Pinterest Shopping Bids Work Better as a Mix Than a Setting

Do not choose one Pinterest Shopping bid strategy and leave it there. Use oCPC to learn whether qualified people will enter the site, oCPM to buy purchase volume at a controlled cost, and ROAS bidding only when the account is returning reliable checkout values. In Pinterest’s own Smart Bidding guidance, the company says a combined oCPC + oCPM approach produced 55.3% higher median ROAS than oCPM alone. That is platform-reported data, not a promise of a 55.3% lift for every store. It is still a useful operating signal: Pinterest Shopping is a portfolio problem, not a dropdown-menu problem.

That distinction matters because the usual launch sequence is backwards. A brand connects a catalog, turns on Performance+, selects a conversion objective, then judges Pinterest after a short run of thin data. When revenue disappoints, the team changes bids, audience, creative, and budget all at once. The account has not failed Pinterest. It has made a result impossible to diagnose.

At Mintec, our rule is simple: a bid should have one job and a known graduation condition. A click-focused campaign proves whether the catalog and landing page can attract the right kind of visitor. A conversion campaign proves whether that traffic can purchase at an acceptable cost. A value-based campaign earns the right to scale only after the purchase values it receives are trustworthy. Mixing those jobs inside one opaque campaign creates reports, not learning.

This is the bid architecture we use to turn Pinterest Shopping from a one-shot test into a controlled ecommerce channel.

The three bidding modes are answers to different questions

Pinterest presents three primary bid strategies under the Catalog Sales objective: ROAS, Conversions (oCPM), and optimized cost per click (oCPC). They are not a ladder where every account must start at oCPC and “graduate” to ROAS. They solve different business problems.

The practical error is treating low CPC as a success metric. A low-cost click campaign can be exactly what an account needs in the first stage, but it cannot settle the question that pays the bills: whether those clicks become margin-positive orders.

Start with signal readiness, not a performance target

Before allocating a dollar to the bid mix, inspect the inputs that bidding will receive. The June 2026 Pinterest Shopping performance checklist is unusually clear on this point: catalog quality, tracking, product-group depth, campaign continuity, and the post-click retail journey all precede bid changes.

Use this five-minute preflight:

1. Purchase value is real. Check a sample against actual orders; wrong currency, discount logic, shipping, or duplicate events corrupts ROAS optimization.
2. Tag and Conversions API are both active. Pinterest reports a 9.0% average CPA improvement and 23.7% more attributed conversions versus tag-only setups. That is directional vendor data, not a guarantee; incomplete measurement is still a blocker.
3. Product IDs reconcile. AddToCart, Checkout, and Purchase IDs must match the catalog.
4. Groups have depth. Pinterest recommends about 500 SKUs per ad group; do not divide a small catalog into statistically empty segments.
5. The promise survives the click. Price, availability, image, and promotion must match from Pin to checkout.

If two or more checks fail, do not call the campaign a bidding test. Call it implementation work and finish it first.

The Mintec three-lane bid mix

The following is a starting architecture for a catalog that already has clean measurement. It is not a universal budget recipe. The allocation should move with conversion volume, inventory margin, and seasonality—but each lane keeps a job that can be evaluated independently.

Lane 1: discovery and diagnostic traffic — oCPC

Allocate 15–25% to a separate oCPC campaign with one coherent product cluster, landing-page type, and a few genuinely distinct creatives. Read it through outbound CTR, product-view rate, and add-to-cart rate—not cheap clicks alone. Weak product views point to Pin/title/page mismatch; healthy views with no carts point to the offer or product page. Set a seven-to-fourteen-day decision window before launch. This lane graduates only when it produces engaged visits and enough lower-funnel events to inform purchase optimization.

Lane 2: purchase-volume engine — oCPM

Allocate 50–65% to oCPM once Purchase events fire reliably. It is the workhorse for a CPA ceiling or catalogs where similar order values make value-weighting less useful. Keep it always on: Pinterest says six-month-plus Shopping campaigns delivered around 25% higher ROAS in its analysis, while campaigns under three months were 5–16% below average. Those are vendor claims, not guarantees, but they support one operating rule: do not reset learning over a three-day report. If delivery is erratic, merge thin groups with the same margin and intent before blaming the bid.

Lane 3: value capture — ROAS bidding

Allocate 15–25% to ROAS only after value data passes the preflight. It is for catalogs where a $25 accessory and a $1,500 sofa should not receive equal priority. Set a target that leaves room to explore, then judge it against contribution margin—not platform revenue that ignores discounts, returns, and shipping. Pinterest reported a median 10% ROAS gain for beta Performance+ ROAS bidders. Use that as a reason to run a controlled test against oCPM, never as proof that the system knows your profitability.

A 21-day test that does not confuse itself

A bid portfolio only works when changes are disciplined. Use this sequence:

The aim is to finish the period knowing what changes next—not to manufacture certainty in 21 days.

What to change when the mix is underperforming

Use the diagnosis before the optimization:

• oCPC gets clicks but weak product views: Rewrite catalog titles around the actual shopper need, inspect image relevance, and make the landing page load faster on mobile.
• Product views and carts are healthy but purchases are weak: Audit price parity, stock, delivery charges, payment friction, and the checkout flow. A higher bid cannot fix purchase resistance.
• oCPM produces purchases but unstable CPA: Consolidate thin product groups, check event duplication, and resist short stop-start cycles that erase the learning you just paid for.
• ROAS bidding spends but value is disappointing: Verify purchase value, remove low-margin items from the test, and compare against the oCPM control. Do not raise the ROAS target as a cosmetic fix.
• All lanes look weak: Return to the catalog and offer. Bidding is an amplifier, not a rescue plan.

Pinterest is not “a cheaper Meta.” It catches intent while people plan, which needs a different rhythm than the feed-based conversion work in our Pinterest Shopping Ads overview and a stronger creative supply than most promotion-only launches. See Pinterest’s newer ad tools for that creative layer, and our Q4 launch-timing analysis for why campaigns should be live before—not on—the first sale day.

The point is controlled learning, not more settings

Pinterest’s bid menu is not the strategy. The strategy is deciding what the account must learn next, giving each campaign a single job, and refusing to scale a result that rests on bad signals.

Start with a clean catalog and complete measurement. Let oCPC diagnose qualified demand. Let oCPM establish purchase volume. Let ROAS bidding compete only after checkout values are credible. Then compare results on margin and incrementality, not on the platform’s most flattering number.

That is how a Pinterest Shopping test becomes a media decision instead of another campaign someone turns off before it had a chance to learn.

Sources: Pinterest Business, “The next step: How to maximize your shopping results on Pinterest” (June 24, 2025); Pinterest Business, “5 best practices for better shopping performance on Pinterest” (June 29, 2026); Pinterest Business, “Less effort, better results: Transform your campaigns with Pinterest Performance+” (April 11, 2025).