Updated
Short answer: Don’t build a pricing model out of automation rules. Use the rules to decide which Rate Shopper strategy an order belongs to, and let Rate Shopper work out what it costs.
That is a division of labour, and it is the point of this page. ShipStation automation rules are good at classification — is this heavy, is it residential, is it going to zone 8, did the customer pay for two-day. They have no way to price anything: criteria compare a field on the order against a value you typed in advance. Rate Shopper is the reverse. It prices well and classifies not at all, because it only ever sees the one parcel you hand it.
Build a static rate model with rules and you are asking each half to do the other’s job.
The older method for this was to price a test data set across USPS zones and weights, note the winner for each combination, and encode those answers as rules. It works. What it produces is a snapshot: the rates as they stood on the day you priced them, applied to every order afterwards.
Rate Shopper has the property that method was trying to fake. It fetches live rates for the eligible services and applies the winner, and in the accounts I work in it resolves at label purchase rather than holding an answer from import — so the parcel gets priced as it stands when the label is bought, after any weight correction, dimension change or address edit. That evaluation point is not documented, so treat it as my observation and check it in your own account by editing a weight and watching what the selected service does.
That is the difference that matters. The staleness a static model has to be maintained against — annual rate changes, a surcharge you didn’t anticipate, an order somebody corrected after import — is not a problem Rate Shopper has.
One rule per shipping option you sell, keyed on the classification, applying the strategy.
Requested Service (Marketplace Value) is the criterion for what the customer picked at checkout. Residential / Commercial, Total Weight, and destination criteria are the other ones worth reaching for.ShipStation documents three ways to apply a Rate Shopper rule — from an automation rule, from the Service drop-down on a single order, or in bulk across a selection. The automation rule is the one that runs without anyone opening the order.
What the rules are deciding here is which strategy, not which service. A free-shipping order and a paid two-day order want different pools of eligible services, and the rule is how an order finds the right one. How Rate Shopper works and how to configure it well covers building those pools.
Two dependencies, and the rules engine will not warn you about either.
The rule that sets package dimensions must sit above the Rate Shopper rule. ShipStation’s setup documentation states this directly, in their words “to ensure that all orders generate the appropriate rate”. Rate Shopper takes the package as an input. An order that arrives at it carrying a default box gets rated on the wrong box, and the answer is confidently wrong rather than obviously wrong.
Anything keyed on a tag your rules apply must sit below the rule that applies it. Rules run top to bottom and later rules override earlier ones, so a classification pass above an action pass is the working shape.
Two cases. Rate Shopper is plan-gated and its availability has changed over time, so some accounts don’t have it. And some operations need a deterministic outcome — the same service on the same kind of order every time, for a promise made at checkout or for a warehouse process that depends on knowing the carrier before the label prints — which a live comparison by definition doesn’t give you.
If either applies, build the model. Compressed:
Assemble a test set. One address from each USPS zone — zones are how USPS expresses distance, priced across nine of them in Notice 123, and the Domestic Zone Chart gives you the zones relative to your warehouse. Then three or four weights: under 1 lb, 1 to 2 lb, over 5 lb. Cross them. If you already have shipments in ShipStation, export them with the zone field included and sample real addresses instead. Add a residential and a commercial address per zone if you want the model to hold up.
Price it. Create a manual store — Settings → Selling Channels → Store Setup → Connect a Store — so the test orders stay out of your real order flow. Enter one order per row of the test set, then open each and use the Rate Browser to see every connected provider’s rate. Record the winner.
Encode it. One rule per unique weight-and-zone combination, action Set Carrier/Service/Package. Most parcels under a pound will be USPS Ground Advantage unless you have a consolidation service connected, so that is usually one rule rather than nine. Where a carrier wins across several zones at several weights, consolidate.
Then re-price the test set when rates move, which for most carriers is at least once a year. The model is only as current as the day you built it — which is the whole reason to prefer Rate Shopper when you can have it.
Operations where the box has to be worked out before the rate means anything — where packing and rating are one decision rather than two in sequence — need logic beyond what either tool expresses, and we build that alongside ShipStation.
Knowing how a carrier prices a package is useful. The harder problem is making the right packing and rate decision on every order, when the correct answer changes with the contents, the destination and the rate card.
String builds and maintains that logic inside existing ShipStation accounts, for operations shipping roughly 6,000 orders a month or more. Below that, ShipStation's native automation and accurate product dimensions will get you most of the way there.
Eight patterns that show up when an operation needs custom shipping logic