01 Apr – 27 Jul 2026 · 118 trading days

The site is
permanently on sale.

Every number here comes from 28,478 order line items, 5,876 orders, the Shopify discount report and UK landing-page sessions over the same 118 days. This version corrects three things the previous cut got wrong: margin was measured against VAT-inclusive revenue, discount-code orders were counted several times over, and a strong April-to-June story hid a sharp July.

01 Revenue waterfall

RRP down to what actually stays in the business. The VAT step matters — it was missing before, and it is the single biggest correction on this page.

Margin is 28.9%, not 40.8%. The earlier figure compared VAT-inclusive revenue against VAT-exclusive cost, which counted £46,290 of HMRC's money as gross profit. On £231,448 of real ex-VAT revenue against £164,529 of goods, the business kept £66,918 across 118 days — about £567 a day before a single fixed cost, wage or ad spend.

Read margin as directional, not audited. Costs here are catalogue standard cost, not landed cost — no freight, duty, storage or payment fees. 2.1% of line items carry no cost at all and count as pure margin, so the true figure is slightly lower again. Deep-discount and outlet stock was often bought in cheaper than catalogue cost, which makes those lines look worse than they are. Every margin number on this page is an estimate built on those assumptions. Revenue, orders, units, sessions and discount depth are measured, not estimated.

02 Margin by product type

Sorted by revenue. Margin is now calculated on ex-VAT revenue, so these are far below the numbers you saw before. "Off RRP" is the realised discount including both automatic markdown and codes.

Product type	Units	Net revenue	Off RRP	Margin %	Margin £

Colour is the only real profit engine. UV Gel Polish Colour returns 44.7% on £80,102 of revenue. Underneath it, three categories are sold below cost once VAT is removed — One Step Gel Polish at −11.4%, Extend Care 5in1 at −9.1% and Extend 5in1 at −24.6%. Nail Drill returns 7.9% and Starter Set 7.2%. Those five categories are 27% of revenue and they are not accidents of a sale — they are the standing list price.

03 Trading states

Every one of the 118 days had something running. These are the three states that existed, with the average realised discount for each — the answer to whether deeper cuts drive better days.

Trading state	Days	Avg off RRP	Net / day	Margin / day	CVR	AOV	£ / session

Depth is not what makes a good day — the event is. Sitewide days average 41.5% off and promo days 41.4% off. Same discount, to within a rounding error. Yet sitewide days earn £2,892 against £2,225 a day, £4.95 per session against £3.40, and £711 of margin a day against £529. Twenty-six sitewide days did the work of forty promo days. Category-markdown days ran the shallowest discount of the three at 38.6% and still held the second-best AOV. Cutting deeper has never been the lever.

04 Discount codes

Rebuilt from order level. The previous version counted a code once per line item and once per order, inflating every row — SOPHIE30 was shown as 463 orders and £21,130; it is 113 orders and £6,420.

Most "codes" are not campaigns. Of 809 distinct codes in the window, only 14 have five or more orders. The rest are one-time codes generated per customer by the signup pop-up and channel links — strings like HVQVMDV8 or 0ce471c3071a, each used exactly once. They are not a promotion you ran; they are one voucher per person. Grouping them as if they were a campaign is what produced the phantom numbers. The old "recorded £" column has been dropped entirely: Shopify only writes an order-level discount amount on some orders, so it showed £2 against a 30% code and measured nothing.

Code	Orders	Net revenue	AOV	Off RRP	Margin %	New cust.	Days live

Trade codes are the only ones that make money. SALON25 posts a £220.97 AOV at 40.5% margin and SALON20 £78.04 at 41.2% — the two best commercial outcomes in the table, from a list of 1,300 people. Flat percentage-off codes do the opposite: FLASH40 returns 18.6%, EXTRA20 19.3%, LAMPS40 7.5%. SOPHIE30 sits in between at 29.2% but brings 43% new customers — the only code on the page that is genuinely buying customers rather than discounting existing ones. Codes touched just 1,227 of 5,876 orders (20.9%); the other four-fifths of the business bought at a price that was already cut.

05 Where people land

74,021 UK sessions across 4,075 landing pages, 01 Apr – 28 Jul. Cart-add and checkout rates are share of sessions that started on that page.

Landing page type	Pages	Sessions	Share	Cart add	Reached checkout

Promotional collection pages are the weakest thing on the site per visit. Promo collections pull a higher cart-add rate — 27.0% against 21.3% for evergreen collections — and then convert to checkout at 9.9% against 10.1%. All that extra browsing turns into nothing. /collections/full-price converts at 19.5%, the highest of any collection on the site, on 241 sessions nobody sends traffic to.

Collection pages · promotional vs evergreen

Collection	Sessions	Cart add	Reached checkout

Product pages people actually arrive on

Product page	Sessions	Cart add	Reached checkout

The three deepest-discount collection pages — up-to-50-off-hot-deals (7.7%), spring-price-cuts (7.1%) and summer-colours-up-to-50-less (7.0%) — sit at the bottom of the checkout table while taking 2,342 sessions. bestdeals is the exception at 1,855 sessions and 12.1%: it is a curated hub, not a percentage. Meanwhile 8,950 sessions land on blog articles and reach checkout 0.6% of the time — 12% of all traffic arriving on pages with no path to a basket. E-files takes 793 sessions at a 6.2% cart-add, the worst of any major collection: high intent, wrong page.

06 Who comes back

First-ever buyers only — customers whose lifetime order count matches their orders inside this window. 2,048 of 4,730.

Starter Set buyers repeat at 18.8% versus 9.9% for colour-only, and return in 12 days versus 29. The razor-and-blade model works. But on the corrected margin basis the entry order nets about £7, not £20 — so the acquisition case is thinner than it looked and depends entirely on the second order arriving. It is still acquisition spend, and it still has never been budgeted as such.

07 Daily rhythm

Weekly is the readable view — 17 full weeks plus a part week. Switch to daily for the spikes. The gold line is margin; bars turn lacquer when the day ran deeper than 44% off RRP.

View

Net revenue Margin Deep-discount day — hover any bar

The shape is four flat months and a falling July. Weeks hold between £3,600 and £5,600 with no sustained climb, then the last three weeks of July drop away. Margin tracks revenue almost exactly, which is the problem: there is no week where the business earned more without selling more.

08 Day log

All 118 trading days. Click any row for what was live, which named codes fired, and the email detail.

Date	State	Sessions	Orders	New	CVR	AOV	Net rev	Margin %	Off RRP	Email

09 Month by month

Net revenue by product type. July shows 27 actual trading days and a full-month estimate built by projecting the four missing days (28–31 July) at July's own average for each weekday.

Product type	Apr	May	Jun	Jul 1–27	Full Julyest	vs Jun

The July estimate is a projection, not a result. It assumes the last four days of July trade like July's other Mondays to Thursdays. If a promotion runs into month-end it will beat this; if the month ends quiet it will miss. Treat £59,910 as the mid-point, not the number.

July reverses the growth story completely. April to June looked healthy — colour up 19.2%, base up 19.4%, top up 17.0%. Then July is tracking £59,910 against June's £77,010, down 22%. The fall is worst exactly where the money is: UV Gel Polish Colour is down 35.4%, the one category returning 44.7% margin, and Nail Drill is down 45.4%. Only Starter Set (+2.8%) and Base (+4.5%) hold. Reading the April-to-June column alone would tell you the consumable engine is compounding; the July estimate says the engine stalled the moment the promotional calendar thinned out.

10 Email

86 sends in date order, oldest first. Gold rows are the top decile by revenue per recipient — the only measure that compares a 1,300-person PRO send with a 22,000-person blast fairly.

Sent	Campaign	Recipients	Open	Click	Orders	Revenue	£/recipient

The strongest emails share one shape: a deadline, a reason, and a named offer. The best send in 118 days was APR 13 "30% extended" at £0.215 per recipient — an extension email, which is a deadline being moved, and it beat the launch. Second is JUN 06 "HOME First 40 Weekend Flash Sale" at £0.170, third JUN 30 "ALL Flash 40% off everything" at £0.142. All three are short, dated, sitewide events. PRO sends occupy four of the top eight on a list of ~1,300 — MAY 06 at £0.169 and MAY 22 at £0.157 beat almost every mass send. Blog blasts to 17–23k return £0.008–£0.036; one returned a single order from 5,917 sends. The pattern is not list size, it is whether the email has an event in it.

11 Anatomy of a great day

The twelve best margin days in 118 on the corrected basis, and what was switched on.

£803

margin/day · two email sends

£660

margin/day · one send

£579

margin/day · no send

Tue

best weekday by margin

Sat

worst weekday by margin

Date	What was live	Net rev	Margin	Margin %	AOV	Sends

Shape A · The volume day

Mechanic: Sitewide, 24–48h
Emails: 2
Typical CVR: 9–14%
Typical AOV: £50–84
Best example: 07 Apr · £5,036 · 41% margin

Everyone buys and baskets stay large. Needs a hard deadline and a final-hours reminder. Lands best Tue–Thu. This is the only shape that produced a £5,000 day.

Shape B · The basket day

Mechanic: Hardware or treat framing
Emails: 1
Typical CVR: 6–9%
Typical AOV: £62–88
Best example: 23 Jun · £87.85 AOV · 39% margin

Fewer buyers, far bigger baskets, and margin holds because the discount sits on one category instead of the whole catalogue.

The anti-pattern

Mechanic: Deep cut on hardware
Duration: Open-ended
Margin: 10–17%
Typical AOV: £40–50
Worst example: 19 Apr · 10.5% margin

Electronics at −50% and "Pick any 3 for £19.99" produced the four lowest-margin trading days in the window. High revenue, almost no profit.

The pattern behind every strong day is an event with a date on it, announced twice. The three best days all ran a sitewide with a second email; the three worst-margin days all ran a deep category cut with no deadline. Note what is not here: no full-price day appears in the top twelve, because there has never been one.