Getting you found on ChatGPT
So the whole opportunity is one sentence: you hold 2.5% of the commercial non-branded organic clicks in your category, and every extra 1% is 88 new customers a year, worth about $40,459 at the lifetime spend your 2020 buyers have already reached. Your branded traffic is not part of this and does not change.
Your own cohort grid gives the ladder, and each rung is simply what customers have already spent. A first order is $99. A first year is $207: everyone who first bought in September 2025, tracked for 12 months. A lifetime is $458 so far: everyone who first bought in 2020, one-time buyers included, tracked for 5.7 years and still spending, so $458 keeps rising. About 1 commercial click in 19 becomes a customer, and the three value columns below price each scenario's new customers at each rung. Shopify cohort analysis Modelled, plus or minus 40%
| Case | Share of the pool | Clicks / yr | New customers / yr | First-order revenue / yr | First-year value / yr | Lifetime value / yr |
|---|---|---|---|---|---|---|
| Today | 2.5% | 4,070 | 217 | $21,530 | $44,947 | $99,469 |
| Conservative | 6.0% | 9,933 | 530 | $52,546 | $109,695 | $242,759 |
| Mid | 10.0% | 16,555 | 883 | $87,576 | $182,824 | $404,598 |
| High | 15.0% | 24,832 | 1,325 | $131,361 | $274,231 | $606,885 |
Read every column as value created per year at that pace, and all of it non-branded. The mid case adds +$66,046 a year of new first-order revenue on top of today: 883 new customers a year, worth $87,576 in first-order revenue, $182,824 by the end of their first year and $404,598 over their lives, with each customer priced at your measured $207 and $458 rungs (customer counts carried unrounded), where $458 is what your 2020 buyers have already averaged in 5.7 years and is still rising, so the lifetime column is a floor. The lifetime money arrives over each customer's life, not inside year one.
We score every brand against the same four pillars, 25 points each. Certified sits at 80 and requires Professional on all four, because a brand with one weak pillar does not get recommended.
Two things are already right and we would not touch them. Your robots.txt gives every AI crawler full access with correct exclusions for cart, checkout and faceted navigation, which is better than most Shopify stores we audit. And your product data carries real GTINs and SKUs on every stocked variant, which is the identifier layer most brands are missing entirely. Your Core Web Vitals also pass on real user data: Google's field measurements put your largest contentful paint at 1,479 milliseconds, interaction at 179 and layout shift at 0.03, all rated fast. Lighthouse lab scores of 68 and 63 are a simulation on a throttled connection and are not what your customers experience, so we have not treated them as a problem.
| Pillar | Points | What is holding it there, and what moves it |
|---|---|---|
| Measurement | 5 / 25 | +15 available. Analytics and Search Console are installed, which is table stakes and does not lift the score on its own. Attribution is materially broken, and we traced it: a second purchase tag on your order confirmation records most orders two or three times under different transaction ID formats, which is why GA4 reports a third more revenue than your Shopify admin. Separately, Klaviyo's non-standard UTMs push $2.03M of email revenue into Unassigned, and the June change in paid spend moved orders onto organic on last click. All three are fixable in week one. There is no AI-visibility tracking of any kind, so nothing tells you whether an assistant named you this month. |
| Technical | 15 / 25 | +5 available. Three specific gates: the review ratings that no crawler can read, duplicate Article and BlogPosting markup on every article carrying contradicting dates, and multiple H1 headings on 41 of 41 product pages. The master FAQ block on your FAQ hub is also invalid, because it contains JavaScript comments, so every parser discards it. |
| Content | 15 / 25 | +5 available. 705 articles clears the volume bar outright, and 40 of 41 product pages carry valid FAQ markup. The gate it fails is connection: the median article contains zero links into the catalogue, and 9 of 12 sampled articles have none at all. |
| Authority | 17 / 25 | +3 available. Your best pillar. Second of seven on domain rank in a verified Australian league. What holds it back, verified on the Australian index: your own site ranks top ten on seven of ten category questions, but no third-party roundup on any of those pages names your products, your review evidence is not surfaced, and there is no knowledge-graph entry for the business. |
The detailed findings, how the scoring works, and where every number comes from are in the appendix.
Women's health is a high-research, high-trust purchase. Buyers arrive with a symptom rather than a product in mind, they ask the same question several times in several ways, and they now ask it in two places: a search box and an AI assistant. The demand is large, it is well documented in your own account, and the AI layer is already answering above the results.
Every volume below is triangulated across three independent sources, Google Keyword Planner, SEMrush and Ahrefs, with the median taken. Ranks are your real Australian positions from your own Search Console over the last twelve months.
| Term | Searches / mo (AU) | Your rank now | AI Overview live? |
|---|---|---|---|
| how to lower cortisol | 6,600 | Not present | Yes |
| magnesium for sleep | 5,400 | Not present | Yes |
| period pain relief | 4,600 | 28 | Yes |
| best collagen supplement australia | 2,900 | Not present | Yes |
| cortisol supplements | 2,400 | Not present | Yes |
| best magnesium for sleep | 2,400 | Not present | Yes |
| hair growth supplements | 1,900 | Not present | Yes |
| menopause supplements | 1,600 | 6 | Yes |
| perimenopause supplements | 1,600 | 9 | Yes |
| dim supplement | 1,300 | Not present | Yes |
| endometriosis pain relief | 1,000 | Not present | Yes |
| pcos supplements | 720 | Not present | Yes |
Of the 87 commercial terms we triangulated, 70 return no impressions for you at all in Australia, 13 sit in striking distance at positions 4 to 20, 2 sit beyond position 20, and 2 are in the top three. 66 of the 87 fire an AI Overview. GSC Ahrefs
AI assistants are sending you customers before anything has been optimised for them. Over twelve months ChatGPT alone sent 1,375 sessions and 43 orders, converting at 3.1%, with Gemini, Copilot, Perplexity, Claude and the smaller assistants adding a further 247 sessions. That is a channel arriving on its own.
We then measured the answer layer directly. We ran 40 of your category's real questions as live Australian searches and read what Google's AI Overview actually cited.
Ranking and being cited have come apart. On "supplements for hormone imbalance women" you sit at position 4 organically while the AI Overview cites ABC News, Women's Health Network, Hers and Innerbody. The four answers that do name you are all condition-specific and practitioner-framed: adenomyosis twice, adrenal fatigue, and naturopath-formulated hormone balance. That is the shape of what you can win.
The most-cited sources are Healthline, YouTube, Medical News Today and GoodRx, and those are not realistically displaceable. The second tier is: Australian clinics, practitioner sites and telehealth. Swisse, Bioglan and Chemist Warehouse are all cited on supplement questions, so a supplement brand being named is clearly not blocked. 40 live AU SERPs, 3 September 2026
One picture, and it is the table above arriving. The three lines are the three value columns, priced at the acquisition pace you have reached: we build to the mid-case pace by month twelve, then graduate to the high-case pace by month 24. Non-branded organic customers only; your branded traffic sits outside all of it, unchanged.
By month six the system is yours, acquiring about 450 customers a year. Month twelve reaches the mid-case pace of 883 a year: $87,576 in first-order revenue, worth $182,824 by the end of their first year and $404,598 over their lives. Month 24, at the high-case pace of 1,325 a year: $131,361 of first-order revenue, $274,231 of first-year value and $606,885 of lifetime value. Lifetime figures price customers at the $458 your 2020 buyers have already spent, measured over 5.7 years and still climbing, not a projection. Modelled
We run this in three phases, and we move through them as the site is ready rather than on a stopwatch. Foundations fixes what is blocking the engines: the review markup they cannot read, the duplicate storefront competing with you, the schema conflicts and the broken FAQ block. Own your territories reconnects and rebuilds the content library around the searches and questions that actually sell, starting with menopause and the periods gap. Authority and compounding builds the third-party citations and the entity signals that decide whether an assistant names you. At the end of six months the workflows, the platform and the system are yours to keep and run. The detail is on the offer tab.
Twelve months to 28 August 2026, from your own Google Analytics. GA4 reports 2,298,515 sessions, 98,631 orders and $12,698,986. Your Shopify admin, read directly, records $9,525,231 on 76,726 orders at a $111.58 average order value over the same twelve months, so GA4 overstates revenue by about a third, exactly as you suspected. We traced the cause: a second purchase tag on the order confirmation fires most orders again under a different transaction ID format, into both of your data streams, and Klaviyo's non-standard UTMs push $2.03M of email revenue into the Unassigned bucket. Both are quick, permanent fixes. The table below is kept for its channel shape, which is directionally sound; every dollar figure in the money model comes from Shopify. Shopify + GA4
| Channel | Sessions | Orders | Revenue | Conv. rate |
|---|---|---|---|---|
| Paid Social | 895,707 | 3,233 | $351,451 | 0.36% |
| Direct | 485,942 | 18,570 | $2,633,584 | 3.82% |
| Paid Search | 264,145 | 20,890 | $2,854,139 | 7.91% |
| Unassigned | 185,944 | 19,113 | $2,373,662 | 10.28% |
| Organic Social | 169,768 | 1,698 | $181,523 | 1.00% |
| Organic Search | 143,005 | 16,601 | $2,167,778 | 11.61% |
| Paid Shopping | 92,988 | 11,192 | $1,189,531 | 12.04% |
| Referral | 14,322 | 1,276 | $172,741 | 8.91% |
| 12,109 | 457 | $63,630 | 3.77% |
Two things in this table need reading carefully, and they are the reason the money model does not run on GA4 revenue. Unassigned holds $2,373,662 that Google Analytics cannot attribute to any channel. And the organic conversion rate of 11.61% is not a search-discovery number: it includes existing customers who search your name and reorder. Appendix F sets out how we separated the two. GA4
The gap between being shown and being chosen, over twelve months of Search Console data, split by what kind of search it was.
| Search type | Impressions | Clicks | Click-through rate |
|---|---|---|---|
| Your own brand | 372,661 | 29,453 | 7.9% |
| A competitor's brand (Ritual Labs, Happy Mammoth) | 432,692 | 5,697 | 1.32% |
| "Happy liver", brand unclear | 206,411 | 9,369 | 4.54% |
| Category demand, named queries | 2,705,519 | 8,436 | 0.31% |
| Category demand, long tail withheld by Google | 2,864,675 | 36,481 | 1.27% |
Google withholds rare queries for privacy, so the final row is reported in aggregate only. We treat it as category demand because those searches are long-tail by definition, while brand terms are a small handful of high-volume ones. GSC
Four pillars, 25 points each. Each pillar is scored on measured evidence from the crawl, your own analytics, and third-party authority data, all gathered on 3 September 2026. Certified requires Professional on all four pillars and 80 or more in total, so a single weak pillar holds the whole score down.
| Pillar | Score | The evidence behind it |
|---|---|---|
| Technical | 15 / 25 | Genuine pass on crawler access: robots.txt runs User-agent * / Allow: / with correct exclusions for cart, checkout, account, sort_by and doubled facet parameters. No AI crawler is blocked. Core Web Vitals PASS on CrUX field data at origin: LCP 1,479 ms, INP 179 ms, CLS 0.03, all rated FAST. Lab scores of 68 and 63 are simulated and are not what users experience. Product identifiers are genuinely strong: real GTINs and SKUs on every stocked single supplement, with offers, price, AUD currency and availability. The 15 products with no GTIN are bundles, digital and non-supplement items. |
| Content | 15 / 25 | 705 articles clears the volume curve outright at 120-plus, and 40 of 41 product pages carry FAQPage schema with 172 valid Question nodes on the FAQ hub. Gate failure, internal linking: median product or collection links per article is ZERO, and 9 of 12 randomly sampled articles have none. The 1,883-word flagship sleep article carries exactly one product link and it 301-redirects to an old handle, so the link is stale rather than broken. 43% of the library is recipes. The commercially relevant education core is roughly 200 articles, not 705. |
| Authority | 17 / 25 | DataForSEO domain rank 35/100 and 467 referring domains, both measured 3 September 2026. This is a DataForSEO 0-1000 rank folded to /100 and must never be presented as an Ahrefs DR. Second of seven on domain rank and third of seven on referring domains in a verified like-for-like Australian league, behind Metagenics on both measures and Fusion Health on referring domains. The link count is not real: 19,717 of 21,021 backlinks, 93.8%, come from five self-owned properties, and 16,480 of those from the staging store. Genuine external links are roughly 1,304. |
| Measurement | 5 / 25 | GA4 property 302925420 and Search Console happyhealthyyou.com.au are both installed and verified, and Shopify analytics is in use. Installed tools are table stakes and lift nothing on their own. Attribution is materially broken: A$2,373,662 of revenue across 185,944 sessions and 19,113 transactions sits in Unassigned over 12 months. The June 2026 paid-media cut reshuffled last-click credit and organic conversion rate appeared to jump from 7.27% in May to 21.46% in July on flat organic sessions. Nobody caught it, so the clean measurement window has to be reconstructed as Sep 2025 to May 2026. |
Total 52 of 100, which places the brand at Structured. Team Empathy AEO Maturity Model
| Points | Pillar | The move |
|---|---|---|
| +15 | Measurement | Fix the measurement layer and set the baseline. Repair GA4 attribution so the A$2.37M Unassigned bucket and the last-click artefact from the paid cut stop distorting the read, capture an organic and AI-visibility baseline, and run a monthly rhythm of acting on it with share-of-voice tracked against the 129-prompt library. This moves behaviour from installed-and-ignored to a monthly action rhythm, which lifts the ladder from level 1 to level 4. |
| +5 | Technical | Clear the three Technical Professional gates. Strip the JavaScript comments from the master FAQPage block so it parses, cut every template to one H1 (41 of 41 products, the homepage and the collections), and write unique meta descriptions across the collection tail. Then server-render aggregateRating into the product JSON-LD, kill the hhy-au-staging.myshopify.com duplicate, and deduplicate the Article and BlogPosting schema. The weighted evidence already sits at 0.85, so the gate is what is holding this pillar down, not the substance. |
| +5 | Content | Connect the content to the catalogue and put a researched plan behind it. Link the 705 articles through to the 44 products and the real collections, starting with the 200-article commercial core, and repoint the stale product link in the flagship sleep article. Then build the cluster plan off the 129-prompt library and add the missing glossary, buying guides and ingredient hubs. Clearing the internal-linking gate alone takes this to 20, and a full cluster plan carries it past that. |
| +3 | Authority | Clean the link profile and earn real Australian citation. Kill the staging store so 16,480 false self-links stop counting, reclaim the 1,856 broken backlinks across 287 pages, prune the guest-post-farm tail, then surface and mark up the 559-review ProductReview asset and go after Tier 1 Australian citation (Australasian Menopause Society, Jean Hailes, healthdirect, ProductReview category listing, Mamamia and ABC). This pillar already clears the Professional gate, so the work is lifting domain rank toward 45 and turning 34 news links into a real editorial footprint. |
We triangulated 164 terms across three independent sources, Google Keyword Planner, SEMrush and Ahrefs, taking the median of the three. No single tool decides a number, because each one is wrong in a different direction. Where two sources disagreed by more than three times, the term is flagged in the workbook rather than quietly averaged.
| Slice of the basket | Terms | Searches / mo (AU) | In the model? |
|---|---|---|---|
| Commercial and solution intent | 87 | 51,095 | Yes |
| Medical definitions and symptom lookups | 71 | 278,040 | No |
| Generic ingredient commodities | 6 | 175,400 | No |
The excluded slices are 90% of the raw volume, and leaving them out is deliberate. Terms like "pcos" at 52,000 a month, "adenomyosis" at 40,500 and "magnesium glycinate" at 79,000 are answered by healthdirect, Mayo Clinic, Healthline and the big marketplaces, and a supplement brand will not displace them. That demand still matters for AI citation and for reaching people early, and Appendix D treats it that way. It is simply not counted as revenue.
| Territory | Your impressions / yr | Your clicks / yr | Your click-through rate |
|---|---|---|---|
| Menopause & perimenopause | 361,495 | 1,332 | 0.37% |
| PMS, periods & cycle | 305,135 | 431 | 0.14% |
| Hormone balance | 257,928 | 871 | 0.34% |
| Sleep, stress & nervous system | 214,290 | 320 | 0.15% |
| Liver & detox | 200,078 | 595 | 0.30% |
| Gut & digestion | 104,075 | 152 | 0.15% |
| Endometriosis & adenomyosis | 97,553 | 1,680 | 1.72% |
| Skin & hair | 58,591 | 63 | 0.11% |
| Weight & metabolism | 28,136 | 50 | 0.18% |
Endometriosis and adenomyosis is the smallest major territory here and by far the strongest performer, at more than five times your non-branded average. One article, "11 tips to manage adenomyosis naturally", earns 4,306 clicks from 190,183 impressions on its own. It is the clearest evidence in your own account that this works when the content is genuinely useful and properly targeted. GSC
Fourteen territories, built from the triangulated keywords and the AI prompt library together rather than from keywords alone. Every count is computed from the source files. Your current Search Console performance by territory is in Appendix C, so the two are never stated twice.
| Territory | Keywords | Commercial terms | AI prompts | Prompts you can uniquely own | Education articles you already have |
|---|---|---|---|---|---|
| Menopause & perimenopause | 30 | 15 | 22 | 17 | 76 |
| Sleep, stress & nervous system | 18 | 15 | 13 | 10 | 40 |
| PMS, periods & cycle | 23 | 12 | 16 | 11 | 34 |
| Hormone balance | 22 | 12 | 16 | 12 | 48 |
| Liver & detox | 12 | 6 | 12 | 9 | 22 |
| Gut & digestion | 11 | 10 | 9 | 5 | 16 |
| Endometriosis & adenomyosis | 13 | 5 | 13 | 12 | 6 |
| Ingredient & supplement education | 6 | 0 | 5 | 5 | 11 |
| Skin & hair | 11 | 6 | 5 | 5 | 31 |
| Weight & metabolism | 4 | 2 | 6 | 2 | 15 |
| Fertility, contraception & postpartum | 0 | 0 | 3 | 2 | 8 |
| Bladder & pelvic | 4 | 0 | 3 | 2 | 6 |
| PCOS | 6 | 2 | 4 | 1 | 6 |
| Thyroid | 4 | 2 | 2 | 0 | 4 |
| Total | 164 | 87 | 129 | 93 | 323 |
You have 705 blog URLs. The most important thing in this appendix is what they are made of.
| What it is | Articles | Impressions / yr | Clicks / yr |
|---|---|---|---|
| Editorial, inside a target territory | 323 | - | - |
| Recipes | 302 | 308,662 | 2,282 |
| Editorial, outside every target territory | 80 | - | - |
Of the 323 education articles that sit inside a target territory, 288 are refresh candidates: they already earn impressions and need commercial connection, internal links into the catalogue and answer-shaped structure rather than replacement. 35 are dormant and earn nothing. That is where the work starts, because reconnecting an article that already ranks is the shortest distance between the traffic you have and the revenue you do not.
| Territory | Education articles | Refresh candidates | Commercial terms uncovered |
|---|---|---|---|
| Menopause & perimenopause | 76 | 69 | 6 |
| Hormone balance | 48 | 46 | 8 |
| Sleep, stress & nervous system | 40 | 34 | 11 |
| PMS, periods & cycle | 34 | 34 | 5 |
| Skin & hair | 31 | 28 | 3 |
| Gut & digestion | 16 | 16 | 6 |
| Ingredient & supplement education | 11 | 11 | 0 |
| Weight & metabolism | 15 | 11 | 1 |
| Liver & detox | 22 | 10 | 5 |
| Fertility, contraception & postpartum | 8 | 8 | 0 |
| Endometriosis & adenomyosis | 6 | 6 | 3 |
| Bladder & pelvic | 6 | 6 | 0 |
| PCOS | 6 | 5 | 1 |
| Thyroid | 4 | 4 | 2 |
| Total | 323 | 288 |
A further 343 articles sit outside every target territory, 263 of them recipes. They carry 575,298 impressions and 3,093 clicks a year. GSC
Measured with DataForSEO on 3 September 2026, Australia. Domain rank is DataForSEO's own 0 to 100 scale and is not the same measure as an Ahrefs Domain Rating, so the two are never mixed inside one table. The competitor set was seeded by hand and each domain verified by loading its site, because automated discovery returned supermarket health portals, not-for-profits and a United States hospital.
| Domain | Domain rank /100 | Referring domains | Organic keywords | Est. traffic / mo |
|---|---|---|---|---|
| happyhealthyyou.com.au | 35 | 467 | 3,270 | 16,075 |
| metagenics.com.au | 39 | 890 | 9,276 | 127,475 |
| herbsofgold.com.au | 27 | 303 | 3,830 | 60,027 |
| carusosnaturalhealth.com.au | 29 | 407 | 3,708 | 47,628 |
| fusionhealth.com.au | 24 | 474 | 1,485 | 11,284 |
| promensil.com.au | 13 | 74 | 643 | 4,060 |
| naturopathica.com.au | 15 | 117 | 414 | 1,146 |
You have already won your shelf. On domain rank you sit second of seven, and on referring domains third. The problem is not that you lack authority against your direct competitors.
Of 21,021 measured backlinks, 19,717 come from five properties you own yourself, and 16,480 of those point at a staging store rather than the site that sells. Genuine external links are closer to 1,304, and only 34 come from news sites. That is a smaller and more fixable base than the headline number suggests, and it is the honest starting point.
AI answers in Australian women's health are assembled from a predictable set of sources: Jean Hailes, which carries 5,707 referring domains, the Australasian Menopause Society, healthdirect, health.gov.au, the Better Health Channel and the TGA's ARTG register. Practitioner bodies including NHAA, ARONAH and ATMS carry weight in a category where formulation credibility is the buying question. On the media side, Mamamia, Body+Soul, ABC and The Conversation are the outlets that get cited back.
Every query was classified into one of four buckets: your own brand and product names including Lisa Curry, a named competitor's brand, the ambiguous "happy liver" searches where the brand intended is genuinely unclear, and everything else. Only the last group enters the model. Google's withheld long tail is counted as category demand and allocated in the same proportion as the named queries.
Your organic revenue in GA4 cannot carry the model, for two reasons we could measure. First, paid media was reduced sharply from June 2026, with paid social falling from 79,063 sessions in May to 11,097 in August. Organic sessions barely moved, but organic orders went from 960 in May to 3,307 in July, taking the apparent organic conversion rate from 7.27% to 21.46%. Orders that had been credited to a paid click moved across to organic on last-click attribution. Second, $2,373,662 of revenue sits unattributed. So the measurement window used throughout is September 2025 to May 2026, before that shift.
Within that window we split organic sessions by the kind of page they landed on, because a subscription brand's branded traffic is largely existing customers reordering rather than search discovery.
| What they landed on | Sessions | Conv. rate | Avg order | Value / session |
|---|---|---|---|---|
| Your brand, or an account page | 29,323 | 23.18% | $142 | $32.90 |
| Your own product pages | 7,055 | 8.56% | $109 | $9.36 |
| A product page, from a non-brand search | 16,853 | 5.16% | $65 | $3.35 |
| A blog article | 32,359 | 0.31% | $107 | $0.33 |
Those per-session values are as GA4 reports them, and GA4 counts 28.5% more transactions than your Shopify admin records orders. So the model scales the conversion rates to Shopify's order truth (multiplying by 0.778), then prices a converting cold click at your measured first-order value of $99.16, since a new non-branded buyer is by definition a first order. That gives a floor of $3.98, an upper end of $6.60, and the model runs at the midpoint, $5.29. The $32.90 branded row is deliberately excluded, because claiming credit for your existing customers reordering would overstate what search can grow.
Roughly one search in five survives ads, shopping carousels and AI answers as a winnable organic click, and you currently take about one in forty of those. Modelled
| Step | Figure |
|---|---|
| Commercial category searches a year (87 terms x 12) | 613,140 |
| Plus the long tail, at 1.35x | 827,739 |
| Winnable organic clicks a year, roughly one search in five | 165,548 |
| What you win today | 4,070 (2.5%) |
| Value of one click, first order | $5.29 |
| Value of each 1% of the pool, a year, first order | $8,757 |
Shopify's customer cohort analysis, read directly, metric "amount spent per customer", cumulative. Your Sept 2025 cohort (1,178 customers) spent $207 each in its first year. Your 2021 cohorts (21,500 customers) have reached $411 each over 4.7 years, and your 2020 cohorts (24,700 customers) $458 over 5.7, with both still climbing, so $458 is a floor. That gives the ladder the tables price on: first order $99, first year $207, lifetime $458 and rising, roughly $55 a year of spend per customer after year one. The 24-month curve in section 4 stacks cohorts on exactly these rates.
Collaborator access to your Shopify admin was granted on 3 September 2026 and the model was re-run on it the same day. What it settled, over the twelve months to 31 August: total sales of $9,525,231 on 76,726 orders across 28,286 customers; new customers produced $1,331,548 gross across 11,977 first orders; the $99.16 the model prices at is Shopify's new-customer average order value, reported net of discounts and shipping (gross works out at $111 a head, so $99.16 is the conservative side), while returning customers produced $8,193,684, which is 86% of revenue, from 19,082 people averaging 3.39 orders and $429 each; and Shopify's own session data shows search referrals converting at 7.19% against 0.25% for social. Every revenue figure in this document is first-order only, and the measured repeat layer sits on top of it, uncounted. Still open: a landing-page level conversion split inside Shopify itself (the shape is taken from GA4 and scaled), and the subscription mix by product.