Somewhere in the last few years, the line “Add shipping protection for $2.59” became as normal in an online cart as the shipping estimate above it. Most shoppers click it without thinking. Most merchants installed it without thinking much harder. And most of the disagreements about whether it is worth it come from people who mean different things by the same three words.
So let us take it apart. Who pays. Who holds the money. Who decides the claim. Where the refund lives. And what changes when the deciding is done by software instead of a person.
- The three business models behind every protection app, and what each one costs you
- Why “protection” is not “insurance”, and why the wording on your policy page matters
- What a claim really costs, worked through on a $64 order
- The six settings that decide whether protection saves hours or creates them
The line in the cart
Package protection is an optional product. The shopper adds it at checkout for a small percentage of the order, typically between one and five percent with a minimum of a dollar or so. If the parcel is lost, damaged, stolen, incomplete or wrong, a claim pays for a refund, a replacement or store credit. That is the whole idea. It has been the whole idea since the first version of it appeared as a checkbox in 2018.
Two things make the offer work or fail. The first is presentation: a clearly priced, one-click line under the store’s own name gets added far more often than a third-party badge with a logo the shopper has never seen. The second is honesty: pre-checked boxes and hidden fees cost conversions, and in a growing number of places they cost fines.
Three models, one checkbox
Behind the same checkbox sit three completely different businesses. Which one you are buying decides most of what follows.
Vendor-held
The app collects the fee, decides the claims and pays the outcomes. The merchant pays nothing and keeps nothing. The shopper files with the vendor, hears from the vendor, and gets a refund or reorder from the vendor. This is the model most people picture, because the biggest names in the category run it. It is also the model where you have the least say in what happens to your customer.
Self-insured
The app sells the protection under your name and passes the entire fee to you. You fund every claim and work every claim, by hand, in a portal. The maths looks wonderful in a spreadsheet: you keep 100% of the fee. The maths looks less wonderful at 5pm on a Friday when there are forty claims in the queue and one person to read them.
Shared
The fee funds a claims pool. The merchant keeps a plan-defined share. The app provides the claim page, the evidence rules, the review and the automation. Claims are decided inside rules the merchant sets, and the outcomes are created in the merchant’s own store. This is BuySure’s model, and it exists because the other two each leave something important on the table: control in the first case, hours in the second.
Ask any app you evaluate one question: when a claim is approved, where does the refund object live? In your Shopify admin, or in theirs?
Why it is not insurance
Insurance is a regulated contract between an insurer and a policyholder, with underwriting, state licensing and a claims adjuster. Package protection is a product a store sells and a promise the store keeps. Every serious app in the category says so plainly, and it matters beyond the legal page. It changes the words you should use: “protection”, “eligible claim”, “resolution”. Not “coverage”, “insured”, “guaranteed”. A policy page that borrows insurance language invites a customer to hold you to an insurance standard.
What a claim actually costs
Take a $64 skincare order. The shopper added protection at 4%, so $2.56. The parcel arrives with a cracked dropper. Here is what each outcome costs the store, roughly, once the protection pool has covered the claim:
- Replacement: cost of goods (say $19) plus shipping ($7). The customer gets the serum they wanted. Cost to the store: $26, funded from the pool.
- Store credit with a 10% bonus: $70.40 of credit, spent back in the store at retail margin. Real cost when redeemed: roughly the cost of goods on what they buy next, plus the bonus. And the customer is back.
- Refund: $64 out. The relationship, for this order, is over.
Now the part nobody prices: the twelve minutes a support agent spends reading the email, asking for a photo, finding the order, deciding, and issuing the refund. At sixty claims a month that is twelve hours. At six hundred, it is a full-time job. That is where the model matters more than the fee.
How a claim gets decided
Three questions decide whether an app saves you time or costs it.
Where is the claim filed? On your domain, under your brand, or on the vendor’s site? A claim page that validates the order before it asks a single question removes half the back-and-forth.
What evidence is required? A claim without a photo is a conversation. A claim with a photo is a record. BuySure requires at least one attachment, and the AI reads it.
Who decides? A vendor’s agents, your agents, or software inside a limit you set? BuySure AI reads the evidence, checks the delivery data, matches the order and tracking, weighs the customer’s history and scores the risk from 0 to 100. Low-risk claims under your auto-approve limit resolve on their own. The rest go to a reviewer or your team with the score and a recommendation. Rejections always come from a person.
The six settings
Every store that runs protection well has made six decisions on purpose. Every store that runs it badly has let them default.
- The name. “Northline Delivery Promise”, not “Shipping Insurance by SomeVendor”. It is your promise.
- The price. A percentage of the cart, shown before the shopper chooses, with a sensible minimum.
- The claim window. Fifteen to ninety days after the order. Thirty for domestic ground, longer for international.
- The waiting period. Two to five days after a delivered scan before a stolen claim can open. Long enough to let pre-scanned parcels arrive, short enough to be fair.
- The auto-approve limit. The value under which a low-risk claim can resolve without a person. Start low. Raise it once you have read a month of decisions.
- The outcomes. Replacement by default when stock allows, store credit with a bonus as the better deal, refund when it is right.
What good looks like from the outside
A customer whose serum arrived broken opens the store’s tracking page, taps “something wrong”, photographs the dropper, picks “replacement”, and gets a reference. Six minutes later a replacement order exists in Shopify and an email says so. Nobody at the store read anything. The claim sits on a timeline with its score and its reasons, in case anyone ever needs to.
That is the standard. Anything that involves the customer explaining the problem twice, or an agent hunting for an order number, or a refund that finance cannot find in the admin, is the old way with a new checkbox on it.
Questions
Does package protection increase cart abandonment?
Not when it is opt-in, clearly priced and easy to skip. Pre-checked or hidden fees are what cost conversions, and in some jurisdictions they are illegal.
Can I self-fund claims without an app?
You can add a product to the cart yourself. What you will not have is the claim page, the evidence requirement, the risk scoring, the timeline and the Shopify-native outcomes. Those are where the hours go.
Is package protection worth it for a small store?
If someone on your team owns “the parcel problem” and it costs them more than a few hours a month, yes. Below that, a clear policy and a good tracking page may be enough for now.
What happens to the protection fee if nobody claims?
In a vendor-held model the vendor keeps it. In a self-insured model you keep it. In BuySure’s model it stays in the claims pool and the merchant keeps the plan-defined share.
