Calculateur CPA — coût par acquisition formule + cible

Across aggregated Google Ads data 2025-2026 referenced continuously, CPA is the most-tracked metric on Google Ads accounts — and the most miscalibrated. The formula is simple (spend divided by conversions), but operational use poses three recurring traps: (1) confusing gross CPA and net CPA, (2) setting Target CPA on a generic benchmark instead of gross margin, (3) optimizing CPA without measuring LTV or overall CAC. The calculator above returns gross CPA. What follows explains how to transform it into business-realistic CPA, how to compare it to gross margin via the 30% rule, and how to apply the 10 levers that cut CPA by 14 to 22% in 30 days.

For cross ROAS / CPA / CPC fundamentals, see our complete ROAS CPA CPC guide. For the 10 CPA reduction levers detail, see CPA reduction guide. For Smart Bidding Target CPA vs Maximize Conversions steering, see Smart Bidding comparison.

CPA formula and quick calculation

CPA (Cost Per Acquisition) is the total cost to obtain a conversion — purchase, lead, signup, quote request. Base formula: CPA = Total spend / Number of conversions. If you spent €1,500 to generate 35 conversions over 30 days, your CPA is €42.86. It's the metric tracked by 90% of advertisers and the main budget arbitrage variable in lead gen and SaaS B2B where conversion value is broadly homogeneous.

The alternative formula is more powerful for steering: CPA = CPC / Conversion rate. This second formulation makes explicit what you can do to lower CPA. Either lower CPC (better Quality Score, negatives, suitable Smart Bidding), or raise conversion rate (landing page, message-match, reduced checkout friction). The second is typically 10x more powerful than the first — doubling your conversion rate (from 2% to 4%) mechanically halves your CPA without touching bids.

Official Google documentation on CPA calculation: support.google.com Cost Per Acquisition. Note that Google Ads displays "all conversions" CPA by default, which can include micro-conversions (page views, cart adds) if you've set them as conversions. For clean steering, only track business conversions with P&L value (purchase, qualified lead, paying signup) — otherwise the displayed CPA in the interface no longer corresponds to actual profitability.

CPA benchmarks by vertical 2026

The orders of magnitude below come from aggregated Google Ads data 2025-2026 (public sources + Google Ads API), cross-referenced with public benchmarks WordStream Google Ads. The ranges correspond to vertical medians — intra-vertical variance remains high based on campaign maturity, tracking quality, gross margin, and account size.

Practical reading: if your CPA sits at the median of your vertical but profitability is bad, two typical causes. (1) Your gross margin is lower than the sector benchmark — verify the discount/promo share in the sales mix and non-media variable costs (logistics, payment, customer service). (2) Your conversions lost post-tracking exceed 15% — typical in fashion e-commerce with product returns or B2C lead gen where 30-40% of leads are CRM-disqualified. The next section details the gross CPA vs net CPA mechanic.

Acquisition CPA vs business-realistic CPA: the metric that matters

This is the nuance that 80% of Google Ads dashboards ignore and that explains why so many accounts show "CPA fine" while the P&L says "degraded profitability". The gross CPA shown by Google Ads is a tracking acquisition CPA — it relates spend to conversions tracked by the conversion tag. Business-realistic CPA includes conversions lost between tracking and final P&L.

Three loss sources between tracked conversion and P&L conversion:

• Product returns in e-commerce. In mode, up to 18-25% of orders are returned (fitting, size, disappointment). In consumer electronics, 8-15%. In beauty/cosmetics, 4-8%. Gross CPA counts the click as converted, P&L counts the net sale. Gross vs net CPA gap: +18 to +25% in fashion.

• CRM-unqualified leads in B2B. In SaaS B2B, 30 to 50% of MQL leads are disqualified to SQL after CRM scoring (company size, function, intent). In high-intent B2C lead gen (insurance, credit, energy), 20 to 35% are rejected post-call (wrong geo, wrong profile, off-target). Gross vs net CPA gap: +30 to +50%.

• Post-conversion drop-offs in services. In premium services (training, consulting, travel), 5 to 12% of tracked conversions drop off before billing. Gross vs net CPA: +5 to +12%.

Net business-realistic CPA calculation: Net CPA = Gross CPA x (1 / final conversion rate). Fashion e-com example: if you track 100 conversions at €25 gross CPA, but 20% are returned and 5% canceled, your final conversion rate is 75%. Net CPA = 25 x (1 / 0.75) = €33.33. It's the CPA to compare to your gross margin.

To steer cleanly, configure adjusted-value conversions in Google Ads: monthly import from CRM or Shopify the actually paid (not tracked) conversions via the Offline Conversion Imports mechanism. Official documentation: support.google.com Offline Conversions. This step is heavy to set up but radically transforms steering quality — Smart Bidding then optimizes on real P&L CPA, not artificial tracking CPA.

CPA / gross margin ratio: the 30% rule

This is the practical rule that separates profitable accounts from those that think they are. Across accounts observed in public Google Ads benchmarks, durably profitable advertisers maintain a net CPA below 30% of gross margin per customer in occasional e-commerce, below 25% in high-retention e-commerce, and indexed on 12-month LTV (not first-purchase margin) in SaaS B2B.

Mechanic: if your gross margin per customer is €40 (€100 basket at 40% margin), you can spend up to €12 in CPA and keep €28 of contribution margin after acquisition. Above €12 CPA, you eat into your contribution margin and rely on repeat purchases to fill the gap — a viable model only if you measure and hold your 12-month retention.

The rule tightens by business profile:

• Occasional e-commerce (retention below 1.5 purchases / 12 months): strict ratio CPA below 30% of gross margin per customer. Without retention to amortize, every acquisition must be profitable at d+30.
• High-retention e-commerce (mode, beauty with 2.5+ purchases/year): ratio relaxed to 35-45% of first-purchase margin, indexed on 12-month LTV or LTV/CAC ratio above 3.
• SaaS B2B: LTV-aware steering mandatory. CPA target below one-third of cumulative 12-24 month LTV. With ARPU €200/month and 18-month retention, gross LTV €3,600 x 75% gross margin = €2,700. Target CPA below €900.
• B2C lead gen: target CPA below 25% of gross margin per converted lead (not per trackable lead — corrected for CRM qualification rate).

Classic mistake: using an "industry standard" CPA / gross margin ratio of 30% type without checking your cohorts' real retention. On referenced accounts, the median gap between assumed retention and measured retention is 25 to 40% — typically advertisers overestimate initial retention because they reason on early adopters, not on mainstream cohorts that scale afterward.

How to set a Smart Bidding Target CPA

The robust 3-step method to calibrate Target CPA without killing delivery at D+14. It's the most common mistake in modern Smart Bidding — setting a Target CPA 30 to 40% below the historical observed level believing the algorithm "will figure it out". Across aggregated Google Ads data 2025-2026 (public sources + Google Ads API), the result in the majority of cases is unambiguous: -55 to -72% volume in 14 days, the algo cuts delivery rather than going below threshold.

Step 1 — Measure real historical CPA over 30 days rolling. Not the 90-day average that smooths variations, but the actual CPA of the last 30 days, segmented by campaign. If weekly CPA varies more than 20% over those 4 weeks, wait for stabilization before any Smart Bidding change — otherwise you're calibrating on noise.

Step 2 — Start Target CPA at +10 to +15% above historical CPA. The goal is to let Smart Bidding exit learning without delivery cuts. This margin is necessary because the algorithm needs an operational window to collect signal — setting Target CPA exactly at historical CPA pins it to the boundary and causes oscillations.

Step 3 — Lower in 10% steps every 14 days as long as volume holds. Watch two metrics at each step: absolute conversion volume (at least 30 conversions/30d to stay out of learning) and weighted CPA by device/audience segment. If volume drops more than 20% at a step, that's the signal you're hitting the floor — go back up a step and stabilize for 30 days before another attempt.

For Maximize Conversions vs Target CPA detail by account profile, see our Smart Bidding comparison. For LTV-aware Target CPA calibration in SaaS B2B, see our CPC calculator and our conversion rate calculator.

10 levers to cut CPA by 14-22% in 30 days

Here is the operational sequence ranked by effort/impact ratio, observed across aggregated Google Ads data 2025-2026. Accounts that apply these 10 levers in 30 days mostly see a CPA drop of 14 to 22% — bigger gains if the account starts from a bottom 25% of the vertical benchmark.

Lever 1 — Quality Score audit on top 20% keywords. Target the keywords capturing 80% of spend with Quality Score below 6/10. Work Ad Relevance (copy/keyword alignment), Landing Page Experience (Core Web Vitals), Expected CTR (RSA pinning). Effect: -24 to -36% CPC on these keywords, so -24 to -36% CPA at constant conversion rate. It's the #1 ROI lever. See Quality Score guide.

Lever 2 — Increase landing page conversion rate. Doubling conversion rate mechanically halves CPA without touching CPC. Work ad-to-landing message-match, single CTA, social proof, Core Web Vitals, mobile-first. Effect: -30 to -50% CPA if initial conversion rate is below 1.5%. See landing pages guide.

Lever 3 — Activate well-calibrated Smart Bidding Target CPA. Method described above. Observed effect: -8 to -18% CPA in 30 days on accounts starting from Manual CPC or Maximize Clicks.

Lever 4 — Tighten match types and add negatives. Audit Search Term Reports of the last 14 days, negatives on top 50 impressed non-converted queries. Effect: -10 to -20% CPA. See 2026 match types guide.

Lever 5 — Layer audience observation then bid modifiers. Customer Match, In-Market, Affinity in observation 30 days, then bid modifiers +20 to +40% on overperforming conversion segments. Effect: -8 to -15% weighted CPA.

Lever 6 — Activate Enhanced Conversions for Web and Offline Conversion Imports. Improves Smart Bidding signal by 12 to 25% by vertical. CPA effect: -5 to -12% in 30 days.

Lever 7 — Test 4-5 new RSAs per ad group. Structured pinning, headlines with H1 keyword, H2 USP, dynamic pool. Effect via Quality Score: -8 to -15% CPC, so -8 to -15% CPA.

Lever 8 — Activate all available extensions. Sitelinks, callouts, structured, prices, image, lead form. +4 to +12% CTR per active extension, mechanical CPC drop, mechanical CPA drop.

Lever 9 — Hourly and geographic bid adjustments. Identify hours and zones with overperforming conversion rate, push +20 to +50% on these segments, lower -20 to -40% on low-converting segments. Effect: -5 to -10% weighted CPA.

Lever 10 — Cut campaigns or ad groups with structurally above 1.5x median account CPA. Reallocate budget toward performing campaigns. Effect: -8 to -15% weighted account CPA. See our 10 Google Ads mistakes and our Google Ads audit checklist.

CPA remains the most useful steering metric in 2026 — provided you read it correctly. The calculator above returns gross CPA. The work starts after: recalculate as net business-realistic CPA, compare to the 30% gross margin threshold, calibrate Target CPA progressively, and apply the 10 levers in effort x impact order. On accounts following this discipline, the observation is stable: -14 to -22% CPA in 30 days, and a durable alignment between media KPI and finance P&L that makes monthly budget arbitrages much easier to present in committee.

FAQ