This is the working, not the plan. Every figure below is tagged with where it came from, so we can go line by line and agree it or correct it. The numbers that actually change the answer are isolated at the end, with what each one does if it is wrong.
is breakeven for September at the £260,154 goal. To get there the ads have to deliver 1.93x new-customer ROAS at £35 per new customer. Today they are delivering 1.60x at £42.
Everything downstream rests on this. Revenue figures include VAT, and two of the largest costs are per order rather than per pound, so a single margin percentage does not work. On the actual mix sold 1 to 08 September:
| Per £100 of sales | Amount | Where it comes from |
|---|---|---|
| VAT | -£16.67 | 20% VAT rate measured Financial dashboard, financials.vatRate |
| Product cost | -£31.61 | 60% blended margin on ex-VAT product revenue measured Dashboard margins.blended, weighted across 10 SKUs |
| Fulfilment | -£8.91 | £4.50 per order, 1.98 orders per £100 measured Dashboard shipping.fulfillmentCostPerOrder |
| Payment processing | -£2.20 | 1.7% plus £0.25 per order measured |
| Contribution | £40.61 | derived This is the 40.6% used everywhere below. |
Cross-check: this same calculation reproduces the financial dashboard's own profit line to the penny for every closed month of 2026. January £9,029, February £4,213, March -£4,947, April -£3,403, May -£1,483, June £5,368, July -£387, August £3,726. If this row is wrong, the dashboard is wrong too.
Fixed overhead is £28,896 a month whatever we spend, so once contribution and overhead are known, everything left is available for advertising. At breakeven, every pound of budget is a pound of profit.
The ad budget currently in the sheet is £76,204, which is £523 under that figure. For a month this size those are the same number, so the budget needs no change.
Ads do not touch this line, so it comes out of the goal first. The forecast is the Recharge schedule multiplied by the share that actually bills, because people skip and churn between the schedule being set and the charge running.
Recharge charges contain two different things: renewals, and the first order when a new subscriber signs up. First orders are completed checkouts, so they always bill, and they belong to new customers rather than to this line. Measuring realisation across both together flatters it. Stripping first orders out, using Shopify order tags, and the two sources reconcile to within £1,000 a month:
| Month | Renewals scheduled | Renewals billed | Realised |
|---|---|---|---|
| April | £54,995 | £49,247 | 89.5% |
| May | £49,511 | £45,367 | 91.6% |
| June | £56,858 | £51,420 | 90.4% |
| July | £64,134 | £57,633 | 89.9% |
| August | £54,673 | £46,142 | 84.4% |
| September 1 to 8 settled | £16,329 | £13,063 | 80.0% |
The rate has fallen every month since May. That is a retention question rather than a forecasting one, but it has to be forecast honestly either way.
| September renewals | Amount | Basis |
|---|---|---|
| Billed 1 to 8 September | £13,063 | Recharge, settled measured |
| Still scheduled 9 to 30 | £40,382 | Recharge queue measured |
| At September's 80.0% rate | £45,368 | assumption our base case, because the trend is downward |
| At August's 84.4% | £47,145 | derived if September stops deteriorating |
| At the April to August pool 89.2% | £49,083 | derived if it recovers to normal |
| Currently in the sheet | £64,427 | Schedule at about 95%, the best month on record |
Also outside the ads. Driven by the email and launch calendar, so it is forecast from what those days actually produce rather than from a monthly average.
| Input | Value | Where it comes from |
|---|---|---|
| Measured 1 to 08 September | £16,251 | Shopify, returning customers excluding renewals measured |
| A day with an email | £2,255 | 4 such days so far this month measured |
| A day without | £1,808 | 4 such days so far this month measured |
| Launch-day multiple | 4.26x | 18 real product launches back to 2024, 95,406 orders measured New-Pott drops specifically: Autumn Scents, New Season, Raku, Teal |
| 12 September, Rust & Amber | £8,655 | derived baseline £2,031 x 4.26 |
| Remaining calendar | 9 + 13 | Email days and quiet days left, from the sheet measured |
| September returning | £66,449 | derived |
The launch lift is a one-day event. Across those four new-Pott drops the day after a launch came back to 1.05x of baseline and the day after that to 0.96x, so no multi-day tail is assumed.
| Value | Note | |
|---|---|---|
| New-customer AOV | £67.83 | Measured this month, 276 customers measured |
| New customers needed | 2,187 | derived £148,337 / £67.83 |
| Banked 1 to 08 September | £18,720 | 276 customers measured |
| Still to find | £129,617 | derived over 22 days = £5,892 a day |
| Current run rate | £2,340 a day | measured so the ads must run at 2.52x the pace of the first 8 days |
| Running now | Needed for the goal | Gap | |
|---|---|---|---|
| Blended ROAS | 4.04x | 3.39x | above the line |
| New-customer ROAS | 1.60x | 1.93x | +0.33 |
| Cost per new customer | £42.29 | £35 | -17.0% |
Blended ROAS reads comfortably today and that is misleading. We are spending £1,459 a day against a base of subscriptions and repeat customers that arrives regardless, so the base is doing most of the carrying. As spend rises toward £2,957 a day, blended falls toward the 3.39x line by arithmetic alone, whether or not anything changes. The only thing that holds it up is new-customer ROAS.
Advertising gets less efficient as you spend more, because the cheapest customers are reached first. We have measured that on this account rather than assuming it away: new-customer revenue per day = 5.694 x (spend per day) ^ 0.824. At the breakeven budget that curve predicts:
| Curve predicts | Goal needs | Gap | |
|---|---|---|---|
| New-customer revenue | £109,669 | £148,337 | £38,668 |
| New-customer ROAS | 1.43x | 1.93x | +0.50 |
| Cost per new customer | £47 | £35 | -26.1% |
| New customers | 1,617 | 2,187 | 570 short |
That is the whole job: a 26.1% improvement in acquisition cost while roughly doubling daily spend. There is one precedent on this account. The Isle of Wight tomato launch in June moved cost per new customer from £54.14 to £41.06, a 24% improvement, at flat spend. Rust & Amber is the equivalent event, which is why the budget steps on the 12th.
| Ad spend | Revenue lands | % of goal | Blended | Profit |
|---|---|---|---|---|
| £43,773 | £180,886 | 70% | 4.13x | £771 |
| £60,000 | £201,375 | 77% | 3.36x | £-7,138 |
| £76,727 the breakeven budget | £221,486 | 85% | 2.89x | £-15,699 |
| £90,000 | £236,894 | 91% | 2.63x | £-22,717 |
On current efficiency the month breaks even at about £45,500 of spend and £183,123 of revenue, which is 70% of the goal. The goal and breakeven only happen together if cost per new customer comes in at £35.
Everything above is either measured or arithmetic, except these. If one of them is wrong, this is what it does to the answer.
Subscription realisation: 80.0% this month
We use September's own rate because the trend has been downward for four months. If it recovers to the April to August average of 89.2%, renewals land £3,715 higher.
If it is wrong: New-customer ROAS needed falls from 1.93x to 1.88x. If the sheet's £64,427 turns out right, it falls to 1.68x, which closes a third of the gap.
Returning-customer forecast: £66,449
Built from 8 days of this month split into email and non-email days, then applied to the remaining calendar. Eight days is a short base.
If it is wrong: 10% either way moves the requirement to 1.85x or 2.02x.
The launch multiple: 4.26x
Median of four new-Pott launches. Rust & Amber is a new pot and a new scent together, which is arguably bigger than any of them, but the sample is four.
If it is wrong: If the 12th performs like an ordinary day, the requirement rises to 2.02x and £6,623 has to be found by the ads instead.
New-customer AOV: £67.83
Measured on 276 new customers this month. It decides how many customers the revenue target converts into.
If it is wrong: A £5 move changes the customer count by about 174, and the CAC target with it.
The response curve: exponent 0.824
Measured on this account from daily spend and new-customer revenue. It sets how much efficiency we lose as budgets rise, and therefore the size of the gap in Step 6.
If it is wrong: It does not change what the goal needs, only how hard the goal is. A flatter curve makes the 26.1% ask smaller.
| Source | Used for | Window |
|---|---|---|
| Shopify (order level) | Revenue, orders, shipping income, new vs returning split, new-customer AOV | 1 to 08 September, plus 95,406 orders back to 2020 for the launch study |
| Recharge | Subscription schedule and what actually billed, by day and status | April to September 2026 |
| Meta and Google Ads | Daily ad spend | 1 to 08 September |
| Klaviyo | Campaign dates, used to identify launch days | 618 campaign days back to March 2023 |
| Financial dashboard | Margin, VAT, fulfilment, payment fees, overhead | Live configuration |
Actuals are complete to 2026-09-08. Built 2026-09-09. Rebuilt from source on every refresh, so nothing here is typed by hand.