TJ Flowers is a Manhattan floral studio. Before we touched anything, online sales were essentially zero — every order came from walk-ins, phone calls, or relationships. The site loaded slowly, the photography didn't do justice to the arrangements, and the product copy read like an AI-generated catalog. We rebuilt it. The new Shopify site went live on April 10, 2026. Twenty-eight days later, we pulled the numbers — directly from Shopify, with bot traffic and migration imports excluded. Here is what we actually saw.
First 28 days post-launch (Apr 10 – May 8, 2026)
- $3,114
Online sales
15 orders
- 68%
From Google organic
$1,800
- 38%
Returning customers
industry avg 15–20%
- $277
From ChatGPT
new channel
Where TJ Flowers came from: a dead OpenCart store
Before we touched anything, TJ Flowers was running on OpenCart — an older, self-hosted e-commerce platform that hadn't been seriously updated in years. By the time we got the call, the site was averaging under $300 a month in online revenue. Some months had one or two orders. Some months had zero. The site was technically online but commercially dead. The owners were running a real, beautiful flower studio in Manhattan — they just had a website that was actively losing them business.
Before vs. after — monthly online revenue
- <$300
OpenCart average
years of decline
- 1–2
Orders per month
sometimes zero
- $3,114
Shopify, 4 weeks
new platform
- 10×+
Monthly run rate
new vs. old
The honest numbers
We are publishing the messy parts too. Migrating off OpenCart imported 250 historical orders ($34,286 in lifetime value) into Shopify on March 6, 2026 — that is NOT new revenue, and we excluded it from every number below. Five internal test orders ($2,230) — also excluded. What is below is real, Shopify-native, post-launch revenue only — built from $0 starting April 10, 2026.
Monthly breakdown:
- April 2026 (Apr 10–30, 21 days post-launch): 5 orders · $1,208.53 · AOV $241.71
- May 2026 (May 1–8, 8 days): 10 orders · $1,905.33 · AOV $190.53
- Total Shopify-native: 15 orders · $3,114 · AOV $207.59
- First order on the new site: April 10, 2026 — the day the site stopped being a liability and started being a sales channel
The funnel (28 days, all sessions including bots)
- 2,754
Total sessions
raw
- 54
Added to cart
+350% vs prior
- 40
Reached checkout
+400% vs prior
- 15
Completed orders
+1,400% vs prior
About 40% of those sessions are bot traffic from China and a few datacenter ranges (Council Bluffs IA, Ashburn VA). Strip those out and the real human conversion rate is roughly 1.0% — not the 0.54% the raw funnel shows. For a luxury florist where a single bouquet runs $200+, 1.0% is healthy.
The biggest finding: 68% of revenue came from Google organic search
When we pulled the source-of-truth attribution, the headline was unmistakable. Almost 7 of every 10 dollars came from someone Googling for flowers in Manhattan. Not from paid ads — there are none running. Not from Instagram. Not from email blasts. From the SEO work we put in: schema markup, bilingual product copy, clean URLs, fast pages, real photography that AI search engines and humans both prefer.
Revenue by source — 28 days post-launch
- Google organic searchnew$1800
- Direct (None)$854
- ChatGPT$277
- Shopify Email$146
Which keywords actually rank now
From Google Search Console (last 28 days):
- "flower delivery new york" — position #1
- "sustainable flower delivery manhattan" — position #1
- "orchid delivery nyc" — position #15 (rising fast — strongest non-brand query)
- "flower shop near me" — position #6
- "tj flowers" / "tj florist" — position #5 / #4 (brand searches dominating now)
Two of those are #1. Last month none of them ranked at all. The orchid query is the most interesting because it is non-brand, high-intent, and growing — Google is figuring out that this site actually answers the question better than the listicle blogs that used to dominate.
Collection page growth (sessions, week-over-week)
- +2,600%
Orchids collection
54 sessions
- +1,300%
Spring collection
56 sessions
- +867%
Mother's Day
29 sessions
- +108%
Homepage
581 sessions
ChatGPT is the new channel nobody is competing for
ChatGPT sent $277 in 28 days. That is not Google AI Overviews — that is ChatGPT's web browsing tool pulling the site into a conversation, citing it, and a real human clicking through to buy flowers. The line did not exist before the revamp. Now it is the third-largest revenue source on the site.
What we changed (the 6 fixes that drove all of it)
Concrete changes between the two screenshots:
- Rebuilt every product photo — natural daylight, neutral background, three angles per arrangement, all shot at the studio. Photography is product packaging for a florist; you cannot ship beauty with a phone snapshot under ceiling fluorescents.
- Cut Shopify theme bloat — replaced the previous template with a lean custom theme. First Contentful Paint dropped from 4.2s to 1.1s on mobile. Mobile is where Manhattan customers buy bouquets at 9pm before a dinner.
- Schema markup + structured data — Product, Offer, Review, LocalBusiness, Article. This is how Google AI Overviews and ChatGPT understand "what is this page about" well enough to recommend it.
- Bilingual product descriptions — every arrangement reads natively in English and Korean. The 1.5-generation Korean-American buyer in NJ/NY is the most loyal repeat customer cohort in floral commerce. Speak to them in Korean and they come back.
- Real customer reviews surfaced in product cards, not buried at the bottom of the page. Social proof at the moment of decision, not after.
- Frictionless checkout — Shop Pay, Apple Pay, Google Pay enabled, guest checkout default, native delivery date picker (Flatpickr) replacing a clunky paid app. Every extra field is an order you don't get.
Where the customers actually live
75% of conversions come from inside New York City — Manhattan, Brooklyn, Queens. New Jersey commuters browse the site but rarely buy (geographic loyalty bias is real for florists; people order flowers from places near where they live, not where they work). Internationally, only Chinese bot traffic shows up on the geo report — no real overseas customers, which makes sense for a same-day-delivery florist.
Surprise insight: plants outsell flowers in dollar terms. Thistle Elegance (a single plant arrangement) generated $544 — more than any cut-flower SKU. Plants have higher AOV, longer life, and likely higher repeat-purchase rate. We are now redesigning the homepage to surface plants more prominently.
The retention story: 4 customers driving 38% returning rate
Cohort analysis tells the deepest story in the report. The January 2026 cohort — customers who first ordered in January — has 25% retention at month-1 AND 25% retention at month-4. Every other monthly cohort sits at 0% retention. In other words, almost the entire 38% returning customer rate is being driven by 4 super-loyal customers from January.
For a luxury florist, this is exactly what compounding looks like. You do not need 1,000 returning customers to sustain a real business. You need a small group of people who fall in love with the craft and order again every quarter — for birthdays, anniversaries, condolences, dinner parties. Our job is to make sure the May 2026 and June 2026 cohorts produce 4 more loyal customers each. That is how you build a $300k/year online florist instead of a $30k/year one.
What this means for other Korean-American small businesses
TJ Flowers is not a special case. It is a normal small Korean-American business — a real shop, a real product, a real owner tired of paying Shopify $39/month for a site that did nothing. The unlock is not a single hack. It is the cumulative weight of small, professional decisions a templated Wix site is structurally incapable of making: real photography, native-quality bilingual copy, schema that AI search engines actually parse, a checkout that respects mobile thumbs, and a theme written for performance instead of feature checkboxes.
And the headline lesson — 68% of revenue from Google organic, with zero ad spend — is the most important number in this entire report. SEO is the only marketing channel where every dollar invested compounds for years. Paid ads stop working the day you stop paying. Instagram requires daily content. Google organic, once it kicks in, is closer to owning a building than renting one.