Fifty-two orders.The best month of the year.
Unseen Footwear
Against the same month last year the business is unrecognisable. Against last month it is quietly better run: more money taken from slightly fewer visitors, and the highest order count of 2026.
On precision. Figures are as Shopify, Klaviyo and Meta report them, to the penny and the single unit. Conversion rate is Shopify’s session-based measure rather than orders divided by sessions, which would read 1.20% because six of the month’s 52 orders came through channels with no web session behind them. Klaviyo attributes £2,870.13 of the month to email and Meta attributes roughly £1,262 to ads; both count some of the same orders, so the two are shown separately throughout and never added together.
Half the month’s traffic came with no referrer at all, which is what a brand with a strong owned audience looks like. A fifth came from social.
| Traffic source | Sessions | Share | Named referrers within it |
|---|---|---|---|
| Direct | 2,172 | 50.03% | — |
| Social | 951 | 21.91% | Instagram 598 · Facebook 349 · Reddit 4 |
| Unknown | 725 | 16.70% | Tradedoubler 688 · Frontpage 5 |
| Search | 474 | 10.92% | Google 451 · Bing 16 · ChatGPT 6 · Yahoo 4 |
| 19 | 0.44% | Gmail 19 | |
| Total | 4,341 | 100% |
Tradedoubler, the affiliate network, sent more sessions than Google and sits outside the Essential scope. Shopify records only 19 email sessions because it reads the last referrer only; Klaviyo credits far more revenue to email using its own click and open window. Section 05 sets out both.
| Step | August 2026 | July 2026 | August 2025 | Year on year |
|---|---|---|---|---|
| Sessions | 4,341 | 4,533 | 1,754 | +147.5% |
| Reached checkout | 123 | 123 | 35 | +251.4% |
| Completed checkout | 33 | 30 | 8 | +312.5% |
| Sessions reaching checkout | 2.83% | 2.71% | 2.00% | +41.8% |
| Checkout completion | 26.83% | 24.39% | 22.86% | +17.4% |
| Pages per session | 2.08 | 1.90 | — | — |
| Bounce rate | 64.99% | 65.10% | — | — |
Every step of the journey improved year on year, and two of the three improved on last month. But 90 checkouts were started and abandoned in August — roughly £9,800 of baskets that reached the payment page and stopped. The abandoned checkout flow caught 73 of them and recovered £196.50. The mechanic works; the recovery rate has never been tuned.
Growth did not come from more people. Sessions fell 4.2% while conversion rose 14.9%. At August’s rate the shop needs 83 sessions to produce one order, so anything that adds qualified visitors compounds and anything that adds unqualified ones does not.
| Landing page | Sessions | Share of traffic |
|---|---|---|
| Homepage | 1,901 | 43.79% |
| /collections/helier | 373 | 8.59% |
| /collections/sale-1 | 236 | 5.44% |
| Helier Black/White product page | 93 | 2.14% |
| /en-je — Jersey storefront | 84 | 1.94% |
| /collections/innovate-series | 80 | 1.84% |
| /collections/men | 74 | 1.70% |
| Helier Grey/White product page | 64 | 1.47% |
Twenty-seven products sold. Helier took 36.22% of product revenue across five colourways and twenty-five pairs; add Marais and two silhouettes account for 59.40% of everything sold.
| Silhouette | Pairs | Revenue | Share | Realised per pair |
|---|---|---|---|---|
| Helier | 25 | £1,985.43 | 36.22% | £79.42 |
| Marais | 15 | £1,270.42 | 23.18% | £84.69 |
| Halkett | 9 | £643.64 | 11.74% | £71.52 |
| Regent | 7 | £642.68 | 11.73% | £91.81 |
| Noirmont | 4 | £316.00 | 5.77% | £79.00 |
| Portelet | 4 | £270.98 | 4.94% | £67.75 |
| Runa, Train, Rozel, Clarn | 5 | £352.02 | 6.42% | £70.40 |
| 27 lines sold | 69 | £5,481.17 | 100% | £79.44 |
Product-level sales total £5,481.17; the remaining £39.99 of the month’s £5,521.16 is non-product revenue. Fourteen of the twenty-seven lines sold exactly one pair, accounting for £1,047.89 or 19.12% of product sales.
Sales are spread thinly and widely. Six areas account for two thirds of revenue; the rest is a long tail of single-order towns.
| Market | Orders | Pairs | Revenue | Share of sales | Average order |
|---|---|---|---|---|---|
| United Kingdom | 36 | 48 | £4,122.13 | 74.66% | £114.50 |
| England | 29 | — | £3,345.62 | 60.60% | £115.37 |
| Scotland | 6 | — | £641.51 | 11.62% | £106.92 |
| Northern Ireland | 1 | — | £135.00 | 2.44% | £135.00 |
| Jersey | 16 | 22 | £1,533.03 | 27.77% | £95.81 |
| United States | 0 | −1 | −£134.00 | −2.43% | refund only |
| Total | 52 | 69 | £5,521.16 | 100% | £108.70 |
| Area | Orders | Revenue | Average order | Share of sales |
|---|---|---|---|---|
| JerseyAll parishes combined | 16 | £1,533.03 | £95.81 | 27.77% |
| SurreyFrensham and Farnham, three miles apart | 3 | £962.87 | £320.96 | 17.44% |
| Glasgow | 4 | £483.51 | £120.88 | 8.76% |
| Leicester | 4 | £387.50 | £96.88 | 7.02% |
| Bromley | 1 | £175.00 | £175.00 | 3.17% |
| Billingham | 1 | £166.50 | £166.50 | 3.02% |
| Six areas | 29 | £3,708.41 | £127.88 | 67.17% |
The remaining 23 orders and £1,812.75 are spread across single-order locations. Areas group Shopify billing cities that sit in the same county or island; a full county roll-up across every order follows next issue.
| Channel | Orders | Revenue | Share |
|---|---|---|---|
| Online Store | 45 | £5,084.17 | 92.08% |
| Flxpoint Enterprise Dropship | 2 | £250.00 | 4.53% |
| LTS | 3 | £241.99 | 4.38% |
| Shopify Mobile | 1 | £0.00 | 0.00% |
| Shop app | 1 | −£55.00 | −1.00% |
Klaviyo credits itself with £2,870.13 in August, 46.88% of the revenue it observes. Campaigns produced 93.15% of that; automated flows produced 6.85%.
| Campaign | Sent | Open rate | Click rate | Revenue | Verdict |
|---|---|---|---|---|---|
| Last Summer SALE | 14 Aug · 17:00 | 23.17% | 0.80% | £1,042.55 | Hero |
| Summer Sale — 45% | 12 Aug · 17:00 | 23.34% | 1.01% | £673.61 | Strong |
| Bank Holiday Archive Sale | 26 Aug · 17:00 | 22.70% | 0.77% | £279.50 | Keep |
| Bank Holiday Pay DayFull price. Briefed and sent the same day | 28 Aug · 18:30 | 22.59% | 0.50% | £270.00 | Scale |
| Sale — The HELIER Edit | 7 Aug · 18:30 | 22.78% | 1.08% | £240.74 | Keep |
| Extended Sale | 16 Aug · 10:00 | 23.51% | 0.93% | £167.23 | Flat |
| Train Focus MailerFirst full-price send after the sale | 21 Aug · 18:30 | 23.18% | 0.54% | £0.00 | No sales |
| Sale — 30% + 15%Highest click rate of the month | 6 Aug · 18:00 | 22.76% | 1.23% | £0.00 | No sales |
| 8 campaigns | 22.59–23.51% | 0.50–1.23% | £2,673.63 |
| Flow | Trigger | Deliveries | Revenue | Per recipient | Verdict |
|---|---|---|---|---|---|
| Abandoned Checkout | Checkout started | 73 | £196.50 | £2.69 | Working |
| Unseen Club welcome | Added from Meta lead form | 209 | £0.00 | £0.00 | Rebuild |
| Welcome Series | Joined Unseen Club | 147 | £0.00 | £0.00 | Rebuild |
| Post Purchase | Placed order | 45 | £0.00 | £0.00 | Rebuild |
| Browse Abandonment | Viewed product | 30 | £0.00 | £0.00 | Rebuild |
| 5 flows | 504 | £196.50 | £0.39 |
Clicking does not predict buying. Across the eight sends the correlation between click rate and revenue is effectively zero (r = −0.02). The send with the highest click rate of the month, 1.23%, earned nothing. The best earner clicked at 0.80%. Optimising subject lines and buttons is optimising the wrong thing on this audience.
Open rate is not a usable signal. Eight sends landed inside a 0.92-point band. A spread that tight is a measurement artefact rather than performance, almost certainly Apple’s Mail Privacy Protection pre-opening messages. We will keep reporting it for continuity; it should not inform a decision.
Email here is retention, not acquisition. Klaviyo credits 46.88% of the revenue it sees. Shopify puts the returning customer rate at 45.65%. Those two matching is not a coincidence — email is largely talking to people who have already bought. Meanwhile 386 flow deliveries aimed at new people earned £0.00. The channel is doing half the work with only half its job running.
Two campaigns delivered in August out of 140 in the account; the remaining 138 are dormant, inherited from the previous agency.
| Campaign | Live since | Objective | Spend | Impressions | Reach | Freq | CPM | Result | ROAS |
|---|---|---|---|---|---|---|---|---|---|
| Prospecting — CatalogueAdvantage+ Shopping | May 2026 | Find new customers | £584.34 | 42,961 | ~14,225 | 3.02 | £13.60 | 7 purchases | 2.02× |
| Unseen Club — Lead form | June 2026 | Grow the Unseen Club | £254.79 | 22,037 | ~11,478 | 1.92 | £11.56 | 134 leads | 0.32× |
| 2 campaigns | £839.13 | 64,998 | £12.91 | 8 purchases | 1.50× |
Meta Ads Manager, 1–31 August 2026, 7-day click attribution. Reach is calculated from impressions divided by frequency and is the one figure in this report that is derived rather than read — it will be read directly next issue. Prospecting frequency of 3.02 against a reach of roughly 14,200 is repetition rather than reach. No remarketing campaign delivered at all in August.
Paid social cost £104.89 per purchase against an average order of £108.70. Before product cost that is £3.81 of margin; after it, the first order loses money. That is normal for prospecting, but it means paid social only works here if the second order arrives — and the second order is supposed to come from the flows that returned nothing.
The lead campaign looks efficient on its own metric and is not. £1.90 a lead is a good cost per lead by any standard, but cost per lead is only worth measuring if the leads go on to buy, and across the campaign’s life they have not. Cheap acquisition of an audience that never converts is more expensive than no acquisition at all, because the cost carries on into the email channel.
Since it launched in June the Unseen Club lead form has spent more than £900 and delivered over 450 leads. In that time it has produced one sale.
That is not a flow problem, or not only a flow problem. Four hundred and fifty people joined the list over three months and one of them bought something. A lead-form ad optimised for the cheapest possible signup collects people who want to be on a list, not people who want the product — so the list has been growing with data that does not convert, and every one of those profiles then sits in the send file dragging engagement down.
The campaign comes off in September. The work moves to the people already in there: rebuilding the welcome sequence so the existing members get a proper introduction, and segmenting the list so the unengaged stop suppressing everyone else’s open rate.
Everything completed on the Unseen account between 1 and 31 August, taken from the project board with recorded time attached.
| Discipline | Tasks | Recorded time | Share of time | Average per task |
|---|---|---|---|---|
| Strategy & proposals | 3 | 5h 00m | 31.25% | 1h 40m |
| Email & Klaviyo | 8 | 4h 20m | 27.08% | 33m |
| Paid social | 2 | 2h 15m | 14.06% | 1h 08m |
| Reporting & analysis | 10 | 2h 04m | 12.92% | 12m |
| Retail & partnerships | 2 | 1h 26m | 8.96% | 43m |
| Website & merchandising | 2 | 40m | 4.17% | 20m |
| Account & operations | 1 | 15m | 1.56% | 15m |
| Content & organic | 1 | untimed | — | — |
| Total | 29 | 16h 00m | 100% | 40m |
Time shown is recorded actual time from the project board. Five of the twenty-nine tasks carry none, four of which are reporting work, so that line is understated. Work that never became a task — calls, messages and ad-hoc requests — sits outside the figure entirely.
Three pieces of work in August were about where the brand goes next rather than what it sold this month. The Unseen Club proposal set out how the membership could become a revenue mechanic rather than a mailing list. The TikTok and Pinterest proposal mapped two channels the brand does not currently use. And the Flannels Liverpool work laid the groundwork for the retail push that went live in September.
None of these appear in the trading figures for August. All three are the reason the September and October figures could look different.
The 28 August send took £270.00 and was the only full-price campaign of the month to earn anything. It was briefed at 07:36 and in inboxes by 18:30. This is what that involves, and it is the same shape for every send.
An occasion, a commercial reason and a shortlist of product. In this case: payday weekend, bank holiday, and a push on the four silhouettes that saw least attention during the sale.
Who receives it, and who is deliberately excluded. Recent purchasers, unengaged profiles and anyone already in an active flow are handled separately.
Subject line, preview text and body written to the occasion rather than to a discount, since there was no discount to lead with.
Campaign photography selected and cropped to template, with product tiles pulled through from the catalogue.
Assembled in Klaviyo against the house template, with links tagged so the revenue can be attributed back to the send.
Rendering checked across clients, links verified, and the send previewed on mobile, which is where most of this list reads.
Timed to the audience, then read at 24 hours and again at seven days once the attribution window has closed.
Eight sends were built to this pattern in August, taking 4h 20m of recorded time between them — an average of 33 minutes each, including the two that were rebuilt after the sale was extended.
Traffic has more than doubled over the year, so absolute revenue is a poor way to compare months. Revenue per 1,000 sessions is the fairer measure, and on it August is the strongest direct-to-consumer month in the record.
| January to August | 2026 | 2025 | Change |
|---|---|---|---|
| Total sales | £55,447.03 | pending | — |
| Orders | 295 | pending | — |
| Pairs sold | 1,001 | pending | — |
| Sessions | 36,547 | pending | — |
| Average order | £187.96 | pending | — |
| Revenue per 1,000 sessions | £1,517.14 | pending | — |
| of which bulk-weighted months | £37,496.47 | pending | — |
| Direct-to-consumer year to date | £17,950.56 | pending | — |
A fixed block from this issue onward, so the year accumulates in one place. The 2026 column is complete; the 2025 comparison needs one further Shopify export covering January to August 2025 and lands next issue. Bulk-weighted months are January, February and April, whose average order values of £466.38, £326.45 and £197.12 mark them as wholesale rather than direct-to-consumer; the final row strips them out so the two years compare like with like.
| Month | Total sales | Orders | Sessions | Conversion | Revenue per 1,000 sessions |
|---|---|---|---|---|---|
| September 2025 | £1,606.60 | 11 | 2,286 | 0.219% | £702.80 |
| October 2025 | £6,208.68 | 55 | 6,467 | 0.742% | £960.14 |
| November 2025 | £2,762.94 | 33 | 8,809 | 0.250% | £313.65 |
| December 2025 | £4,075.90 | 38 | 7,383 | 0.352% | £552.07 |
| January 2026bulk weighted, average order £466.38 | £22,893.10 | 49 | 7,331 | 0.286% | £3,122.75 |
| February 2026bulk weighted, average order £326.45 | £8,042.59 | 26 | 5,899 | 0.254% | £1,363.38 |
| March 2026 | £2,746.71 | 22 | 3,597 | 0.500% | £763.61 |
| April 2026bulk weighted, average order £197.12 | £6,560.78 | 42 | 3,948 | 0.608% | £1,661.80 |
| May 2026 | £3,285.85 | 28 | 3,406 | 0.470% | £964.72 |
| June 2026 | £2,019.18 | 28 | 3,492 | 0.344% | £578.23 |
| July 2026 | £4,377.66 | 48 | 4,533 | 0.662% | £965.71 |
| August 2026 | £5,521.16 | 52 | 4,341 | 0.760% | £1,271.86 |
Twelve months to August 2026, Shopify. January, February and April carry bulk orders at average order values of £466.38, £326.45 and £197.12 and are not comparable on a direct-to-consumer basis. Excluding those three, August is the highest revenue per 1,000 sessions in the record and the highest conversion rate.
The brand sells outside a sale window. October 2025 took £6,208.68 from 55 orders — more orders than any month since, August included — on a recorded discount rate of 5.6%. Whatever October had, it is repeatable, and it is worth establishing what it was before the autumn drop.
Traffic alone does not do it. November 2025 was the highest-traffic month in the record at 8,809 sessions and produced £2,762.94 from 33 orders, a conversion rate of 0.250%. More visitors without a reason to buy produced a third of August’s revenue per thousand.
Four pieces of work for September, in the order we would take them. Each one is chosen because the August data points at it directly.
Segmentation first, and the lead-form signups are where it starts. Over 450 profiles have joined through the Unseen Club ads and produced one sale between them; sending to all of them every time is what holds open rate in a flat 0.92-point band and puts deliverability at risk over time. We will build engagement tiers, suppress the persistently unopened, and give the active segment a reason to stay active.
Four flows returned £0.00 on 431 deliveries. We would not attempt all four. The Meta lead welcome comes first — 209 deliveries, and paid social is spending £1.90 a head to fill it. The main welcome series second, 147 deliveries to the same kind of new subscriber. Browse abandonment third: only 30 deliveries but the highest intent and the smallest build. Post purchase can wait, because returning customers are already 45.65% of orders.
123 sessions reached checkout in August and 33 completed, so 90 baskets stopped at the payment page. Before changing anything we want the team to review the current on-site journey end to end and come back with a prioritised roadmap rather than a list of tweaks — what to test, in what order, and what each test would need to prove.
Working with David and James on how we support the storytelling around the autumn range. August’s best-performing send and best-performing ad were both single-product pieces on Helier, and the site’s second most-entered page was the Helier collection. There is an appetite for one story told properly rather than several told lightly, and the plan should be built around that.
Two of them cost build time and no media. The other two are planning work that shapes what October and November look like rather than September. Between them they address the three things August exposed: an email channel doing half its job, a checkout losing nine baskets in ten, and a product story that works when it is told and is not being told often enough.