The Delivery Gap: What Your Product Has to Pay For

The gap your product has to pay for

A seller shipping from China is not competing against an empty shelf. They are competing against whatever the customer could plausibly buy instead, delivered on whatever timescale that alternative now offers. In the second quarter of 2026 the large U.S. retailers made that timescale substantially shorter: Home Depot said more than 65% of its in-stock parcel deliveries are same day or next day and that it had launched three-hour delivery nationwide on tens of thousands of products; Target said its stores fulfil more than 95% of its sales and that it moved nearly 30% more same-day and next-day units than a year earlier.

That does not make direct-parcel dropshipping unviable. Plenty of products still travel perfectly well by post. What it does is raise the price of being generic. The difference between your delivery and the local alternative is a real cost the customer bears, and something in your offer has to pay for it.

Call that difference the delivery gap. It is not only extra days. It is uncertainty about the arrival date, a slow first tracking scan, a long tail of late parcels, harder replacement, missed birthdays and deadlines, more effort to return, and less trust before the customer has bought anything at all.

What pays for it is the product gap: a design that is genuinely unavailable locally, a meaningful specification or material difference, personalisation, a better problem-solution bundle, a demonstrable functional improvement, a sufficiently better complete price, better proof for a narrow use case, or value the customer cannot substitute.

A useful way to hold the two together is a decision model rather than an equation: net offer advantage = product advantage + price advantage + proof advantage − delivery penalty − trust penalty − return penalty. It does not produce a number. It produces the right argument, which is that a delivery penalty has to be paid for out of something, and "more advertising" is not one of the terms.

What the customer is actually comparing

Before measuring anything, write down the comparison the customer is really making. For each of your top three destination countries, record the three closest credible alternatives and their complete proposition.

• Product match: exact, close substitute, or merely visually similar.
• Specification: material, dimensions, capacity, included parts, compatibility.
• Total delivered price: item, shipping, tax and any accessory the customer must buy to use it.
• Delivery promise: the advertised range, and the observed delivery where you can see it.
• In-stock status: real current availability, not a cached listing.
• Reviews: count, rating and the complaints that repeat.
• Return and replacement: cost, time and difficulty.
• Trust: recognisable retailer, marketplace protection, familiar payment options.
• Intellectual property: brand, design, patent or licensed-character exposure on either side.
• Customer mission: urgent replacement, planned purchase, gift, novelty or recurring need.

A similar-looking listing is not automatically a substitute. But the burden is on you to write down why it is not. If the only difference you can name is your own marketing, you are selling a commodity into a faster market.

The mission line matters more than sellers expect. Urgency and slow fulfilment are the worst possible combination, and a strong need makes it worse rather than better. A customer who needs the part to finish today's repair is not a customer you can serve from a two-week parcel, no matter how good the price is.

Measure the route as a distribution, not an average

Most sellers carry one number in their head: roughly how long shipping takes. That number is almost always a median dressed up as a promise, and customers do not experience a median. They experience their own parcel, and the ones who complain, refund and leave reviews are disproportionately in the tail.

Track each SKU × variant × country × route through every stage, not end to end: order to purchase, supplier processing, warehouse receipt, quality control and packing, dispatch, first tracking scan, international transit, customs, final mile, delivered date, failed delivery, loss, and return to sender. Stage-level timing is what tells you where to intervene. An end-to-end number tells you only that you have a problem.

From that, calculate a small set of figures and keep them per route rather than per store.

• P50 delivery time: half of parcels arrive by this point.
• P90 delivery time: nine in ten arrive by this point.
• Tail gap: P90 minus P50 — the single best summary of how unpredictable a route is.
• Tracking activation lag: order or dispatch to first carrier scan.
• Dispatch reliability: share of orders dispatched inside the promised handling window.
• Delivery-promise hit rate: delivered inside the displayed range, divided by delivered orders.
• WISMO rate: "where is my order?" contacts per 100 orders.
• Refund, reshipment and chargeback rate, split by delivery band.

A wide tail gap is usually a supplier-processing or handoff problem rather than a carrier problem, which is why the stage split matters. If you are choosing between routes rather than diagnosing one, the controlled method is in how to run a controlled shipping route test, and the reason two tracking numbers can describe the same parcel is in tracking handoffs.

Price the delivery-gap tax

Once you can band your orders by delivery time, the delivery gap stops being a feeling and becomes a cost. For each band, compare checkout conversion, pre-dispatch cancellation, WISMO contacts, refund probability, reshipment cost, chargebacks, review sentiment and repeat purchase.

Then estimate the expected fulfilment-failure cost per order: refund probability × refund loss, plus reship probability × reship cost, plus chargeback probability × chargeback cost, plus variable support contacts × cost per contact, plus expected loss and damage cost.

That figure is the delivery-gap tax. It is usually larger than sellers assume, because most of it is booked in places that do not look like shipping — support time, refunds, and the orders that never converted because the delivery estimate was vague.

Now rebuild contribution properly. Post-fulfilment contribution per order = net revenue − product cost − checking and packing − international and last-mile shipping − payment fees − expected fulfilment-failure cost − variable discounts and commissions.

And then the number that should actually drive the decision: contribution per session = conversion rate × post-fulfilment contribution per order − acquisition cost per session.

Contribution per session is the right metric because it is the only one that lets a faster, more expensive route win. A faster route lowers margin per order. It can also raise conversion and cut refunds, reships and support by more than it costs. Measured per order, it looks like a loss. Measured per session, it may be a clear gain. It may also not be — which is the point of measuring rather than assuming.

What actually counts as a product gap

The product gap is the half sellers skip, because it is harder to buy than a faster route. It is also the half that decides whether anything else works.

• Genuinely unavailable design, not a slightly different colourway of something on a local shelf.
• A specification or material difference the customer can verify and would pay for.
• Personalisation or made-to-order manufacture, which makes local immediacy irrelevant because no local alternative exists.
• A better problem-solution bundle — the complete set rather than the item plus three things the customer has to source themselves.
• A demonstrable functional improvement you can show rather than claim.
• A sufficiently better complete price, counted after delivery and tax, not before.
• Better proof and explanation for a narrow use case the mass-market listing does not serve.

A product is most exposed when it is visually generic, local substitutes are plentiful, local delivery is one or two days, the need is urgent, variant or compatibility risk is high, returns are difficult, and the delivery promise on your own product page is vague. Any two of those is a warning. Five is a rejection.

The specification half of this is a sourcing job, not a copywriting job. Turning a difference into something a supplier can quote and a warehouse can check is covered in how to write a product sourcing brief suppliers can actually quote.

Put the two gaps in a matrix

With a product-gap judgement and a measured delivery gap, the decision becomes readable.

• Strong product gap, small delivery gap: scale carefully, and protect the specification and supply consistency that created the advantage.
• Strong product gap, wide delivery gap: validate on direct parcel first, then test pre-stock or a hybrid model against contribution per session.
• Weak product gap, small delivery gap: test only where you have a measurable price, audience or conversion advantage.
• Weak product gap, wide delivery gap: usually reject. Advertising will not repair the offer, and faster shipping will only make the loss arrive sooner.

The bottom-right cell is where most failed products live, and it is the cell sellers most often try to fix with media spend.

The stock-transition gate

Moving a product from direct parcel toward China pre-stock, or from pre-stock toward local stock, is a cash decision disguised as a service decision. Four conditions should be sufficiently true before stock moves anywhere.

• Demand concentration: a small number of SKUs, variants and countries produce most of the orders.
• Demand stability: the pattern survives more than one campaign, one creative and one short trend.
• Service penalty: the current delivery measurably harms conversion, refunds, support or repeat purchase — measured, not assumed.
• Economic recovery: the expected service benefit exceeds storage, handling, inventory and obsolescence cost.

Do not use a universal order-count threshold. Anyone who gives you one has not asked what you are selling. A stable one-SKU replenishable product and a trend-led product with thirty variants carry completely different stock risk at the same order volume. The number of variants, the shelf life of the trend, and the share of orders sitting in the top few SKUs matter more than the order count.

The operating comparison between the three models — direct parcel, China pre-stock, and an overseas warehouse — including what each one does to cash, is in dropship, pre-stock or overseas warehouse, and the replenishment arithmetic is in a practical reorder-point calculator.

Why "just ship faster" and "just hold stock" both fail alone

Paying for the fastest route on every test is the most common wrong answer. It increases packed cost, destroys contribution, exposes you to dimensional-weight pricing, and still does nothing about supplier-processing delay, stockouts or wrong variants. Worst of all, it hides the real problem when the product was never differentiated enough to justify a wait in the first place.

Holding local stock is the second wrong answer applied too early. It locks cash into unproven variants, creates obsolescence, splits demand across countries, adds receiving, storage and fulfilment charges, introduces forecasting and replenishment risk, and strands slow-moving tail variants that nobody will buy at any price.

The correct sequence is unglamorous. Establish whether the product gap is real. Measure the current delivery penalty. Identify where the delay actually occurs. Compare operating models on contribution per session, including cash tied up. Then move stock only where demand is concentrated enough to justify it.

One thing worth doing immediately, before any of the above: state the honest delivery range on the product page and at checkout. It costs nothing, it is the single cheapest reduction in the trust penalty and the WISMO rate, and it converts a proportion of your refunds into orders that were never placed — which is a better outcome than it sounds.

A seven-day version of this audit

The full method is more work than most sellers will do in one sitting. A week is enough to get the decision.

• Day 1 — choose the real decision. Test or reject a product, scale or hold, change route, pre-stock, move stock locally, redesign a bundle, change price, or fix a defect source. Do not start with "find insights."
• Day 2 — build the local-substitute table for your top three countries.
• Day 3 — build the route distribution: dispatch and delivery P50/P90, tracking lag and promise hit rate, by SKU, country and route.
• Day 4 — reconstruct post-fulfilment contribution including refunds, reships, chargebacks, support and actual packed shipping.
• Day 5 — segment outcomes by delivery band, variant and country.
• Day 6 — model the three operating models on contribution per session, including cash tied up and obsolescence.
• Day 7 — make one bounded decision, and write down what would reverse it.

The output should be one sentence, not a report. Reject the generic product. Keep direct parcel because the product gap is strong. Pre-stock only the two dominant variants. Test a faster route in one country. Re-source to cut handling time. Improve the specification instead of paying for speed. Pause scaling until P90 stabilises.

Where this fits our side of the work

A useful China-side comparison needs more than a product screenshot and a factory unit price. Separating supplier processing time from route transit is what tells you whether a delay is a sourcing problem or a shipping problem. Quoting the customer-ready packed unit — including every accessory, insert and protective layer — is what makes a landed-cost comparison honest. Checking whether the exact variant is available from the factory or already sitting in stock is what decides whether pre-stock is even an option for this product.

Those are the comparisons we run when someone sends us a product to quote: exact product against merely similar alternative, factory processing against route transit, standard against faster route, direct parcel against pre-stock, and core local stock against long-tail direct parcel. What that costs and what the quote covers is set out in what fulfilment actually costs, and the sourcing-to-delivery workflow is in how it works. If you want the delivery gap on a specific product measured rather than estimated, send the product, the destination markets and the variants you actually sell.

The purpose is not to move every seller into a warehouse. It is to find out which constraint is actually binding — product differentiation, supplier availability, processing reliability, packaging, route, country fit, inventory placement, or the economics of the offer itself. Those have different fixes, and only one of them is shipping.

Evidence boundary

The retailer delivery figures cited here are company-reported statements from the Q2 2026 official releases and official-hosted earnings-call transcripts of Home Depot and Target, and describe those companies' own networks in the United States.

Where management associated faster delivery with higher conversion, that is their interpretation, not a randomised estimate; no company disclosed a delivery-time elasticity, and none of these figures transfers directly to an independent store.

The measurement method, the contribution definitions, the matrix and the stock-transition gate are RyanFulfil's operating framework, built from route quoting, dispatch, pre-stock and delivery-exception work, and they are decision models rather than formulas with a correct answer. No universal delivery threshold or order-count threshold for holding stock is implied or intended.

Route eligibility, customs treatment, tax and product-safety rules change and are destination-specific — check the current position for your product and market before acting.