How every cost is sourced, dated, and computed

BestPrintOnDemand exists to answer one question honestly: what does it actually cost to fulfill a given product through each print-on-demand supplier, all-in. To keep that answer trustworthy, every figure on the site is tied to a named source, carries a confidence tier, and is stamped with when it was retrieved. This document is the full account of how we source, decide, compute, and refresh, so that no number ever has to be taken on faith.

We rank sources in a strict hierarchy and always prefer the highest available tier. First is a supplier's own pricing API: a first-party endpoint that returns the exact base cost for a specific variant (Printful, Printify, Gelato, Prodigi, and Gooten each expose one). Second, when no API figure is available, is the supplier's own public pricing page, read directly from the product listing. Third, and only when the first two are gated or silent, is a reputable third-party reference (a trade publication, a cost-analysis site such as PodVector or SynCost, or a detailed platform review), which we always cross-check against the supplier before we display it. We never let a lower tier override a higher one, and we never invent a figure to fill a gap: if no real source states a number, we return null and mark it unverified rather than guess.

Printful, the largest supplier, exposes its base costs through a PUBLIC catalog API that needs no key at all (GET api.printful.com/products/{id} returns every variant's base fulfilment price), so Printful base costs are pulled live and refreshed automatically in every deployment, with zero setup. The other suppliers gate pricing behind a free-account key or token (Printify uses a bearer token, Gelato and Prodigi an API-key header, Gooten a RecipeID). That is account-gating, not merchant-gating: no paid plan, sales channel, or store connection is required, just a free account. For those suppliers, until the operator connects a key we do not fabricate live prices: the figure falls back to the cited, code-owned catalog (each still linked to the supplier page it came from), and the operator back-office shows which connections are live and which are awaiting a key.

We decide real-time versus scheduled per supplier, based on how each API is shaped and how fast the underlying number actually moves. Base blank costs change rarely, so we fetch them on a schedule (daily or weekly is ample) and store them, rather than hitting a supplier on every page view. Two suppliers price through a live quote rather than a static table (Prodigi via POST /quotes, Gooten via POST /price); we snapshot those on the same schedule for the comparison table, and can additionally call them live for an exact shipping quote at the moment a reader needs one. Volatile, request-specific values (live shipping rates for a specific destination, current stock availability) are fetched on demand or driven by supplier webhooks where offered, never cached as if they were stable. In short: slow-moving costs are scheduled and stored, fast-moving values are pulled live.

Every figure carries one of three confidence tiers, shown the same way on every surface. Confirmed means we read it from the supplier's own API or pricing page. Reported means it comes from a reputable third party because the supplier gates the figure, and we could not confirm it first-party. Unverified means no real page states it, so we show nothing rather than a guess. Color is never used alone to signal a tier; a text label always accompanies it. Alongside the tier, each live figure carries its own fetched_at timestamp and, where the supplier states one, a source_event_at, so a reader can always see how fresh a number is and when the supplier last changed it.

We keep two clocks and never let one stand in for the other. The editorial catalog (product and supplier identity, print methods, the cited base costs, this copy) is code-owned and changes only when the code is edited; it carries a last-verified date and is never presented next to a live number as though it were live. Live data carries its own retrieval timestamp. Ingestion is idempotent and keyed on the exact variant, so re-pulling a price updates the same record rather than duplicating it, and we retain a history row on every change so a product's cost trajectory stays auditable. A figure whose fetched_at has aged past its freshness window is flagged as stale, not shown as silently current.

The site earns revenue through affiliate relationships with some suppliers, and we are explicit about how that works so it never distorts the underlying data. When a reader clicks through to a supplier and signs up or orders, the supplier pays us a referral commission under that program's published terms (the affiliate-network sources in the registry link each program). Estimated affiliate earnings shown in the back-office are exactly that: estimates, derived from published commission rates and benchmark conversion figures, clearly labeled as estimates rather than booked revenue. That relationship also earns an active partner a bounded, disclosed boost in the default Best value ranking, so a genuinely competitive partner can sit just above a slightly cheaper alternative. The boost is bounded (landed cost is always the largest factor, so a materially cheaper option still wins), every partner is labeled 'Partner', and it never changes a base cost or a fact. A 'Cheapest' sort always ranks by price alone, so the lowest verified cost is one click away, affiliate or not.

Refresh runs on a schedule and every run is recorded. A bearer-protected refresh route (called by the platform scheduler) invokes each connected supplier's fetcher, writes any new figures through a guarded ingest path, and logs a feed run with per-supplier status: connected, degraded (for example, a supplier whose key is not set or whose endpoint failed), the number of rows written, and any error. The operator can see, at a glance, which feeds ran, when, and whether each supplier is live. Base costs, shipping, and margins change and vary by size, color, print method, and region, so the site is an informational cost-comparison tool, not a print supplier and not financial or business advice. We always tell the reader to confirm the final price on the supplier's own page before listing a product. That combination (a clear source hierarchy, honest handling of gated prices, tiered confidence, dated retrieval, and disclosed affiliate economics) is what makes the comparison trustworthy.