# Ochre & Silk: model workings

How every dollar figure in the Ochre & Silk Search & AI strategy report (14 October 2026) was
built, step by step, so you or your own AI assistant can recompute it. The market, the clicks and
the share are **Australia only and non-branded**. The value per click comes from Shopify's
Google-organic figures over twelve months, split into branded and non-branded. Every dollar is
**gross** (sales as your Shopify attribution report records them, before returns; homewares returns
are not measured yet), in AUD.

## 1. The inputs, and where each one comes from

| Input | Value | Source |
|---|---|---|
| Measured keyword core | 93,515 searches / month | Google Keyword Planner, SEMrush and Ahrefs for Australia, median of three, October 2026. Core = terms you stock with direct buying intent; plural and word-order variants counted once. The full measured basket is 252,000 / month |
| Long-tail factor | x 1.35 | Standard Team Empathy uplift from the measured core to the long tail around it |
| Winnable click share | 20% of searches | The share of searches that end in an organic click once ads, shopping carousels and AI answers take theirs |
| Non-branded clicks, all countries | 11,319 a year | Search Console, sc-domain:ochreandsilk.com.au, 13 Oct 2025 to 12 Oct 2026: named non-branded queries plus the queries Google withholds (long tail, counted as non-branded) |
| Australian share of non-branded clicks | 77.8% | Same source, named non-branded queries split by country (applied to the withheld queries too) |
| Non-branded Australian clicks today | 8,805 a year | 11,319 x 77.8% |
| Branded clicks (excluded) | 1,347 a year | Queries containing Ochre & Silk, its misspellings, or Imprint House |
| Shopify Google organic | 14,706 sessions, 166 orders, A$108,944.85 | Shopify Analytics > Attribution > Channels, last 12 months, last non-direct click |
| Non-branded conversion rate | 0.9020% (shown as 0.902% in the report) | 118.6 non-branded orders / 13,149 non-branded sessions (split explained in section 5) |
| Non-branded order value | A$638.26 | A$75,697.39 / 118.6 orders |
| Value per non-branded click | A$5.7569 (shown as A$5.76) | Conversion rate x order value |
| First-order share | 67.5% | Same Shopify row: 112 of 166 orders came from new customers |

## 2. The market chain

1. 93,515 x 1.35 x 12 = **1,514,943 searches / year**
2. x 0.2 = **302,989 winnable organic clicks / year** (25,249 a month)
3. 8,805 / 302,989 = **2.91% share today** (shown as 2.9%)
4. 1% of the pool = **3,030 clicks / year** = 27.3 orders = **A$17,443 gross / year**

For context, the full Australian category, sized from Ahrefs' keyword universe across 18 product
seeds, scaled to the triangulated basis and with the long tail, is about 554,587
searches a month. The model uses only the measured core.

## 3. The scenarios

Each case grows the share of the winnable pool; the value per click is held at today's measured figure.
Revenue = share x 302,989 x A$5.7569.

| Case | Share | Clicks / yr | Orders / yr | New customers / yr | Non-branded revenue / yr (gross) | New revenue / yr (gross) |
|---|---|---|---|---|---|---|
| Today | 2.9% | 8,805 | 79 | 54 | A$50,690 | - |
| Conservative | 6.0% | 18,179 | 164 | 111 | A$104,656 | +A$53,966 |
| Mid | 12.0% | 36,359 | 328 | 221 | A$209,313 | +A$158,623 |
| High | 20.0% | 60,598 | 547 | 369 | A$348,855 | +A$298,165 |

Conservative, mid and high are 6%, 12% and 20% of the pool. "New revenue" is the rounded case minus
the rounded today figure. The today row uses the measured clicks: 8,805 x A$5.7569 = A$50,690.

## 4. The value ladder (new customers only)

The scenario table counts every non-branded order, first and repeat. The ladder counts only the new
customers each case brings (orders x 67.5%), so repeat revenue is never counted twice.

- **First-order value: A$638.26, measured.** The non-branded order value. Assumes a first order is the same size as the average order (Shopify gives no first-order split on this report).
- **First-year value: A$797.82, modelled.** First order plus a quarter of a repeat order in year one.
- **Lifetime value, 5 years: A$957.39, modelled.** First order plus half a repeat order over five years.
- Why those assumptions are conservative: Google organic took 54 returning orders for every 112 new ones in the last twelve months (0.48 per new order, in a single year). Both modelled rungs sit well under that.
- **Re-run trigger:** Shopify Analytics > Customer cohort analysis (cumulative amount spent per customer). The two modelled rungs are replaced with the measured cohort curve.

| Case | New customers / yr | First-order revenue / yr | First-year value / yr (modelled) | Lifetime value, 5 yrs (modelled) |
|---|---|---|---|---|
| Today | 54 | A$34,200 | A$42,750 | A$51,300 |
| Conservative | 111 | A$70,612 | A$88,264 | A$105,917 |
| Mid | 221 | A$141,223 | A$176,529 | A$211,835 |
| High | 369 | A$235,372 | A$294,215 | A$353,058 |

## 5. How the branded / non-branded split was made

Shopify does not label orders branded or non-branded, so the A$108,944.85 was split with two measured inputs:

1. Branded share of Google organic visits = Search Console's branded click share (10.6%, incl. Imprint House).
2. Branded visits convert 3.37x better than non-branded and order 1.1x larger, read from GA4 (Google-organic visits landing on the homepage, where branded searches land, against every other landing page, twelve months).
3. Solved so orders and revenue add back to Shopify exactly: branded 47.4 orders / A$33,247 at 3.04%; non-branded 118.6 orders / A$75,697 at 0.9%.

Cross-check: GA4's direct landing-page revenue split puts non-branded at 59.3%, or A$64,575 on Shopify's total. The report quotes the band A$64,575 to A$75,697. Pricing all 11,319 non-branded clicks at A$5.76 gives A$65,162, inside the band at the low end: Shopify records more Google organic visits than Search Console records clicks, so pricing per click leans low.

## 6. The compounding curve (mid case arriving)

0%, 10%, 35%, 70% and 100% of the mid case's new revenue (A$158,623) live at today, month 3, month 6, month 9 and month 12.

| Point | Value (non-branded gross / yr) |
|---|---|
| Today | A$50,690 |
| Month 3 | A$66,552 |
| Month 6 | A$106,208 |
| Month 9 | A$161,726 |
| Month 12 | A$209,313 |

## 7. Smell tests run before publishing

- Non-branded converts (0.90%) below branded (3.04%), as it should.
- Value per non-branded organic click (A$5.76) sits below Google paid's value per session on the same Shopify report (A$6.17).
- No secondary-market rate is used: GA4 shows US organic visitors converting at 1.75% on 32 orders, above Australia's 0.97%. That is small-sample and plausibly the Leanne Ford halo, so the model runs on the blended Shopify rate only.
- The mid case (3,030 non-branded clicks a month) stays well under what comparable Australian lighting stores already earn on Ahrefs' estimates (Zest Lighting ~15,277 a month).
- GA4 records 16% less Google-organic revenue than Shopify over the same months, and its organic sessions fell to 0.6-0.8x Search Console clicks from June 2026 (a tracking change). Shopify is the load-bearing source.

## 8. What the model leaves out (all upside)

- **AI prompts.** 134 AI-chat questions in the library, counted separately and not priced.
- **Offshore demand.** About 2,514 non-branded clicks a year from outside Australia, spread across the US, UK, NZ and Europe, not modelled.
- **Adjacent demand.** Broad terms you stock with mixed intent (dinner sets, table lamps, home decor) sit outside the core.
- **Conversion improvement.** Held flat at today's figure.

## 9. Known limitations

- Shopify figures come from one channel screenshot; the table continues below the visible rows, so total store revenue is not known (the visible rows sum to at least A$390,931.50).
- The Shopify window (about Sep 2025 to Aug 2026) and the Search Console window (13 Oct 2025 to 12 Oct 2026) differ by about six weeks.
- No returns rate yet, so every figure is gross; a net view follows once returns are shared.
- No cohort data yet, so the first-year and lifetime rungs are modelled, as labelled.
- The cluster set is a first proposal, to confirm together; the counts recompute if it changes.