Everything is $1–$3. The store holds $0 of inventory.

The research · researched 2026-10-22

How a $0-inventory dollar store actually works

The four questions — where products come from, how they’re sold, how they reach the customer, and how the profit is collected — answered from current supplier, carrier, platform and FTC sources. Every figure below is from that research; where a number could not be verified, it says so.

Research · Sourcing

Where the products come from

Product cost is not the problem — factory prices for the right micro-items really are $0.01–$0.25. The problem is that no supplier can pick, pack and deliver one $1 item to a US home for the ~$0.50 total a 50%-margin $1 item would allow. Only two routes are truly $0-inventory: supplier-catalog dropshipping (CJ, AliExpress/DSers) and curated US dropshippers at roughly double the unit cost — and a route only counts if its terms let us hold supplier payment until the customer’s payout has landed in our bank.

*Every supplier-side time in this table starts only when that supplier is paid. A checkout-to-door promise always adds the payout wait first — ~2 business days, up to 5 for a new account — then the payment-time stock/price re-check, supplier processing and transit.

What one item actually costs, delivered

Research · Tariffs

The end of duty-free small parcels

Research · Selling

The storefront: Shopify, WooCommerce, or custom

Sales tax: Shopify includes automated calculation and liability tracking, but economic-nexus collection duties fall on the seller on any platform. Tax is not costed in the tables below (it varies by state and volume) — one more drag on “net.” Platform and fee figures were verified against the official Shopify and Stripe pricing pages via index fetch on 2026-10-22 — the research browser could not load shopify.com live (DNS failure), so those two pages were not verified in a live browser session.

Research · Fulfilment

From checkout to the customer’s door

Research · Profit collection & unit economics

The math: can an order ever net 50%?

Payment fees first, because they set the ceiling. Stripe and Shopify Payments both charge 2.9% + 30¢ per successful domestic card charge. The percentage is painless; the fixed 30¢ is not — it is 30% of a $1 sale all by itself. PayPal through Shopify runs from 3.49% + 49¢ (52.5% of a $1 charge), and putting external Stripe on Shopify Basic instead of Shopify Payments adds Shopify’s 2% third-party fee on top. Stripe also keeps its fee on refunded charges and charges $15 per dispute.

Assumptions for every table (deliberately optimistic — reality is worse): product cost is 25% of list price ($0.25 per $1 item — achievable only at factory or best CJ/AliExpress pricing, not at US curated dropshippers); one consolidated parcel per order; shipping actually paid is $5.00 per parcel (below the documented $6.66–$8.40 USPS commercial benchmarks — a negotiated light-mailer rate that could not be publicly verified); card fee 2.9% + 30¢ on the total charged. Excluded: platform and app fees, returns, chargebacks, tax, advertising and your time. Margin = net ÷ total charged to the customer.

Table A — Free shipping for the customer (store absorbs $5.00/parcel)

Even if the products were free, the $1 and $5 rows still lose money — the $5 parcel and the 30¢ fee exceed the takings. With free shipping, 50% net only appears at a ≈$30+ basket.

Table B — Customer pays $4.99 shipping per order

Table C — Table B plus the always-on 10% discount (the model this store runs)

Fixed costs push Table C further down: Shopify Basic plus ~$30 of apps is $69/mo — $0.69 per order at 100 orders/mo (the $15 row falls to ≈44.5%, the $10 row to ≈36%), and a 5% returns allowance takes roughly 5 more points. Add the Plinko wheel on top and the $15 case falls from 48.2% to ≈46% — the expected reward cost is 7.75% off product plus a 5% chance of a free $3 item riding in the parcel, but the cart caps the combined discount at 15% of the basket, so the reward layer can never add more than 5 points beyond SAVE10 on one order. Reward cost always moves margin one direction: down.

Try it yourself

Margin calculator

The research formula, live. Set a basket, a discount, a shipping charge and a sourcing route, and see the exact order economics — including the smallest basket at which 50% net becomes possible under your settings.

Shipping paid is held at the research’s optimistic $5.00/parcel assumption (documented US commercial benchmarks run $6.66–$8.40). Returns, chargebacks, sales tax and advertising are excluded — every real-world addition lowers the result. Raising the shipping charge is the strongest lever, but charging “shipping” above cost is economically just raising the price, and springing it late in checkout invites FTC and state drip-pricing scrutiny.

Research · Real photos for every product

5–10 real photos per item: yes — if the catalog is picked backwards

The requirement stands exactly as set: every product gets 5–10 real pictures of the exact item it honestly is — front, back, angle, in-use, scale, detail, what’s in the box, packaging, and the variants actually sold. No AI images, no stock photos of a similar item, no other variant’s shots padding the count. The follow-up research verdict: achievable — conditionally. It works only if the catalog is reverse-selected: start from supplier listings that already pass a 5-real-photo gate, and drop or replace every SKU that doesn’t. It is not achievable for an arbitrary pre-chosen 2,000 items, because the photos have to already exist — and under this store’s hard rule (we never hold product, ever) nobody here can shoot the ones that don’t.

Where the photos actually come from

Nobody at scale hand-downloads images: the supplier’s listing media imports with the product. On CJ, List / Add to Import List pushes the listing — photos, description, price — into Shopify in seconds, in bulk. AliExpress via DSers works the same way (Import List → edit images → Push to Shopify), and DSers’ free tier covers 3,000 products, enough for this whole catalog. The binding constraint is what a listing contains: AliExpress listings are built on a maximum of 6 main images, plus description images and per-variant images — so a good listing yields 8–12 raw images, of which typically 5–8 are honestly usable for the exact variant sold, after deleting duplicates, pure-text infographics and other-variant shots. A thin listing yields only 3–4 — and fails the gate. Every card in the store above carries its own shot list (open “photos required” on any product) showing exactly which shots that item needs.

Verification without ever receiving a product

Supplier photos are treated as unverified claims until they pass this stack — none of which involves product coming to us:

· Supplier gate: AliExpress stores over 95% positive feedback with an Item-as-Described score of 4.7+, established 1–2+ years; on CJ, vetted catalog products with fulfilment history, not fresh listings.
· Review-photo cross-check (the highest-value check): the listing gallery is compared against photos buyers posted in reviews of that exact listing — real colour, stitching and packaging. If review photos materially diverge from listing photos, the SKU is dropped. Review photos are for internal verification only; the buyer owns their copyright.
· Internal consistency: main vs description vs variant images vs written specs. A listing that contradicts itself fails.
· CJ QC photos at fulfilment: for the first customer order of every new SKU, CJ photographs the actual unit in its warehouse before it ships; if it doesn’t match the published gallery, the SKU is paused and the customer refunded proactively. The customer receives product — we never do.
· CJ custom photography (top sellers only): CJ shoots photos/video of the actual product in its own studio as a paid service — priced per product (product + labour + optimising), photography in 3–7 business days, a studio product video cited at $50. At a per-SKU quote this is a tool for the top 50–100 proven sellers (budget roughly $750–$4,000 at CJ’s quotes), never for all 2,000.

Rights: there is no blanket licence

CJ’s dropshipping agreement puts trademark and copyright responsibility entirely on the seller and grants no licence to reuse listing photos; AliExpress’s terms grant a licence to AliExpress, not downstream to dropshippers. So the workflow is: written supplier permission, kept on file, per supplier — a message confirming we may use the listing’s photos and video to sell units bought from them, saved with date and store ID. No reply or a refusal means that supplier’s products don’t list. A watermark burned into a photo is a reject signal, not an editing job. And only unbranded, generic goods are listed at all — if a photo shows a recognisable brand or character, enforcement lands on us, the seller of record.

The scale math: 10,000–20,000 images

Staged photo plan

Stage 0 (Week 1): permission templates sent to the CJ agent and top AliExpress suppliers; rights log started; audit checklist adopted. Stage 1 — pilot, 100 SKUs (Weeks 1–2): measure the pass rate at the 5-photo gate, audit minutes per SKU, and review-photo mismatch rate. Proceed only if pass rate ≥60% and mismatch <5% — otherwise cut the 2,000 target or change the category/supplier mix; never lower the photo standard. Stage 2 (Weeks 3–5): first 500 live, in batches of ~100, first-order QC-photo checks running. Stage 3 (Weeks 6–12): to 2,000 at pilot-measured capacity, failures replaced from candidate overage. Stage 4 (ongoing): CJ custom shoots for the top 50–100 sellers; permissioned customer review photos grow compliant sets from 5 toward 7–8.

Research · Pricing shipping the fun way

Fill the Box: one box, one price — and shipping shrinks as you fill it

The model the storefront above now runs. The box is the unit: one shipping charge per order, never per item, and the charge visibly shrinks as the basket grows — a deterministic progress meter in the cart (“$3.20 away from $2.99 shipping” → “$1.50 away from FREE shipping”) turns the threshold into a game the shopper tries to win. No randomness anywhere near the price: a spin-wheel for a shipping discount would be prize + chance tied to a purchase, the illegal-lottery pattern — the always-on discount stays deterministic and any random reward stays free-entry and separate (see Reward rules).

*Worst-case contribution = total charged − card fee (2.9% + 30¢) − $5.00 parcel − product cost, with product cost at the store’s listing cap of ≤50% of shelf price and no discount applied. The live cart above models the buying target instead — product cost at 25% of list with SAVE10 applied — and shows +$12.02 net (52.3%) on the same $20 basket. Same parcel, same card fee: the two sets of figures differ only in the cost basis, and each is labelled wherever it appears. In the live store, tiers are applied to the basket after discounts, and the cart meter shows the live tier as you shop.

Why thresholds, and why this shape

Orders with free shipping average ~30% higher value; 58% of consumers add items to qualify for a free-shipping threshold and 93% say free shipping encourages them to buy more. Practitioner guidance is to set the free threshold ~30% above the typical basket and show a progress bar — for this store, a $20–$23 target basket puts free at $30, with the $2.99 step at $25 rewarding the first stretch. The models that lost: free shipping on everything (a $15–$20 order contributes only $1.76–$4.12 — one $15 chargeback fee erases 4–8 orders); per-item shipping (wins the arithmetic, loses the store — it taxes exactly the basket-building a dollar store needs, and 48% of cart abandoners cite extra costs as too high); shipping included in price (needs ~+$0.50 per item, breaking the $1–$3 promise); and a $9.99/yr unlimited shipping club — it nets $9.40 per member-year after card fees against $5.00 of unrecovered parcel cost per order, a break-even of 1.88 orders per member per year; a monthly-ordering member burns $60 of shipping against $9.40 collected. A club is only discussable later at ≥$29.99/yr with a $25 minimum basket and a disclosed fair-use cap. Fallback if the tiers confuse: flat $4.99, free at $35 — at $35 the contribution is $11.19, the safest free threshold in the research.

The three conditions that make these numbers true

(1) One parcel per order — CJ-consolidated or single-supplier fulfilment only; a 10-item AliExpress basket from 6 sellers is up to 6 parcels and 6 shipping charges, which silently destroys the model. The storefront enforces this in the cart: once a box holds items from one supplier route, an item from a second route cannot be added to it — the shopper is told why and offered a fresh box (a second order) for the other route instead. (2) The ≤50% buying cap — no SKU lists whose product cost exceeds 50% of its shelf price. (3) A real parcel line ≤$5.00 at the weights actually shipped, confirmed in writing per SKU from CJ before launch. That third condition is an open question, stated plainly: a CJ in-account calculator test (Oct 2026, 500 g) showed China→US lines at $10.79–$12.07 and even US-warehouse fulfilment at $9.79 — above the $5 planning figure. If the signed quote lands at the $6.66–$8.40 US benchmark range instead, the tiers shift to $5.99 / $3.99 / free at $35 and the math is re-run before trading. Legally, the tiers are built to comply by construction: shipping charged never exceeds the real parcel cost (shipping may be excluded from an advertised total only if it reasonably reflects actual cost; “handling” charges may not be excluded at all — so this store has no processing, service or handling fee, ever).

When the cash actually moves — payout first, supplier second

Day 0 — customer pays; we have not been paid. Shopify captures the card and deducts 2.9% + 30¢, but under the owner’s definition, “get paid” means the money has landed in our bank account. Until that happens the order is “Awaiting payout”: the supplier is not paid, the supplier order is not released for processing, no CJ/DSers balance is topped up, and no float, credit line, advance or store funds bridge the gap. There is no supplier float to hold or replenish — that is the point of the rule.

~2 business days later (up to 5 for a new account) — payout lands. Only then is that order’s supplier bill paid, from that order’s received funds, and only then does supplier processing start. Reconcile the Shopify payout to the order before releasing payment, and pay no supplier invoice that cannot be tied to a landed payout. Operating cash may pay the store’s own platform and app bills; it may never be advanced to a supplier for a customer order.

The known CJ consequence: CJ holds an item for only 24 hours on an unpaid order, so that hold will expire before we are allowed to pay. Nothing is reserved for us during the payout wait. At payment time we re-check live stock and the full supplier price: if the item is still available and the total cost is still at or under the store’s ceiling, we pay and processing starts; if it is unavailable, or the price has moved above the ceiling, we cancel and issue the customer a prompt full refund — no substitute, no silent delay and no surprise extra charge (the FTC Mail Order Rule requires honest shipment timing and refund handling when an order cannot ship as promised). How often this will matter is unproven until the pilot: the 100-SKU pilot must measure stockouts and price-over-ceiling failures across the real 2–5 business-day payout window before the model scales.

A later chargeback still costs the order amount plus a $15 fee once a supplier has legitimately been paid after payout; the rule removes fronting, not post-payment dispute risk. The lean fixed stack — Shopify Basic $39/mo + DSers $0–$19.90 + Matrixify $20/mo when bulk-editing — totals ~$39–$100/mo, about 5–12 orders a month at ~$9 worst-case contribution just to cover it.

Research · Worldwide markets

Worldwide: local warehouse to local customer, or don’t go

The governing fact everywhere: at $1–$3 prices, the math only survives when the supplier’s own stock in the customer’s country ships domestically. We hold nothing and ship nothing in every market, and the payout-first rule applies in every market: supplier payment waits until that customer’s payout has landed in our bank, so every local delivery window includes the payout wait before supplier processing and transit. CJ’s overseas-warehouse model only counts where CJ or its suppliers already stock the SKU as their own catalog stock. Where “US warehouse” turns out to mean our pre-stocked inventory (secondary sources cite minimums of 10 pieces per variant, 100 total), that SKU is out of scope for us, however fast it would arrive. Confirming which SKUs are unit-by-unit catalog stock with stock/price stability across the payout window — in writing, from a CJ agent — is the single most important open question in this whole plan.

Margin basis in this table: the margin checks use the worst-case basis — product cost at the 50% listing cap, no discount — the same basis as the Fill-the-Box table above. The live cart models the 25% buying target with SAVE10 applied, which is where the higher hero figures come from (+$12.02 net, 52.3%, on a $20 US basket). Money per market: price in round local numbers, set manually — auto-conversion turns $1.99 into £1.57, a price no discount shopper trusts. Shopify Markets charges 1.5% per transaction for currency conversion from a US store; do not enable Managed Markets (a further 3.5%) at these basket sizes. Rewards law abroad (headline): UK free prize draws need a genuine free entry route and CAP Code compliance; Australia’s chance-based trade promotions stay permit-free nationally if the total prize pool is kept under A$3,000; Canada needs a no-purchase entry route and odds/prize disclosures. The always-on deterministic discount travels everywhere; the random rewards launch US/UK/AU only.

Kill criteria — decided before launch, not after

Photo-gate pass rate under 60% in the 100-SKU pilot → cut the catalog target, never the standard. “Not as pictured” refunds over 5% of orders for a supplier or category, two weeks running → delist it. Average parcels per order above 1.3 → pause multi-supplier SKUs. Supplier DDP quote above $6.50 on over 20% of live SKUs → re-price, re-source or delist. Contribution under $2 per order for two consecutive weeks in a category → kill the category. A payout that has not landed → the supplier stays unpaid and processing does not start; never bridge the wait with float, credit or store funds. If the payout wait makes the published delivery promise impossible, notify the customer and refund under the published policy. Any supplier requiring a deposit, a prepaid balance, an advance, or payment within 24 hours of order placement to keep an item → incompatible; remove the supplier/listing. Payout-time stockout or price-over-ceiling failures become frequent in the pilot → shrink to suppliers/listings proven stable across the payout window, or do not launch that route. Chargebacks approaching 1% → freeze scaling. Any customer charged duty/VAT on delivery in a market sold as DDP → suspend that market immediately.

Research · Discounts & rewards legality

Always-on discounts and falling rewards: what’s legal

Research · Verdict

Verdict & the model worth building

• $20 minimum order, hard-gated at checkout — the threshold that clears 50% on that order (52.3% before fixed costs and returns). The earlier $15 gate was profitable but let a $15 order through at 48.2%, under the rule.
• Fill-the-Box shipping (see Shipping): one charge per order — $4.99 at $20–$24.99, $2.99 at $25–$29.99, free at $30+ — and only while the product-cost ratio holds at ≤$0.25 per $1 in the live cart model (≤50% of shelf at the listing cap).
• Payout first — hard rule, no exceptions: a supplier is paid only after that customer’s Shopify payout has actually landed in our bank account. Checkout capture is not payment. No float, credit, deposit, prepaid supplier balance or advance is ever used. The order timeline is customer pays → payout wait (~2 business days, up to 5 for a new account) → stock/price re-check → supplier paid from the received funds → supplier processing starts → shipping. If the re-check fails, the customer gets a prompt full refund.
• Bundles are single items: a 50-sticker pack, a 12 hair-clip set, a balloon-and-ribbon party set — each listed as one $2–$3 SKU. A 50-pack of factory stickers costs $0.50–$1.00 total, not 50 × a single-unit cost. This is the only way a $3 shelf price carries real margin.
• Buying ceiling — checked twice: ≤$0.20 of product cost per $1 of list for $1 items; ≤$0.60 all-in for a $3 bundle, both at listing and again at supplier-payment time after the payout lands. If the price has moved above the ceiling during the payout wait, do not pay — refund the customer promptly instead.
• One warehouse, one parcel. A basket split across two suppliers means two $5 parcels and an instant loss.
• Supplier: CJ Dropshipping primary, DSers/AliExpress secondary — selected for stable stock and stable prices across the 2–5 business-day payout window, not just the lowest day-one quote. Nothing is ever held or shipped by us — no pre-stocking a CJ warehouse, no micro-stock at a consolidator: a SKU whose only fulfilment route needs our stock does not get listed. Any supplier requiring payment within 24 hours of order placement to keep an item is incompatible with the payout-first rule. (An earlier draft of this model allowed tiny bulk buys at a consolidator; the owner’s never-hold rule removes that, and the bundles are supplier-packed multipacks instead.)
• Platform: Shopify Basic, Shopify Payments only — never an external gateway on Basic (+2%).
• Catalog mix stays 90 / 5 / 5 at the listing level — 90% of listings at $1 (basket-fillers), 5% at $2, 5% at $3 (the engineered bundles). Margin is tracked on the order, never per item.
• Returnless refunds for anything defective, wrong or lost — return postage exceeds the value of everything in the store.
• No paid ads at launch. Any acquisition cost above ~$1–$2 per order erases the margin in Tables B and C. Traffic has to be organic, social, SEO and email.
• Promotions: one honest always-on code ≤10%, already inside the pricing math, plus a compliant entry reward — all entry paths free (arrival drop, Play button, AMOE on request and exit claim — no purchase, payment, email or account to enter or claim), published odds, capped prizes, redeemable only on $20+ baskets — or a guaranteed basket-tier reward instead of chance.
• Not built on: Temu (reseller orders prohibited), Kole/DollarDays case packs as single items (the unit is the pack; shipping starts $6.95), print-on-demand (base cost exceeds the cap).

Open questions the research could not close

· Whether specific CJ SKUs shown as “in the US warehouse” are available unit-by-unit without a private-inventory (pre-stock) order — CJ’s public material is ambiguous and secondary sources describe pre-stock minimums. The never-hold rule turns on this; get it in writing from a CJ agent before listing.
· How often CJ/AliExpress stock disappears or the price moves above the ceiling during the 2–5 business-day payout wait, after CJ’s 24-hour unpaid hold has expired — unproven until the pilot. The pilot must report the payout-time stockout rate and price-breach rate separately; a route that only works by paying inside 24 hours is not viable under this rule.
· The store’s real DDP parcel cost — the $5.00 planning figure, the $6.66–$8.40 US benchmarks and a $9.79–$12.07 CJ calculator test all disagree. The Fill-the-Box tiers are correct for a $5 line and must be re-cut from signed per-weight quotes before launch.
· CJ’s exact custom-photography price per SKU — CJ quotes per product (only a $50 product-video figure is published). Get quotes for the top sellers in the pilot phase.
· How often supplier photos actually misrepresent the shipped item — no credible published rate exists; the photo pilot measures our own rate instead of citing one.
· Whether H&J Closeouts runs any single-unit dropship program — no reliable published program was found; verify with them directly.
· The exact current tariff stack for any specific item — rates changed four times in 2025–26 and vary by product code; the ≈35–37.5%-above-base tracker range is a range, not a quote. Get a broker quote per SKU.
· Whether a sub-$5 US light-parcel rate can actually be negotiated — public benchmarks are $6.66–$8.40; the $5.00 in every table is an optimistic assumption, and at $6.66 the 50%-net minimum basket rises to ≈$27.

Provenance

Sources

Research conducted 2026-10-22 from the sources below, with a business-first follow-up the same day covering real product photos, shipping pricing, cash timing and worldwide markets. Prices and fees are 2025–2026 figures as cited in each section. CJ/DSers specifics from the follow-up are search-index evidence — the follow-up’s live check of CJ’s blog failed on DNS — so every load-bearing CJ/DSers term is slated for in-account confirmation before any money is spent.

• Shopify official pricing — Basic $39/mo ($29 annual), 2.9% + 30¢, PayPal from 3.49% + 49¢, third-party fee 2%: shopify.com/pricing
• Stripe official pricing — 2.9% + 30¢, $15 dispute fee: stripe.com/pricing
• Photos — CJ/AliExpress import, image counts, rights: cjdropshipping.com · storeautomator.com (max 6 main images) · CJ Dropshipping Agreement (IP clause) · CJ photo/video shooting service · Shopify Community — supplier permission · alidrop.co — supplier vetting & review photos · Shopify product media limits · Matrixify pricing
• Shipping pricing & “free” claims: FTC Guide on “Free” (16 CFR 251) · olshanlaw.com — drip pricing · redstagfulfillment.com — free-shipping data · charle.co.uk — thresholds & progress bars · Walmart+ pricing
• Cash timing: folio3.com — payout timelines · onrampfunds.com — payout holds & funding gaps · CJ order flow — supplier processing starts only after the supplier is paid · FTC Mail Order Rule guide
• Worldwide: avask.com — EU €3/HS-line duty 2026 · landmarkglobal.com — UK £135 VAT rule · stripe.com — AU GST registration · cbsa-asfc.gc.ca — Canada courier thresholds · omnisend.com — UK Chinese-marketplace shopping · ecdb.com — EU cross-border spend · atomixlogistics.com — de minimis suspension · lawpath.com.au — AU promotion permits
• De minimis timeline: fashionlawjournal.com · fulfill.com/section-321 · icpainc.org · importcostcalculator.com
• CJ Dropshipping costs, shipping, US warehouse: super-ds.com · easync.io · blog.cjdropshipping.com · landed example: gp-fulfillment.com
• AliExpress / DSers: ecommerceparadise.com · dodropshipping.com
• YunExpress / China parcel costs: sourcingxpro.com · ecommpanion.com · nextsmartship.com
• USPS Ground Advantage: goshippo.com · simplfulfillment.com
• Kole Imports dropship program: support.koleimports.com · closeout examples: groupon.com/coupons/koleimports
• DollarDays: thecleverbusiness.com · salehoo.com · dollardays.com
• Alibaba factory prices: accio.com (custom stickers) · accio.com (vinyl stickers) · made-in-china.com
• Temu no-dropshipping / local pivot: cedcommerce.com · accio.com · importify.com
• DHgate: sourcelow.com · ecommercefastlane.com
• Shopify vs WooCommerce costs: commercegurus.com · 365dropship.com
• Returns / returnless refunds: dropified.com
• FTC sweepstakes & promotions law: consumer.ftc.gov · blog.brandmovers.com · PCH case: thesocialmedialawfirm.com
• FTC deceptive pricing, 16 CFR 233.1: law.cornell.edu
• Shopify gamified-discount apps (Plinko / spin wheel): apps.shopify.com/prizo · easyappsecom.com