What does it realistically cost to have an agency build a custom Shopify Function for one discount use case?

Is there a market rate for an agency-built discount Function?

No. There is no market rate to report for one custom Shopify Function, the code Shopify runs at checkout to apply your discount rule. In our judgment, any article that gives you one invented it or averaged a handful of anecdotes. Whether an agency or an in-house developer should build it is the staffing decision that comes first; the price question starts once an agency is on your shortlist.

The Shopify Partner Directory is the closest thing to a price list, and it shows why no rate exists. Partners can display a self-set starting price, many write "Contact for pricing" instead, and buyers can filter by price range. On the first page of the directory's custom apps and integrations listing, the starting prices shown run from $75 to $3,000 (Shopify Partner Directory, October 2026). Each figure is a partner's own starting price for its services, so it is not a price for your rule. Shopify's own instruction is to "communicate directly with the partner, set your project terms" (shopify.com/partners, October 2026).

What do Deploi's decision records estimate for one discount Function?

Deploi's Build vs. Buy record on custom discount types puts an illustrative $8,000 to $25,000 on one new production discount Function (October 2026). Its reference scenario is a Shopify Plus store selling $20 million to $100 million a year, with a rule no packaged app expresses. Two other records price the cases around it: porting a rule that already ran as a Shopify Script, and the alerting a live Function needs. Shopify Scripts, the older checkout code, stopped running on 30 June 2026 (Shopify Help Center, October 2026).

The Scripts-to-Functions record adds that a one-rule store overpays at $10,000 (illustrative), so one ported rule sits at the bottom of its band.

Which parts of the discount rule move the price most?

Three facts about the rule set most of the price: the data it reads, its collisions with other discounts, and whether its numbers are hardcoded or configurable.

The data it reads. The rule runs as a Shopify Function, custom code that Shopify executes at checkout. A rule that reads customer tags, product tags, a metafield (custom data) table, the market and the delivery method has five inputs, any of which can be missing. Shopify caps the query that fetches the rule's data at 3,000 bytes excluding comments (shopify.dev, October 2026), so scoping works inside a hard ceiling.

Collisions with other discounts. In our judgment, stacking, precedence and mutual exclusion are where the build time goes. A one-sentence rule often needs four more sentences on how it collides with other discounts. A brief that settles those collisions gets a tighter quote.

Hardcoded or configurable numbers. Tiers baked into the code are cheaper to build and more expensive forever, because every change ships as a new app version (Shopify app versions guide, October 2026). Tiers stored in metafields cost more once and turn each later change into a merchandising edit. Merchants make that edit in a settings screen the agency builds with Shopify's function settings component (Shopify developer docs, October 2026). Deploi's decision record on Function settings judges that hardcoding stops being cheap at about 8 promotion changes a year (October 2026).

Which technical limits and add-ons push the price further?

Two more drivers sit outside the discount rule: the largest cart the Shopify Function must survive, and the alerting that ships with it.

Cart size and language. Shopify gives each Function run 11 million instructions, 128 kB of input and 20 kB of output for carts up to 200 line items, and scales those limits proportionally above 200 (shopify.dev, October 2026). Shopify recommends Rust for large carts, because JavaScript runs into instruction limits sooner (Shopify developer docs, October 2026). A rule that must survive a 200-line wholesale cart is a different engineering problem from one that runs on a three-line direct-to-consumer cart.

Alerting. A Function with no alerting is cheaper and is not finished. Shopify keeps each run's logs in the Dev Dashboard (its developer console) and truncates them after 1 kB (Shopify monitoring guide, October 2026). The same guide describes no merchant-facing alert, so somebody has to build one and price it. Ask every quote whether it includes alerting; in our experience, most do not. Deploi's Build vs. Buy record on Function observability scores this layer CUSTOMIZE: keep Shopify's run logs and build the alert. Its illustrative $12,000 to $40,000 in year zero covers two or three live Functions, so alerting for one discount Function costs less (October 2026).

What should an agency quote for one discount Function itemize?

Deploi's checklist: an agency quote for one discount Function, the custom code that applies your discount at checkout, should itemize seven lines, so you can see where each agency put the work and the risk:

1. Discovery and a written rule specification, signed off before the build.
2. The Function itself, with the language named and the reason for it.
3. The configuration surface: hardcoded numbers, or settings your team edits in the Shopify admin.
4. Acceptance testing against named cart scenarios, including collisions with other discounts.
5. A staged rollout plan.
6. Logging, alerting and a runbook (the written steps for when the rule fails), or an explicit note that they are excluded.
7. Who owns changes after launch, and at what rate.

A quote that is one line and one number is not cheaper. It is the same work with the risk moved onto you.

When should you not commission a custom discount Function?