Split Shipments: Managing Customer Expectations Across the Order Journey

WISMOlabs Research Series: Context-Aware Post-Purchase Experience

Let customers know items may arrive separately, then make it clear what has arrived and what is still to come.

Split-shipment communication can start with a simple notice: “Items in your order may arrive separately.” Once a split is confirmed, tell the customer that the order will arrive in separate packages and make the remaining deliveries easy to follow.

The customer-facing message should answer: Has my whole order arrived, and if not, what is still to come? A delivered scan on one package should not turn into “your order has arrived” while another item is outstanding. Customers should be able to tell that the rest is coming without having to ask “where is the rest of my order?”

WISMOlabs examined shipment records from selected retailers with recorded multi-package/split shipment orders. In the summer 2026 cohort, carrier mix and the timing of package records showed recurring relationships with deliveries spread across different days.

The practical implication is to keep customer informed as packages move separately, leading with a short order summary, then offering tracking details and reliable arrival estimates when available. The research measures recorded shipment and delivery patterns; the guidance before purchase and during fulfillment is a best practice operating recommendation, not a measured checkout or conversion result.

What is a split shipment, and when is a delivery partial?

A split shipment is one order fulfilled through multiple packages. A partial delivery is the period when some of that order has arrived and an expected item or package remains outstanding.

The distinction matters. Several packages can arrive on the same day. Conversely, an order with multiple packages can leave a customer waiting days between arrivals. Counting packages tells you how many parts the experience must explain; it does not tell you how long the order will remain partially delivered.

In this research, multiple recorded carrier and tracking – number combinations identify multi-package order groups. That is a useful tracking-data measure, but it does verify the physical contents and/or a reason for each split.

Explain the customer impact, not the packing decision

Two common operational reasons for splitting an order are cost and logistics: keeping packages within size or weight limits to avoid extra charges, and shipping items from different locations. An order can involve both. These are operational examples, not measured causes of the splits in this study.

Those decisions can stay behind the scenes. The useful message is that items will arrive separately and what that means for the customer. Explain the reason only when it helps them make a decision or take an action. Relevant charges, delays or changes to the promised experience still need to be clear.

Start with a simple notice and add details when they are known

If separate shipments are possible, a short notice before purchase or in order confirmation can set the expectation: “Items in your order may arrive separately”.

When fulfillment confirms the split, change “may arrive separately” to “will arrive separately.” Add tracking as it becomes available. If the customer needs to know about a delay or take action, explain that change. Otherwise, keep routine packing decisions out of the message.

Journey stageCustomer questionCommunication priority
Before purchase or in order confirmationMight items arrive separately?A short “may arrive separately” notice where relevant.
Once the split is confirmedWhat should I expect?Say the order will arrive in separate packages.
Shipment and transitWhere is my order?Link available tracking and show whether the rest is on the way or still being prepared, when verified.
Partial delivery or an exceptionHas everything arrived? Do I need to act?Say “portion of your order has arrived,” show what remains and highlight any required action.
Final delivery or another agreed resolutionIs my order complete?Confirm completion only when everything is delivered or otherwise resolved.

WISMOlabs journey framework: proposed communication practices.

These are opportunities to keep the customer informed. Use existing confirmation and shipping messages where they do the job. Keep the order summary consistent across tracking, notifications and support.

Which fulfillment patterns were associated with arrivals on different days?

WISMOlabs compared split orders within the same retailer, then took the median of the eligible retailer differences. This gives each eligible retailer equal weight in the summary rather than allowing the largest shipment volume to determine the result.

The outcome was the share of split orders whose retained packages had delivery dates on more than one calendar day. Timing comparisons required a recorded delivery milestone for every retained package.

Figure 01 summarizes the observed relationships and their limits.

Figure 01 · Observed associations

Carrier mix and record timing showed recurring associations

Median within-retailer difference in multi-day arrival share. Scale: 0–60 percentage points (pp).


Package-count counterexample: three or more packages versus two showed no consistent direction across eligible retailer comparisons. This is not a measured zero effect.

View chart data and definitions
Retained split orders with a delivery milestone on every retained package
Within-retailer comparisonMedian difference in multi-day arrival shareOperating use
Multiple recorded carriers versus one+26.7 percentage pointsKeep each package’s expected arrival visible.
Different package-record creation dates versus the same date+40.9 percentage pointsKeep the rest of the order visible before all tracking is available; do not infer dispatch.
Three or more packages versus twoNo consistent directionUse package count for layout, not as a timing-risk measure on its own.
Source: WISMOlabs retained shipment and delivery records, summer 2026. Each eligible retailer is weighted equally; eligible sets differ by comparison. No industry prevalence or causal estimate.

Multiple carriers call for package-specific expectations

The carrier comparison supports a clear design priority: an order summary should preserve each package’s delivery expectation. A single headline date can become misleading if it describes only the first or last package.

The positive relationship remained when the analysis was restricted to two-package orders. It therefore was not explained solely by comparing orders with different package counts. That still does not establish a carrier performance ranking or a reason to consolidate carriers.

For example, a hypothetical order update could say:

Part of your order has arrived. Your jacket was delivered today. Your shoes are traveling separately and are currently expected Friday. Track your order below.

Use the actual item assignments and current delivery evidence. If the remaining package has no reliable expected date, state that plainly instead of supplying a date from other packages in the same order.

Keep the rest of the order visible before all tracking is available

Check whether more items remain before treating the visible package as the whole order. That check belongs in the retailer’s systems; customers do not need to see an explanation of incomplete package records.

The relationship stayed positive in the sensitivity checks, but its size changed when orders with conflicting shipping and delivery timestamps were excluded. Use this as a reason to inspect data readiness, not as a calibrated prediction of delivery timing.

For example, when the retailer has verified that the remaining items are still being prepared:

Part of your order is on the way. The rest is still being prepared and will ship soon.

No final package count is needed for that message. Use “Package 1 of 2” only when the total is confirmed. If the remaining items’ status is unknown, say they are awaiting an update rather than assuming they have shipped.

More packages require a usable layout, not an automatic risk label

The package-count comparison did not support a universal “more packages means more staggered arrivals” rule.

For a long package list, summarize progress first, keep outstanding packages easy to find, and collapse completed details when appropriate. For a long wait between deliveries, emphasize the remaining items, current expectation and available action. An order can need either or both.

Use product category, order value and a customer context to refine the response

Industry labels are a starting point for questions, not a substitute for the order’s contents. The current data does not support a reliable vertical benchmark or a comparison of split shipments by order value.

Retailers can use their own item, value and fulfillment data to decide when extra detail or support is useful. This is internal prioritization, not a requirement to display every field to the customer.

Order contextCustomer question to answerUseful information, when available
Main product and required accessory arrive separatelyCan I use what has arrived, or am I waiting for an essential part?Item-to-package mapping and verified product dependencies.
Several apparel items arrive separatelyWhich pieces arrived, and where is the rest?Item names, variants and quantities in each package.
Custom products or an event-related orderWill the remaining items arrive in time for their intended use?Production status, current delivery expectations and any customer-provided deadline.
Time-sensitive or perishable contentsWhich package needs attention now?Package contents and verified handling or receipt instructions.
Higher-value items remain outstandingWhich part of the purchase is still undelivered?Valid order and item values alongside package status.

For financial prioritization, distinguish two measures:

Revenue share associated with split orders is the value of eligible split orders divided by the value of all eligible orders in the same period. Count each order’s value, even if it has several shipments. This measures the commercial share of orders receiving the split-shipment experience; it does not measure lost or at-risk revenue.

Outstanding value within an order is the value of items still awaiting delivery or another agreed resolution divided by that order’s value, using a consistent valuation policy. Recalculate it as items arrive or are resolved. Do not assign value equally to packages unless their contents and values justify that allocation.

Value also misses some practical dependencies. An inexpensive accessory can be essential to using a more expensive product. Combine value with item importance, time sensitivity and unresolved exceptions when deciding what information to emphasize. Keep basic clarity and access to help available to every customer.

Show order summary above the package details

A useful order summary should show what arrived and what remains. Put available tracking underneath, with arrival estimates and actions where useful. Item names or images can help when package contents are reliably known, but a basic separate-package notice should not depend on complete item mapping.

Order situationSuggested headlineDetail that must remain visible
Some items have shipped; others are being preparedPart of your order is on the wayAvailable tracking and a short note that the rest is still being prepared.
A package arrived; another is outstandingPart of your order has arrivedDelivered contents and the status of everything still outstanding.
One package has an exceptionOne package needs attentionAffected contents, the required action and the status of unaffected packages.
All expected items were deliveredYour order has been deliveredDelivery details for every expected package.

WISMOlabs operating framework. These recommendations have not been evaluated for support or revenue impact in this study.

If item-to-package mapping is unavailable, do not guess which item arrived or ask customers to interpret a data gap. Show verified package statuses and make clear that the order is not yet complete when items remain. Keep a route to help visible for items that appear genuinely missing.

Define completion before sending a final message

The checks behind a simple completion message belong in the retailer’s systems. The authoritative order record should account for every expected item and quantity. An order can close only when nothing remains awaiting fulfillment, delivery, pickup or another agreed resolution.

A replacement label does not resolve the customer’s order. Link the replacement to the original package, suppress misleading duplicate progress, and keep the replacement outstanding until it reaches the required resolution. Similarly, a failed attempt, return-to-sender event or refund request is not automatically a completed customer resolution.

Use “your order has been delivered” only when all expected items were delivered. If some items were cancelled, refunded or collected, use language that accurately describes those outcomes.

Do not infer completion merely because every currently known tracking number is terminal. The fulfillment system must also confirm that no expected items or packages remain unaccounted for.

Figure 02 shows the checks behind that completion rule.

Figure 02 · Completion framework

Figure 02 · Completion framework

A delivered package does not close the whole order

Proposed internal checks behind a simple customer message.

Account for every item

Check the order record for anything still awaiting shipment or delivery. Customers do not need the packing plan.

  • Expected items and quantities
  • Anything still to ship
  • Anything still to arrive

Resolve what remains

A replacement stays outstanding until resolved. A failed attempt or return event alone is not customer resolution.

  • Packages awaiting delivery
  • Replacements and pickup
  • Approved cancellations or refunds

Choose accurate wording

Use “delivered” only when all expected items were delivered; otherwise describe the actual resolution.

  • Outstanding: “Part has arrived”
  • All delivered: “Order delivered”
  • Mixed resolved outcomes: “Order complete”
Completion ruleEvery expected item must be accounted for, and nothing may remain awaiting fulfillment, delivery or another agreed resolution.
Framework: WISMOlabs whole-order completion model. Proposed operating logic, not a measured customer or financial result.

Coordinate notifications across the order

When sending a package update, include enough order context to prevent confusion. After a delivery, say whether the rest of the order is still outstanding. After an exception, identify the affected package without implying that the whole order is delayed. Just like orders shipped in a single package a split does not require an email for every internal fulfillment or carrier event.

Coordinate this across WISMOlabs, email platforms and any retailer-managed messages. Choose an owner for each message type, deduplicate repeated events, and use the same order-completion rule in every channel. Where nearby routine events are combined, preserve urgent exceptions and customer actions.

A hypothetical exception update might read:

The package containing [items] needs your attention: [verified issue and action]. Your other package is [verified status]. View the whole order for the latest details.

Only promise another update when the configured workflow will provide it.

Measure the business result through missing-item contacts and resolution

The intended benefit is fewer avoidable “where is the rest of my order?” contacts and fewer negative reviews that mistake a separate delivery for an incomplete order. A simple split-shipment notice may help customers make that distinction.

Test the proposed experience against the current experience using:

Related missing-item contacts during the partial-delivery window per 1,000 eligible split orders.

Classify the contact reasons separately: another package still traveling, an item awaiting fulfillment, an item absent from a delivered package, and a delivered package the customer cannot locate. These situations require different responses.

Alongside contact volume, measure repeat contacts, time to documented resolution, unresolved orders, premature whole-order delivery messages and notification complaints.

As a secondary measure, classify reviews mentioning an incomplete order. Separate confusion about a pending delivery from verified missing items; a complaint is not automatically a misunderstanding. Because only some customers leave a review you should track review coverage.

Where feasible, assign the experience at order level before the first partial delivery and keep it consistent across the tracking pages and messages. Use a common follow-up period. Retain unresolved orders in the evaluation and identify open cases separately in resolution-time reporting.

If a randomized test is unavailable, use a concurrent comparison group and report changes in carrier mix, package count, arrival timing and order value where available. Those changes can affect contact demand independently of the communication.

Translate a validated reduction into handling value only after the outcome is measured:

avoided related contacts × fully loaded handling cost per contact

Subtract implementation and ongoing costs. Distinguish support capacity released from cash expenditure actually reduced. Revenue, loyalty and repeat purchase require their own measured outcomes.

Figure 03 separates the shipment evidence available today from the outcomes needed to establish business value.

Figure 03 · Measurement framework

Measure customer outcomes before claiming savings

Shipment records establish delivery patterns. Business impact needs linked outcome data.

Shipment evidence

Available in this study: retained package records and delivery milestones.

  • Carrier mix
  • Record-creation timing
  • Multi-day arrivals

Customer outcomes

Not measured here. Test contacts, resolution and reviews that confuse separate deliveries with missing items. Retain unresolved cases.

  • Related missing-item contacts
  • Repeat contacts
  • Time to documented resolution

Handling value

Not measured here. Apply verified contact reduction to handling cost, then subtract implementation and ongoing costs.

  • Contacts avoided × handling cost
  • Capacity is not cash savings
  • Revenue and loyalty need separate tests
Framework: WISMOlabs split-shipment outcome measurement plan.

Study scope and methods

WISMOlabs analyzed retained records from selected retailers confirmed to submit shipments through its API and with recorded multi-package orders. Orders entered the summer 2026 study cohort by their earliest retained shipment-creation date; all retained sibling records were included through the study snapshot.

An order group combines retailer and nonempty order reference. A recorded package is a distinct carrier and normalized tracking-number combination. Retailers needed at least 1,000 retained order groups and 100 split groups for the base comparison set. Each factor comparison required at least 100 split orders with complete retained delivery milestones in each group within a retailer.

Results are the unweighted median of within-retailer percentage-point differences. The eligible retailer set can differ by factor. Delivery milestones were complete for 97.7% of split orders in the base comparison set; orders missing a retained package’s delivery milestone were excluded from timing results. Those results do not describe unresolved or undelivered orders, which may have a different experience.

Retailer identities are withheld for confidentiality.

Questions retailers ask about split-shipment communication

When should retailers tell customers that an order may ship separately?

Use a short notice before purchase or in order confirmation when separate shipments are a relevant possibility. Retailers do not need to know the eventual packaging plan. Once a split is confirmed, say the order will arrive separately and add tracking details as they become available.

Do customers need to know why the order ships in separate boxes?

Usually, the communication can stop at how the split affects delivery. Warehouse locations, size limits and packing economics need not appear in routine updates. Add a reason only when it helps the customer understand a relevant change or decide what to do.

Should a partial delivery be marked “delivered”?

The delivered package can be marked delivered. The order should remain partially delivered while expected items or packages are outstanding. Show both states clearly.

Should every package trigger a separate email?

Use meaningful package events to decide whether an update is useful, then include the whole-order context. Coordinate senders and suppress duplicate information. This study does not determine an ideal message count.

Does using one carrier prevent partial deliveries?

No. The study found an association between multiple recorded carriers and multi-day arrivals within eligible retailer comparisons. It did not establish that using one carrier prevents staggered deliveries or improves carrier performance.

How should retailers track the value of split orders?

Calculate the value of split orders as a share of all eligible order value, counting each order once. Separately calculate the value still outstanding within each order. Both need reliable transaction and item data; neither is a measure of lost revenue.

How can retailers tell whether clearer communication saves support work?

Compare related contacts and documented resolution outcomes against a suitable baseline or control group. Include unresolved orders and access-to-help guardrails. Shipment milestones and message handoffs alone cannot establish a saving.

Keep the message simple from purchase to the last delivery

Start with awareness that items may arrive separately. Once the split is confirmed, make that clear. When a package arrives, say whether more is still to come. The customer does not need the packing plan or the reason behind it to follow the order.

Keep internal completion checks thorough and customer updates brief. Test whether this reduces confusion, related contacts and misunderstood partial-delivery reviews, while keeping help easy to reach for real problems.

Explore WISMOlabs’ shipment tracking software for connecting order context, carrier events and customer communication. For a broader evaluation checklist, see the shipment tracking software guide.

About Author
Picture of WISMOlabs Team
WISMOlabs Team
We help eCommerce and subscription brands grow revenue, boost repeat sales, cut support costs, and build loyalty through smart post-purchase technology.
Split shipment research - managing customer expectations
Split Shipments: Managing Customer Expectations Across the Order Journey

Split shipments can confuse customers when one package arrives before the rest. WISMOlabs research examines how carrier mix and shipment-record timing are associated with staggered deliveries – and why clear, order-level communication matters. Learn what to tell customers before fulfillment, during transit and after a partial delivery, without exposing unnecessary operational detail.

When ETA is not a promise
When an ETA Is Not a Promise: A Retailer Playbook for Confidence-Aware Delivery Messaging

WISMOlabs analyzed anonymized operational data from more than 2.5 million shipments across 140 retailer entities to examine carrier ETA availability and accuracy. The study explains why an ETA is evidence, not automatically a retailer promise, and provides transparent methodology, benchmarks, limitations, and a practical framework for confidence-aware post-purchase messaging for ecommerce retailers.