #bundles anchor; the lab CTA links to #lab-report. Both anchors must work on the live page.
Carry utm_source, utm_medium, utm_campaign, utm_content, utm_term and fbclid from the advertorial URL onto every CTA link.
The Meta pixel fires PageView on both pages, and ViewContent / AddToCart on this page, before any spend.

5. Acceptance (Ambiw sends James these)

Mobile (390px) and desktop (1280px) screenshots of the live hero, problem section, bundle cards and subscribe toggle.
One test subscription order per tier: full price on order 1, the order-2 renewal price in Skio.
An advertorial CTA clicked with ?utm_source=meta&fbclid=test lands on #bundles with both parameters intact.
Meta Pixel Helper shows PageView and AddToCart on this page.

Gift subscriptions upgrade Added 1 Oct 2026. Hard cutoff Sun 25 Oct, dev deadline Fri 23 Oct. Brief not yet written.

Three gift subscriptions, all built on the current Beans and Ground gift subscription structure. Live before the Christmas gifting window.

#	Product	Change
1	Beans and Ground Gift Subscription	Rename the current monthly gift subscription to this name. Structure stays as it is.
2	Mushroom Coffee Gift Subscription	Rename the Lion's Mane gift subscription. Rebuild it on the Beans and Ground structure, with the gifting-length slider in place of the fixed 3 months it has today. When it goes live, change the desktop nav tile's "3 months" line under Mushroom Coffee Gift Subscription to match (James 1 Oct).
3	Coffee Pods Gift Subscription	New product. Same structure as Beans and Ground, including the gifting-length slider.

Hard cutoff: Sunday 25 October. Dev deadline Friday 23 October, the last working day before it.

Also part of this launch: the Coffee Pods Gift Subscription needs its SKUs and ShipBob setup (product, SKU mapping and inventory) created before it goes live. Check the renamed products' SKU names and ShipBob mapping still match after the renames.

This tab holds the brief once it is built: prototype first, then "What changed vs live", then "Needs James before Ambiw builds it", then dev instructions after sign-off.

Images and videos for this page, original size: Asset library →

Blog article plus the new BC blog template (EXP-011). Replaces /blogs/blog/how-to-use-coffee-pods, same URL. v9, 8 Oct: James Bellis author line + About the author box (Person schema), one summary box as contents (folds to 4 on mobile, desktop rail appears after it), badges centred by an automated pixel/baseline check, real footage image pairs every 300 words or fewer, darker body text, product-first callout. Open full page →

What changed vs live Live: balancecoffee.co.uk/blogs/blog/how-to-use-coffee-pods

Change	Why
Generic pods explainer replaced by a Nespresso Original brew guide	Our pods only fit Original machines. The new cohort wants better cups, not pod history.
Seven habits up top, then one section per habit	Convenience buyers skim. Each section ends with a "Try this" step.
Americano rule, programming by taste, preheat	Your three tips. No ranking article explains why the lungo button tastes bitter.
Aluminium vs compostable section	Makes our freshness case with cited evidence, plus the England food-bin ban on compostable pods.
Recycling: empty, rinse, council kerbside	We are not a Podback member, so the article never sends readers there.
Flat white video embedded, FAQ, product CTA	Milk hack for latte drinkers. FAQ targets bitter-taste and Vertuo searches.
9 numbered steps, at-a-glance spec card	James 8 Oct: numbered lists are easier to follow. Spec card answers "just give me the numbers".
Tables and lists swapped for components	Cup-size visual, cup-size slider, Americano animation, storage do/don't tiles, aluminium vs compostable cards, troubleshooting picker, recycling steps. All text stays in the HTML for Google.
Inline product callout after step 4, 50/50 end banner	EXP-011 brief: products surfaced in context, a real button instead of a text link.
Live nav, logo and scrolling top bar	Pulled from top-bar.html (the nav source file) at build, so the blog matches the real site and updates with it.
Contents panel on the right, Balance Journal style	First 5 sections, "+7 more", reading progress %, related guides. Mobile keeps the open in-article list.
Schema: Article, Breadcrumb, Video. No FAQ schema	Google stopped showing FAQ rich results in May-June 2026 and HowTo in 2023.

Blog template: type scale and heading rules Applies to every blog post

Level	Used for	Font	Desktop / mobile
H1	Article title, one per page, Title Case	Canela Regular, +0.01em	56px / 40px
H2	Each main section (numbered step), sentence case	Canela Regular, +0.015em	42px / 30px
H3	Sub-section inside a step, sentence case	Area SemiBold	22px / 19px
H4	Minor label inside a sub-section	Area SemiBold, capitals, tracked, fern	13px / 13px
Body	Paragraphs and lists	Area Regular, 1.7 line height	19px / 17px
Standfirst	One line under the H1	Area Regular, grey	23px / 19px

Canela is for the H1 and H2 only, never under 28px, and always clearly bigger than the Area copy beside it. Cards, numbers and H3s are Area. Header is full width: breadcrumb, H1, standfirst and dates, then a 2:1 hero across the column, then body + contents rail. White page background; cards sit on the light tile colour. Byline shows Published and Updated dates (both read by Google), no author tag. Desktop gets a sticky contents rail. On mobile the contents list sits open in the article, because collapsed ones get missed (NN/g). Reading column 700px.

Needs James before it goes live

Sign off the copy and the template, or send notes on any section.
Template rollout: run as EXP-011 (all posts, 50/50 Intelligems test) or ship straight to every post.
URL decided (James, 8 Oct): same URL, /blogs/blog/how-to-use-coffee-pods. The new article replaces the old body there and keeps its rankings.
The flat white video is approved in Studio but not posted yet. For the live post, upload it to Shopify Files or link the posted Reel.
Live PDP FAQ says pods can go "via Podback schemes". We are not a member. That line needs removing on the store.

Images and videos for this page, original size: Asset library →

Meta pre-sell advertorial: Javvy's a39 listicle structure (9 reasons), run on our purity angle. All CTAs go to the Clean Coffee Bundle offer page. Open full page →

Compare: v2 from the rebuilt copywriter not live, for your pick

Written from scratch on the new advertorial method (one belief per reason). Legal is flag-only: 5 red and 21 amber lines are listed in legal-review-v2.md for your call. Images still sit in the current page's slots, so some pictures won't match their new reason yet.

Open v2 full page →

Story map: what each reason does

Same order and media slots as Javvy's a39 page. Read top to bottom as one story. Last column is the verdict.

#	Heading now	Job in the story	What it teaches	Objection it handles	Moves the reader to	Verdict
1	You heard what might be in your cup. You were right to ask.	Hook: recognition	Other people share the worry	"Am I being paranoid?"	Reads on, feels understood	Keep
2	Mould has months to grow before you brew	Cause	Drying, shipping and storage let mould grow; roasting doesn't remove it all	"Roasting kills it"	Believes the problem is real	Keep
-	(missing)	Agitation: why it matters	What the toxin is and what regulators say it does (WHO: kidneys in animal studies; IARC Group 2B; UK limit 5 µg/kg, no duty to test)	"So what if it's there?"	Wants a way to check	Gap: you asked for real problem-agitation-solution. The page names the cause, then jumps straight to trust
3	Can you trust what's in your cup right now?	Trust gap	You can't see, smell or taste it; the bag won't say	"How would I ever know?"	Wants proof	Keep. Last line repeats reason 4 ("sent ours to the experts")
4	We source some of the world's finest coffee and send it to an independent lab	What we do	Quality beans, dry storage, ISO 17025 lab	"Who checked it?"	Trusts the method	Keep. Legal check on "world's finest" open
5	Mycotoxins checked to 0.3 parts per billion. Ours came in below.	Proof 1: mould toxins	The main threat, measured precisely	"Is it actually clean?"	Core fear answered	Keep
6	Lead, arsenic and cadmium came in below the limits too	Proof 2: metals	A second risk they may not have known	(none new)	More confidence	Repeat: 6 and 7 are one beat ("and the other tests passed"). Merge to free a slot for the agitation gap
7	600+ pesticides screened. Residues within legal limits.	Proof 3: pesticides	Broad panel, honest result	(none new)	More confidence	Merge with 6. CTA "See the three roasts" does not fit a pesticide reason
8	Don't take my word for it. Read all 26 pages.	Verify it yourself	The full report is public	"Anyone can print clean on a bag"	Trust confirmed	Keep
9	A farm label tells you how coffee was grown. A lab report tells you what's in the bag.	Objection: organic	Certification checks farming, not the finished bean	"I'll just buy organic"	Sees why lab beats label	Keep. Overlaps 8 on "labels don't prove it"; keep 8 on the report, 9 on certification
10	Clean gets people to switch. Taste keeps them.	Objection: taste	Fresh roast, 3 roasts to try	"Healthy coffee tastes worse"	Wants to try it	Keep. Gap: no customer voice on taste anywhere on the page. Add one real review here
11	3 bags, 50 cups, 74p a cup. 60 days to decide.	Objection: price and risk	Price per cup, no subscription, 60-day refund	"Too expensive" / "What if I don't like it?"	Ready to buy	Keep
12	Reaching #12 means you want to know what's in your coffee.	Close	The offer and gifts	(none)	Clicks through	Keep

Proposed standard advertorial framework

Fixes the two gaps and the one repeat above. Every section has one named job. This goes into the rebuilt copywriter as the default for any advertorial; the rebuilt copywriter's re-run of this page will follow it, shown old vs new here.

#	Job
1	Hook: recognition
2	Cause
3	Agitation (legal fact bank lines only)
4	Trust gap
5	What we do
6	Proof: the main fear
7	Proof: everything else, one reason
8	Verify it yourself
9	Objection: the alternative they'd buy instead
10	Objection: taste, with a real customer voice
11	Objection: price and risk
12	Close and offer

Needs James before Ambiw builds it

More CTAs than our advertorial rule allows. The 8 Sep rule is one CTA per advertorial. This page copies Javvy: 9 inline links, an offer button and a sticky button. Keep, or cut to the offer box only?
Gift value. Page says "up to £73" (your call). The offer page script totals the gifts at £91.97. Pick one figure for both pages (brain-d0p9).
Byline "James B." with your photo, linked to your bio. No quote is put in your mouth. Approve, or run it unsigned.
Offer box photo includes the Mushroom Duo, which is only a gift on 3 x 1kg. Keep, or swap to a bags-only shot.
Live store variant reads "45 cups" for 3 x 250g. At your 15g dose it is 50 cups (74p at £36.99). Ambiw to relabel.
Better footage to shoot: BC sample sealed for the lab (3), back-of-bag label (2), phone showing the report page (6), roast-date stamp (8).

Images and videos for this page, original size: Asset library →

Rebuild of balancecoffee.co.uk/pages/healthy-coffee-science. Keep the same URL handle when it goes live. Open full page →

What changed vs the live page

#	Change	Why
1	Leads with proof, not the founder	The buyer comes for proof. The live page makes them read about James before they see a lab result.
2	Hero: "You can't see mould in your coffee. A lab can."	Purity is the #1 buying reason (Sep 2026 survey). Gap, proof and outcome all sit above the fold.
3	New problem block: why coffee needs testing	Explains drying, storage and the ochratoxin A legal limit in plain words, with no invented stats.
4	Lab results as 3 stat cards + link to the full report	"No detectable" mould, mycotoxins, lead, cadmium, arsenic; 600+ pesticide compounds within legal limits. Matches report 26-LA45200.
5	7 steps cut to 5: sourcing, storage, testing, roasting, mushrooms	"Science-backed research" step removed: it was a health-outcome promise with no content behind it.
6	Banned wording removed	"Toxin & mould-free guarantee", "100% free from", "chemical-free", "top 1% of farms", "verified by nutritionists", "antioxidants", "adaptogenic".
7	New: comparison table, reviews strip, 5-question FAQ, sticky mobile CTA	Answers the doubts ("every batch?", "what does no detectable mean?") before they leave.

Needs James before Ambiw builds it

Hero image badge says "No detectable mould, ochratoxin A, aflatoxins, lead, cadmium or arsenic". The 29 Sep lab-claim ruling bans "no detectable" wording (the hero sub line was already removed for this, 1 Oct). Remove the badge or reword it?
Clemmie Rose's exact title: the brain has both "Registered Nutritionist" and "Nutritional Therapist". Page says "Nutritionist".
Direct farmer relationships: the live page claims them, no source found. Page says only that you visit the farms.
Legal check applied (44 fixes). 1,800mg Lion's Mane, 20% less caffeine and climate-controlled storage kept (confirmed by James 26 Sep). Pesticide count reads "600+ compounds" (James, 29 Sep 2026: 600+ everywhere).
Nothing goes to Ambiw until James signs off.

Images and videos for this page, original size: Asset library →

Review prototype. Use the page-context buttons and brand-colour swatches inside the frame to compare placements and complete section colourways.

What changed vs live

Surface	Change	Why
Homepage	A four-card proof section after Shop Bestsellers.	Current homepage uses broad absence claims. This makes the testing scope clear at a glance, while preserving product discovery near the top. No lab-report CTA is included here: this is a fast trust cue, while the Lab Results page holds the detail.
Lab results	The same cards lead; five detailed report categories sit below in expandable rows.	Visitors get a readable answer first and can inspect the measurements and original report.
Science page	The proof section can replace its current three-stat lab block.	One set of report-backed results across the site avoids competing summaries.
Product page	A compact version sits below the purchase area.	The purchase flow stays clear while proof is close to the buying decision.

Report count and claim basis

The August 2026 report has five laboratory categories: mould and yeast; ochratoxin A; aflatoxins; heavy metals; pesticide residues. The four shopper cards combine the two toxin categories and lead with chemical pesticide checks. James confirmed monthly independent sample testing on 5 Oct 2026; the displayed result is still the August 2026 sample. It records 611 pesticide result lines: 560 individual analytes and 51 legal sum definitions, so the customer-facing line is “600+ pesticide checks,” not “611 different pesticides.” It also records five individual mould toxins plus a total aflatoxins result, three metals, and two microbial results: 622 reported result entries across the five categories (611 + 6 + 3 + 2). Some pesticide entries are legal sums of other analytes, so this is not a claim that 622 independent substances or tests were run. Pesticides had one quantified residue definition within its legal limit; the other named groups were below the laboratory’s measurement level. Report: 26-LA45200, 21 August 2026.

Needs James before Ambiw builds it

Approve the shared four-card direction, colourway and placement across the homepage, lab results, science and product pages. This is a CRO Hub review prototype; nothing has changed on Shopify.
Confirm whether the lab page should keep every technical method line visible under the expandable results or rely on the downloadable original report.
Current homepage “toxin-free”, “zero mould” and “pesticide-free” wording conflicts with the 29 Sep claim ruling. Correct that shared baseline before any homepage experiment.
Final copy and claim sign-off precede the Ambiw handover. The report covers one submitted beans and ground sample; it does not identify a SKU or establish every-batch results.

Images and videos for this page, original size: Asset library →

Rewrite of balancecoffee.co.uk/pages/our-story as your first-person mould story. Open full page →

What changed vs the live page

#	Change	Why
1	New story arc in your voice	Before (the coffee guy) - the fog - mould - why coffee - building it - the lab - a note from you.
2	Your mould exposure and symptoms	Brain fog, tired-but-wired, broken sleep, headaches, low mood, overload. Framed as your experience, never as something the coffee fixed.
3	"Why coffee?" education block	Drying and storage explain the mould risk; ochratoxin A has a legal limit in coffee (EU 2023/915, EFSA 2020). No invented percentages.
4	Lab proof with the report page image	Turns the story into evidence: no detectable mould, mycotoxins, lead, cadmium or arsenic; 600+ pesticide compounds within legal limits.
5	Removed unsourced claims	"Top 5% of farms", "85% of coffee", "25% more paid", "free from nasties", "healthiest coffee available".
6	Sign-off note + "not medical advice" line	Keeps the story honest and inside health-claim rules.

Needs James before Ambiw builds it

How long the symptoms lasted ("about two years"), whether you moved out of the damp, and your caffeine habits wording.
Clemmie is happy to be named as backing the coffee.
The crash clip for the After section (placeholder on the page).

Replaces the Shopify native cookie banner. Shown over a screenshot of the live homepage. Open full page →

What changed vs the live banner

#	Change	Why
1	380 x 142px card bottom-left on desktop, no icon, slim 40px buttons (was about 1000 x 240px across the page). Phones: bottom sheet, 138px, about 16% of the screen	Bottom-left notices got the most interaction and top bars the least (Utz et al. 2019, 14,135 visitors, 78% mobile, German non-shop site). Blocking centre modals did not lift consent (Nouwens et al. 2020, 40 people). Google exempts legally required notices but asks they take a small fraction of the screen
2	Area heading "Cookie Policy", one-line body	Plain statement title, no question mark: James (28 Sep) - a question creates hesitation and pulls the visitor off their journey. No published test shows a title wording moves consent either way. Every study that changed banner wording found no or small effects (Utz 2019, 32,225 visitors, 78% mobile; van Bavel 2016, 602 people; Laine 2021). Buttons, position and visual design move consent (Ma & Birrell 2022; Berens 2022). Never use loss wording on Reject
3	Reject and Accept on the first layer, same size, both filled pills	ICO: refusing must be as easy as accepting. Hiding Reject lifts consent 22-23 points but is non-compliant in the UK
4	Manage choices = small grey link beside Privacy policy (no extra row); choices panel with 4 categories, all optional ones off by default	UK GDPR: no pre-ticked boxes, granular choice
5	Non-blocking fixed overlay, no page lock, async load	No cookie wall (ICO); an overlay does not shift layout (web.dev)

Signed off

James signed off this version on 28 Sep 2026: light-green Reject, dark-green Accept, title "Cookie Policy", grey Privacy policy + Manage choices links. Build exactly this.

Dev instructions - cookie pop-up (read first) Signed off 28 Sep 2026. Source: consent-modal.css + consent-modal.js on google-bundle (view source).

#	Area	Instruction
1	Replace the banner	Switch OFF the Shopify native cookie banner (Settings > Customer privacy > Cookie banner). Add consent-modal.css and consent-modal.js to the theme (theme.liquid, JS deferred). Keep the bc-cc class prefix.
2	Consent API	Consent is written with window.Shopify.customerPrivacy.setTrackingConsent: preferences, analytics, marketing, and sale_of_data = marketing. The script loads the consent-tracking-api feature and shows the pop-up only when shouldShowBanner() is true (regions that need consent, until the visitor chooses).
3	No tracking before consent	No Meta, Google Ads or other marketing pixel fires before Accept. Check every app pixel and Custom Pixel respects Shopify consent. Test in a UK session: load the site, do nothing, confirm no marketing requests in the network tab.
4	Layout	Desktop: 380px card fixed bottom-left, 24px from the edges, about 142px tall. Phones (600px and below): full width minus 8px each side, 8px from the bottom, about 138px tall. Non-blocking: no overlay, no scroll lock, no close X.
5	Design	White card, 14px radius (12px phones), soft shadow. Title "Cookie Policy" Area 800 15px. Body Area 400 13.5px grey #555. Privacy policy (/policies/privacy-policy) and Manage choices are grey underlined text links on the body's last line. Buttons 40px pills, uppercase Area 800 13.5px: REJECT ALL Morning Mist #D1E0D7 with Obsidian text; ACCEPT ALL Forest Fern #516E5A with #F5F5F5 text. Same size, side by side, Reject on the left.
6	Choices view	Manage choices swaps the card to 4 rows: Strictly necessary (Always on), Preferences, Analytics, Marketing. Toggles start OFF. Buttons: Reject all, Accept all, Save my choices. Back link returns to the main view.
7	Footer link	Add "Cookie settings" to the site footer with the attribute data-bc-cookie-settings. It reopens the choices view (legal requirement: withdrawing consent must be as easy as giving it).
8	Performance	Load async, CSS-positioned. Run a mobile Lighthouse check and confirm the pop-up is not the LCP element and adds no layout shift.
9	Do not	Do not change copy or colours. Do not ship the preview switches (?cc-demo, ?cc-prefs) or the review frame (consent-modal.html). Do not pre-tick any toggle.

Images and videos for this page, original size: Asset library →

Open on its own page →

Dev brief EXP-018 - new buy box vs live, mobile A/B test (read first) 28 Sep 2026. Arm B = the "New" column below. James creates and launches the test in Intelligems.

Raised: 28 Sep 2026 by bc-cro-agent, after James's sign-off ("launch a/b test brief for dev team please", 28 Sep).

For: Ambiw (Shahid, Saleha), routed by James. Agents hold no Shopify or Intelligems write path. James creates and starts the Intelligems test.

Bead: brain-9d74.

1. What we are testing

The redesigned buy box (Variant D on the preview) against the buy box live today, on phones only.

Arm	What the visitor sees	Preview
A - Control	The live PDP, unchanged	balancecoffee.co.uk/products/stability-blend
B - Variant	New buy box + new FAQs, as built on the preview (Variant D: with the month-1 free gifts strip)	google-bundle.pages.dev/pdp-faq-stability?view=mobile&variant=d

Build only the "New" column. The tools bar, the "Now" column and "Developer notes" toggle on the preview are review-only.

Hypothesis. If we show the phone buy box with cups and price per cup on each size tile, one forward-only savings line and a simpler subscribe card, then more orders will be 1kg and more will be subscriptions, with revenue per visitor at least flat, because each step up states one concrete saving instead of a block of numbers.

2. Risk zone and call path

Hot zone. The buy box changes pricing display and subscription selection.

Item	Setting
Risk flag	Hot
Call path	Full Call: 14+ days AND revenue per visitor (RPV) probability to beat baseline 95% or more
Monitoring	Daily for the first 3 days, then every 2-3 days
Kill rule	Kill if control RPV leads by more than 1.5x after 48 hours, or variant revenue is more than 15% below control for 48+ hours in a row with 100+ visitors per arm
Package exception	Hot zones normally test one variable. This is a package test (whole buy box vs live) under the low-traffic rule: phone PDP traffic cannot support one test per element. Sub-choices are judged later on click data, not extra arms
One Hot test at a time	James confirms in Intelligems that no other Hot-zone test (sidecart, cart, checkout, pricing) is live at launch

3. Intelligems setup (James, under 10 minutes)

Field	Value
Test name	EXP-018 PDP buy box v2 vs live (mobile)
Test type	Template test: control = current product template, variant = `product.buybox-v2` built by Ambiw
Targeting: device	Mobile only
Targeting: pages	These 9 product pages (all share the 250g / 500g / 1kg + grind buy box): stability-blend, darkfire-energy, halcyon-decaf-coffee, rotate-espresso, level-up-blend, aurora-reserve-single-origin, kenya-single-origin, indonesia-single-origin, papua-new-guinea-single-origin
Excluded	lions-mane-mushroom-coffee (no grind, different price ladder), bundles, subscription-only products, christmas-coffee, gift subscription
Split	50 / 50
Primary metrics	1kg share of coffee orders; subscribe rate (share of orders that start a subscription)
Decision metric	RPV (net revenue per visitor)
Guardrails	RPV not below control; conversion rate; AOV
Start	A Monday, with no promo or email offer live on these products
Hard stop	Sun 22 Nov 2026, before any Black Friday promo

4. How the variant is delivered

Recommended: an alternate product template (product.buybox-v2.json) with a new buy box section and the new FAQ section. Intelligems template testing serves it to the variant arm. Why: it keeps the variant's markup, JS and CSS out of the live template, so control stays byte-identical and the revert is one click. A CSS/JS overlay on the live template is not recommended: this change replaces the selector, pricing logic and subscription card, not only styling.

5. Build requirements

Source files: balance-coffee/marketing/lp-html/pdp-faq-compare.js (markup and logic) and pdp-faq-compare.css. Copy source: balance-coffee/marketing/pdp-faq/stability-pdp-faq-copy-2026-09-25.md. Where the preview's Developer notes and this brief disagree, this brief wins.

All prices, cups and savings come from the product's own Shopify variants and selling plans. Nothing is hard-coded to Stability Blend. The same template must render correctly on all 9 pages.
Cups per bag come from the variant option name ("250g (15 cups)"). If brief 009's cup-count correction lands first, the tiles follow it automatically.
Order tracking: on add to cart, both arms write the visitor's Intelligems test group to a cart attribute _exp018_arm (control / variant). Without it, 1kg share and subscribe rate cannot be split by arm (Shopify orders carry no test group or device). Ambiw confirms the Intelligems JS call that returns the group.
The subscription add to cart uses the product's existing Skio selling plan, same as live. No change to plans or prices.

6. Acceptance criteria (variant arm, at 390px)

#	Check
1	No text below 13.5px (only the step numbers in circles may be 13px)
2	Size tiles show weight, "(N cups)" and price per cup for the selected plan. Tiles have 2px extra padding top and bottom so the 1kg tile sits clear of the Best Value badge
3	20px between the size tiles and the savings line on phones
4	Savings line shows the forward saving only, like-for-like, one line, and tapping it selects the next step. One-time: 250g "Upgrade to 500g: save £X", 500g "Upgrade to 1kg: save £X", 1kg "Subscribe: save £X". Subscription: 250g and 500g "Upgrade to ...: save £X", 1kg "Best price: just Xp a cup". No "£X saved" part. Stability Blend values: £4.99 / £9.99 / £6.00 one-time; £4.24 / £8.49 / 57p subscription
5	Subscribe & Save card has no "Save £X" pill; 15% sits in the title
6	Price per kg: small grey £/kg under every price (subscribe and one-time, all sizes). Required by the Price Marking Order per the legal audit of 26 Sep (brain-mb15), which overrode James's earlier removal. Remove only if bc-legal-agent confirms a small-trader exemption
7	One-time row: grey "Full price" inline left of the price on 250g only. Row height is identical across sizes
8	Free gifts strip inside the Subscribe & Save card: indented to the card title text (28px phones), 60px thumbnails stacked over text on phones, struck value plus green FREE label. Barista course value £29.99; ritual mug shows FREE only. Footer link "View upcoming gifts"
9	One-time is preselected. Tapping anywhere on the subscribe card selects subscription
10	Add to cart reads "ADD TO CART - £{selected price}", no struck price
11	Delivery line above the button: "Freshly roasted - arrives in 2-3 working days". Trust row under the button: "60-day money-back guarantee" and "Free UK delivery over £25", then payment icons
12	FAQ order: purity ("What makes Balance Coffee different?"), cost, taste/origin/caffeine, shipping and returns. Paragraphs max 3 lines on phones, answers use full card width. Caffeine footnote sits inside the green box. No source note on the cost FAQ. Shipping FAQ guarantee line matches the live refund policy wording at launch
13	Control arm renders the live PDP unchanged, apart from the hidden cart attribute
14	Checked on iPhone Safari and Android Chrome on 3 of the 9 pages, one blend, one decaf, one single origin

7. Forecast and call date

Input	Value	Source
Mobile PDP views, 4 largest in-scope PDPs	14,146 in 90 days = 157 a day (Level Up and single origins not in the pull)	GA4, 17 Jun - 14 Sep 2026, `bc-menu-bestseller-slots-2026-09-15.md`
GA4 capture rate	About 40% of Shopify orders, so real visitors likely 157-393 a day	same file
1kg share baseline	43% of single-bag web orders (428 of 991), all devices	Shopify, 17 Jun - 14 Sep 2026, `bc-coffee-order-mix-analysis-2026-09-15.md`
Subscribe rate baseline	Data gap	Pull from Shopify before launch

Forecast (example launch Mon 12 Oct 2026):

Milestone	Date	What it can show
First Full Call check	Mon 26 Oct (day 14)	Only a very large RPV effect (60%+)
Hard stop	Sun 22 Nov (day 41)	1kg share shift of about 9-14 points; RPV only if lift is 39%+

RPV will almost certainly not reach 95% before Black Friday. At 157-393 visitors a day, 41 days gives 3,200-8,100 visitors per arm. RPV needs roughly 5x the visitors of conversion rate. The mechanism metrics read faster because each order counts. Decision for James at launch: at the hard stop, ship the variant if 1kg share or subscribe rate is up at 95% and RPV is not below control; otherwise revert. This is an exception to the Full Call and needs his explicit yes.

Formula (to refresh with live Intelligems numbers after 7 days): visitors per arm needed = 7.84 x 2 x p(1-p) / (p x lift)^2 for conversion; x5 for RPV. Days = visitors per arm x 2 / visitors a day.

8. Sizing

Scenario	RPV lift on mobile single-bag revenue	Annual £
Conservative	+2%	about £1,900
Target	+5%	about £4,800
Aggressive	+10%	about £9,600

Base: 991 single-bag web orders in 90 days x 70% mobile share (GA4 purchase share) x £34.56 AOV = about £7,980 a month. Estimate: Shopify orders carry no device field. Subscription lifetime value from a higher subscribe rate is not included.

ICE: Impact 7, Confidence 6, Ease 5 = 210. Risk: Hot.

9. Open items before launch

Item	Owner
Month-1 gifts (barista course + ritual mug) confirmed as what Skio sends on a first subscription order; barista course at £29.99 in store (brain-d0p9). If not confirmed, build Variant B (no gifts strip) instead	James
Shipping FAQ guarantee line matches the refund policy live at launch (brain-5gqd)	James
No other Hot-zone test live in Intelligems	James
Subscribe rate baseline pulled	bc-data-analyst-agent

10. FAQ icon set for other PDPs

Added 1 Oct 2026 at Ambiw's request. Three new icons for the Ingredients, How to Brew and How is Balance Coffee Healthier? FAQs on other product pages, in the same format as the EXP-018 FAQ icons.

Format: white line glyph, 24 x 24 viewBox, 2px round stroke, on a solid Fern (#40685D) circle. Circle 32px with a 17px glyph on desktop, 28px with a 15px glyph on phones (.nw-ic in pdp-faq-compare.css). Use the same .nw-ic wrapper so all FAQ icons match.

FAQ heading	Status	File
What makes Balance Coffee different? (purity)	In EXP-018 build	faq-icon-lab.svg Inline SVG `<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="#fff" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M9 3h6M10 3v6L4.5 19a1.5 1.5 0 0 0 1.3 2h12.4a1.5 1.5 0 0 0 1.3-2L14 9V3"/><path d="M7 15h10"/></svg>`
Taste, origin and caffeine	In EXP-018 build	faq-icon-cup.svg Inline SVG `<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="#fff" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M4 9h13v5a5 5 0 0 1-5 5H9a5 5 0 0 1-5-5z"/><path d="M17 11h1.5a2.5 2.5 0 0 1 0 5H17"/><path d="M8 3.5c0 1.5 1.5 1.5 1.5 3M12 3.5c0 1.5 1.5 1.5 1.5 3"/></svg>`
Cost	In EXP-018 build	faq-icon-pound.svg Inline SVG `<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="#fff" stroke-width="2.2" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M16.5 6.5A3.5 3.5 0 0 0 9.5 7v4.5c0 3-1 5.5-3 7.5h11"/><path d="M6.5 12.5h7"/></svg>`
Shipping and returns	In EXP-018 build	faq-icon-truck.svg Inline SVG `<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="#fff" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><path d="M1 4h13v11H1zM14 8h4l4 4v3h-8z"/><circle cx="5.5" cy="17.5" r="2"/><circle cx="17.5" cy="17.5" r="2"/></svg>`
Ingredients	New, 1 Oct	faq-icon-bean.svg Inline SVG `<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="#fff" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><ellipse cx="12" cy="12" rx="6.5" ry="9" transform="rotate(35 12 12)"/><path d="M16 6.3C13 8.5 14.5 11 12 12s-1 3.5-4 5.7"/></svg>`
How to Brew	New, 1 Oct	faq-icon-brew.svg Inline SVG `<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="#fff" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M4.5 3h15l-4.5 6.5h-6z"/><path d="M12 9.5v2"/><path d="M6 13.5h12v3a4.5 4.5 0 0 1-4.5 4.5h-3A4.5 4.5 0 0 1 6 16.5z"/></svg>`
How is Balance Coffee Healthier?	New, 1 Oct	faq-icon-sprout.svg Inline SVG `<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="#fff" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" aria-hidden="true"><path d="M12 21v-9.5"/><path d="M12 14.5C12 10 9 7.5 3.5 7.5c0 5 3 7 8.5 7z"/><path d="M12 11.5C12 6.5 15 4 20.5 4c0 5.2-3 7.5-8.5 7.5z"/></svg>`

Copy check before build: the heading "How is Balance Coffee Healthier?" uses "healthier", which is on the banned-claims list in this hub (Meta Offer Page brief, section 2). The icon is neutral and works with any heading. The heading wording needs bc-legal-agent sign-off before it goes live.

Desktop 1440 Shop menu open on load. Hover or click Shop / Learn to try it.

Desktop 1440 - Mushroom Elixirs hovered tiles swap per category

Desktop 1280 Shop menu open

Desktop 1440 Learn menu open

Dev instructions - desktop nav (read first) 22 Sep 2026. Everything needed to build arm B. Prototype source: /top-bar (view source for markup and CSS).

Mobile A favourites first

Mobile B categories first

Dev notes (open threads)

Brewing guides hub Rebuild of /pages/coffee-brewing-guides. 12 guides in 4 groups. Bean-to-Cup and Filter Machine are coming soon. Open full page →

What changed vs live Live Meta landing today: /products/espresso-taster-pack

Decisions in force (James, 2 Oct, on Neel's structure)

Gift ladders

First-order profit, break-even ROAS and profit at the 2.2x max CPA, version A gifts

Same, version B gifts (Scoop added to every first order, from late Nov)

Supplier table in full, version A: orders 1, 2 and 3+ (subscriber rows at 15% off)

Supplier table in full, version B

Inputs and gaps

Needs James before Ambiw builds it

Dev brief - Clean Coffee Bundle offer page DRAFT until James signs off. Drafted 23 Sep 2026, updated 1 Oct 2026 to match the current page.

1. Key copy (check against staging)

2. Do not carry across from the old page or the PDP template

3. Bundles and subscription (Skio)

4. Advertorial CTAs and tracking

5. Acceptance (Ambiw sends James these)

Gift subscriptions upgrade Added 1 Oct 2026. Hard cutoff Sun 25 Oct, dev deadline Fri 23 Oct. Brief not yet written.

Calm Coffee PDP launch Added 1 Oct 2026. Brief not yet written.

What changed vs live Live: balancecoffee.co.uk/blogs/blog/how-to-use-coffee-pods

Blog template: type scale and heading rules Applies to every blog post

Needs James before it goes live

Dark Fire Pods launch Added 1 Oct 2026. Brief not yet written.

Compare: v2 from the rebuilt copywriter not live, for your pick

Story map: what each reason does

Proposed standard advertorial framework

Needs James before Ambiw builds it

What changed vs the live page

Needs James before Ambiw builds it

What changed vs live

Report count and claim basis

Needs James before Ambiw builds it

What changed vs the live page

Needs James before Ambiw builds it

What changed vs the live banner

Signed off

Dev instructions - cookie pop-up (read first) Signed off 28 Sep 2026. Source: consent-modal.css + consent-modal.js on google-bundle (view source).

Asset library For Ambiw. Every image and video used in the designs in this hub, at the highest resolution we hold.

Dev brief EXP-018 - new buy box vs live, mobile A/B test (read first) 28 Sep 2026. Arm B = the "New" column below. James creates and launches the test in Intelligems.

1. What we are testing

2. Risk zone and call path

3. Intelligems setup (James, under 10 minutes)

4. How the variant is delivered

5. Build requirements

6. Acceptance criteria (variant arm, at 390px)

7. Forecast and call date

8. Sizing

9. Open items before launch

10. FAQ icon set for other PDPs