The engine is
already built.
We found where your next profitable, repeat orders are waiting.
Baseline: 13 months to 30 April 2026
Three rules run through the whole deck.
Your brand name already drives real sales.
This is not a repair job. The engine works. You are the most-admired brand in South Africa, two years running, ahead of Nike and Adidas at home. This deck shows the next steps that turn that strength into a bigger, more profitable business: the wins first, then the model.
What the data confirms.
People come looking for you by name.
Your team's testing has already moved the numbers.
A customer list few shops your size can match.
| Group | Customers | % of list | Orders each | Lifetime spend |
|---|---|---|---|---|
| Champions | 61,386 | ~10% | 3.5 | R3,494 |
| Loyal | 85,232 | ~14% | 3.1 | R3,182 |
627,982 customers, 1,065,766 online orders all-time, R984.8M spent all-time. The top 23% of customers drive 45% of all orders and 49% of all spend.
You win online sales without cutting prices.
Your main range sells online at full price, with almost no discount codes used (six orders in 13 months, about 0.0035%). The whole plan protects margin, because the brand is wanted, not pushed. We build on that, we do not change it.
The paid-social base is built right. This next part is about scaling it.
The layers to switch on next.
The headline ad return looks great, but it hides the real cost of new buyers.
Ads report a 27 times return. But most of that just collects sales you would get anyway: ads on your own name return 117 times. About 74.8% of search spend goes on people typing your name. Only about R24,000 a year goes to finding people who do not know you yet. Split spend on your name from spend on new people, and you can see the real cost of a new buyer for the first time.
About R137,668 can move to channels that work, with no new budget.
Spent on local ads at today's returns, that is worth roughly R0.9M to R1.5M a year. Also, about 69% of the actions the account counts are not sales. Removing them helps the ads aim at real sales, at no cost.
The same story on Meta, sharper
Facebook and Instagram run near parity with search on their active months, about R84,000 each (paid social began in September 2025, and is recently nearer R92,000 a month). The signal is strong and the click rate is high. But the ad return tells the same story as search, only sharper. The account reports a 45 to 53 times return. That number counts a sale to Meta whenever someone just saw an ad in the day before buying, and your brand is so strong that many of them would have bought anyway. Counted on clicks only, the true return is closer to 6 to 11 times. Switching the report to a click-only basis does not change the business. It gives you a new-buyer number your team can act on. One thing to fix this week: the account is paused against a small unpaid balance, about R6,500, so paid social is dark right now. Settle it and delivery resumes. It is a billing item, not a build fault.
Three moves on Meta, all from settings and data you already own.
The one channel your audience lives on that you have not entered.
No spend, no tracking, no audience there yet. Your core customer is South African, Gen Z, 16 to 28, and 88% on a phone. The Scorpion Kings collaboration reached 105,000 people at a 24.45% open rate, so the right cultural moment lands with this crowd. And because you convert so well on search, every TikTok view that makes someone search your name feeds the channels that already sell. Brand first: you approve every post, and we grow it in steps that match results, not all at once.
Checkout sales dropped in the last 30 days. Fixing this protects what you have.
Most of the loss is at two checkout steps. One of them is about payments.
| Checkout step | Finish | Should be |
|---|---|---|
| Start checkout to entering payment | ~47% | 60% to 75% |
| Entering payment to buying | ~42% | 70% to 85% |
The payment step reads as a card or gateway issue (bank declines, 3DS), a fix with your payment provider, not a checkout-design issue. Above that, raising the baseline rate toward benchmark is genuine growth, run alongside your team's own tests.
Your repeat-buyer engine is built. The only thing left is switching it on.
The 30/60/90-day cohort read asked for at the scoping session, located.
| Window | Buy again | What it shows |
|---|---|---|
| Month 1 | 5.72% | the leak starts here |
| Month 3 | 3.79% | still slipping |
| Month 6 | 2.89% | levels off, low |
The biggest leak is the first 30 days: most buyers who lapse, lapse early. A post-purchase flow inside the first month is built to catch them, which is why month one is the single biggest retention win.
Turning on the built email flows, with the numbers.
| Lever | Today | Benchmark | What it could add |
|---|---|---|---|
| Buy again in month one | 5.72% | 12% to 18% | the biggest win |
| Email flows | R4.4M (1 live) | 20% to 40% of sales | R6.6M to R8.1M / year (ceiling) |
| Win back old buyers | 174,502 asleep, none contacted | wake 3% to 5% | 7,900 to 13,100 orders |
These are industry benchmarks, not promises. The R6.6M to R8.1M is a ceiling extrapolated from a single live flow, not a forecast. December 2025 brought the most new buyers (37,200) but the fewest came back (4.89%); a post-purchase flow is built to fix exactly that.
Own the tracking that records your sales.
Your setup is better than most shops your size. The tracking is smart, the data is being saved, and the tracking pixel is set up the right way, with consent. But two tracking boxes load on every page. You can see one. You cannot see the other, and that is the one feeding your ad results. The fix: your own tracking container, on GALXBOY credentials, with your e-commerce team holding admin from day one. We design and configure it; you own and control it. This is simply owning your own data, which a brand your size should.
The path to
1,000,000.
1,000,000 orders is a 3 to 4 year goal, not a quick promise.
You do about 428,500 orders a year now. One million is about 2.33 times that.
The safe near-term goal is about 472,500 orders in Year 1 and about 549,500 in Year 2. We would rather promise 472,500 we can hit than 1,000,000 that needs everything to go right. One million is where these steps lead once they grow and are funded.
Every number checked against Shopify, your source of truth.
| Channel | Orders / year | Sales / year | Order value | Order share |
|---|---|---|---|---|
| Online (Shopify) | 159,100 | R143.9M | R904 | 37.1% |
| In-store, 14 stores | 266,200 | R225.0M | R845 | 62.1% |
| Other | 3,200 | R3.2M | n/a | 0.7% |
| All channels | 428,500 | R372.1M | R868 | 100% |
13 months of real data, turned into a yearly figure. Shopify is the one source of truth for orders and sales, and every other tool is checked against it. Each store: about 19,000 orders and R16.1M a year (Kimberley opens in 2026).
We split online orders into new and repeat, because each grows in a different way.
Every extra order sits in one bucket only.
| Bucket | What it counts | Why it cannot overlap |
|---|---|---|
| A. New buyers | First orders from new visitors | Counts new buyers only, not repeat |
| B. Repeat buying | Extra repeat orders above today's 71,000 | Never counts the base again, never in A |
| C. Win-back | Old buyers brought back | They buy nothing today, so every order is extra |
| D. Checkout | More checkouts finished by today's visitors | Happens at the end, A brings visitors at the start |
| E. In-store | Shop orders, minus sales that just moved online | Separate from all online |
The tricky pair: B is someone buying again later. D is someone finishing one visit. The same order is never counted in both.
One bucket per step. No step counts another's orders.
The safe buckets reach about 751,500. We are honest about the rest.
| Lever at its best case | Extra orders | How sure |
|---|---|---|
| New traffic: visits to about 25M, TikTok at full scale | +130,000 to +160,000 | low, needs spend |
| Stores: grow to 26 to 28 shops | +60,000 to +80,000 | medium |
| Repeat buying up toward 15%, win-back running well | +30,000 to +40,000 | medium |
By Year 3 to 4 the safe buckets reach 751,500 (1.75x). The last 248,500 orders to reach 1,000,000 only come if all three of these hit their best case at once. We say so plainly.
Every online order comes from exactly one place.
| Scenario | Online | In-store | Total |
|---|---|---|---|
| Baseline | 159,100 | 266,200 | 428,500 |
| Year 1 | 191,100 | 278,200 | 472,500 |
| Year 2 | 240,100 | 306,200 | 549,500 |
| Year 3 to 4 | 362,100 | 386,200 | 751,500 |
Every online order is either a starting order or one extra from A, B, C or D. None counted twice. Totals include about 3,200 orders from other channels.
Sales are half the story. After ad spend and the cost to serve an online order, here is the contribution, honestly.
| Year | Extra online sales | Ad spend | Gross profit after ads (53% PROXY) | True contribution after cost-to-serve (PROXY) |
|---|---|---|---|---|
| Year 1 | ~R29M | ~R2.3M | ~R13.0M | ~R8.9M |
| Year 2 | ~R73M | ~R3.2M | ~R35.5M | ~R25.0M |
| Year 3 to 4 | ~R183M | ~R17M | ~R80M | ~R53.6M |
Every figure is a labelled PROXY until your cost data lands. Gross profit after ads is before the cost to serve an online order (shipping, payment, returns, about R130 each); taking that out gives the true contribution, the right-hand column, about R8.9M (PROXY) in Year 1 and still positive. Final profit needs your cost of goods. TikTok budget sits inside ad spend.
We face the order-value question head on. Growth keeps discounts flat or lower.
A careful plan needs a real worst case, not a slightly smaller best case.
| Lever | Base Year 2 | Worst case Year 2 |
|---|---|---|
| A new buyers | +28,000 | +18,000 |
| B repeat | +18,000 | +13,000 |
| C win-back | +20,000 | +12,000 |
| D checkout | +15,000 | +12,000 |
| E in-store | +40,000 | +35,000 |
| Total | 549,500 (x1.28) | 518,500 (x1.21) |
Worst case, with both orders and order value down: Year 2 sales about R433M against R479M flat, and gross profit after ads about R21M at the 53% proxy (R18M to R23M across the band), still positive. The shakiest lever is new traffic, and the near-term plan leans on it the least.
We separate what the plan protects from what it adds.
Four numbers turn this into a real profit model.
We are asking for only four numbers, not your full costs. With them, the proxy falls away and the model shows real contribution. The next step is to add them and turn every sales line into a contribution line you can steer by.
| Cost of goods, by category | Replaces the 53% PROXY |
| Card and payment fees | Part of the real cost per order |
| Packing and delivery cost per online order | A cost that grows with orders |
| Online returns rate | Real sales after returns, and the true cost of one more order |