Public data, 16 Oct 2026. I can't see your conversion rates: anything not measurable from outside is marked unknown or as my hypothesis.
Orange outline = the stage where I measured the biggest leak.
| Stage | What I found | Why it matters |
|---|---|---|
| Named in an answer | Blocky 0/408, Clone.Store 4/408 category answers — all 4 ChatGPT with search, on "copy a store" questions. With search off, neither assistant knew either brand even by name (0/15 each). | Assistants recommend what their sources mention. None of those sources (best-of lists, Reddit, Shopify docs, App Store pages) mention you. |
| Described correctly by name | Clone.Store: ChatGPT 11/15 correct, Claude 3/15. Blocky: ChatGPT described "Blocky Fraud Filter", a different app, 3 of 3 times. Claude quoted an old 1.8★ (live: 3.8★). | A merchant who has heard of you and checks with an AI gets a wrong or negative answer — losing people close to buying. |
| Click → visit | ChatGPT tags links utm_source=chatgpt.com. claude.ai sends no referrer, so Claude visits look like "direct" traffic. | Without a "Where did you hear about us?" question, Claude's contribution stays invisible in any analytics tool. |
| Stage | What I found | Why it matters |
|---|---|---|
| Found in search | Organic rank: #45 "theme sections", #68 "page builder", #83 "sections". Not in the top 120 for "ai store builder" and 6 other searches. For "blocky", #1 is "Blocky Fraud Filter IP Blocker" (4.7★, 359 reviews). | A results page shows 24 apps, so Blocky first appears on pages 2–4. Shopify says Built for Shopify and positive reviews both lift ranking — the top 5 for "ai store builder" all have the badge. |
| Listing → install | Unknown — the listing has no GA4. Visitors see the thinnest listing in its set: 4 screenshots, no video, no free plan, 3.8★, no AI. | My hypothesis: it converts poorly next to competitors with free plans, video and more reviews. GA4 turns this into a number. |
| Install → paid | Unknown. blockyapps.com says "Get Started For Free" and "no credit card required"; its FAQ says $39.99/month after a 3-day trial. | My hypothesis: "free" wording lifts installs but also trial cancellations and "surprise charge" reviews. Check trial → paid in the Partner Dashboard. |
| Stays and reviews | 4 reviews in the last 90 days vs 57–286 for the leaders. The newest is a 1★ about content deleted after cancelling. | Review velocity is the most visible gap between Blocky and the apps above it — and it's what ChatGPT reads. |
Counted 16 Oct 2026 from each listing's reviews sorted by newest.
| Stage | What I found | Why it matters |
|---|---|---|
| Reach | Off-site footprint = one YouTube creator (4 videos, 10k–38k views each) + 1 Reddit comment. Brave (Claude's search) holds 2 clone.store pages, Google showed 4. Bing showed none (rough check — Bing Webmaster Tools gives the real count). | Few routes in. ChatGPT and Claude can only cite pages their search providers hold. |
| Site → sign up | Headline: "Built & launched free. In 3 minutes." FAQ: "$40 a month"; an old meta description: "no cost". No Pricing, About or Terms page (all 404). Mobile fails Core Web Vitals (LCP 4.1 s, TTFB 2.2 s). | My hypothesis: "free" brings sign-ups that drop at the $40 step, and assistants repeat "free". Missing pages leave "is it legit?" unanswered. |
| Sign up → build | Setup ~3 min, AI is real. "From popular brands" then takes 20–30 min, stuck at "10%" for 19+ min. An unfinished draft disappeared. | A stuck bar after a "3 minutes" promise looks like failure; some users leave before the build finishes. |
| Build → connect → pay | "Connect a store" installs Blocky. My hypothesis: Shopify's install screen shows "Blocky", not Clone.Store, at the moment of paying. | A different name at payment looks like bait-and-switch — and matches the 1★ "does not build anything with AI". |
| Publish → keep | Default output reuses the reference store's photos and invents reviews and claims (see the build). | A merchant who sees the risk won't publish; one who doesn't may face claims later. Retention and word of mouth suffer. |
You probably have site-level funnel data already; first check it's arriving.
| # | Step | Why | How you'll know |
|---|---|---|---|
| 0 | Turn on tracking (~3 h, week 1): GA4 + Measurement Protocol on the listing, Search Console and Bing Webmaster, check PostHog, UTMs, "Where did you hear about us?", funnel events. | Every later step needs a before/after number. | Weekly listing views → installs → paid |
| 1 | State one price, the same way everywhere (headline, FAQ, meta, listing, Blocky site). | "Free" vs "$40/month" vs "no cost" hurts trust at payment; assistants repeat whichever page they read. | Paywall → paid; no "free" in AI answers |
| 2 | Publish Pricing, About and Terms pages, incl. what can and can't be copied. | They answer "is it legit?" for people and give assistants something to cite. | Sign-up rate; AI answers to "is Clone.Store legit" |
| 3 | One name at the payment step: listing and install screen say Clone.Store. | Removes the name switch right before payment. | Connect → paid; no new "no AI" reviews |
| 4 | Ask for a review after a merchant's first successful build (Shopify's in-app prompt, no incentives). Check the cancel flow. | Review velocity is Blocky's widest gap. | Reviews per 90 days (baseline 4) |
| 5 | Fix the build progress display and the draft warning. | Keeps users through the 20–30 min wait. | Build started → completed |
| 6 | After step 0 shows mobile drop-off: speed up clone.store on mobile. | For conversion, not AI — no assistant links speed to citation. | Mobile vs desktop sign-up rate |
app_installed · onboarding_completed · first_section_published · trial_started · subscription_activated · app_uninstalled (+ reason)
signup_completed · build_started · build_completed · connect_store_clicked · store_connected · publish_paywall_viewed · subscription_started
store first-touch utm_source and referrer on the user
These become actions 0, 1, 3 and 5 in the plan.