Tarification de l'API Google Maps en 2026 : ce que coûte vraiment votre Store Locator

L'API Google Maps facture entre environ 2 et 40 dollars pour 1 000 requêtes, selon le SKU et le palier tarifaire. Un Store Locator standard à 10 000 sessions par mois revient à environ 285 $ ; à 100 000 sessions, environ 4 295 $ ; à 500 000 sessions, environ 17 975 $ — principalement du fait des éléments Place Details Pro et Route Matrix. Le crédit mensuel gratuit ne couvre que les petits projets ; les abonnements couvrent les niveaux Essentials et Pro, mais pas Enterprise.

Ce qui a changé en mars 2025

Google a remplacé son crédit mensuel forfaitaire de 200 $ par des plafonds d'utilisation gratuite par SKU et a introduit trois formules d'abonnement. Ce changement modifie en profondeur la façon dont les coûts s'accumulent :

Avant (avant mars 2025) : un crédit unique de 200 $ s'appliquait à l'ensemble des usages API. Une équipe utilisant Maps, Geocoding et Places pouvait absorber environ 200 $ d'utilisation mixte avant de commencer à payer.

Après (à partir de mars 2025) : chaque SKU dispose de son propre plafond gratuit — 10 000 événements pour les SKU Essentials, 5 000 pour Pro et 1 000 pour Enterprise. Les usages gratuits ne se cumulent plus entre API. Un Store Locator utilisant quatre API Essentials bénéficie de 40 000 événements gratuits au total — mais seulement 10 000 par API.

The Hidden Math: What a Store Locator Really Costs

Most pricing guides list individual SKU rates. But no application uses a single SKU. A store locator page typically triggers four APIs per user search:

• Dynamic Map - display the map (1 load per page view)
• Autocomplete - user types an address (session-billed when terminated with Place Details at Pro or Enterprise tier → keystrokes are free; per-request at $2.83/1K when used without session tokens or when the session is abandoned)
• Place Details (Pro) - resolve the user-entered address into coordinates (1 call per search, terminates the autocomplete session)
• Route Matrix - calculate distance to candidate stores (1 call per search, element count = number of candidate stores × 1 origin)

Store data itself - names, opening hours, phone numbers - is typically hosted by the retailer (JSON file, CMS, or internal API). Google Places is not called to display company-owned stores. A retailer showing the 5 closest stores from a user search generates, at minimum, per session: 1 map load + autocomplete keystrokes (free in session) + 1 Place Details Pro + 1 Route Matrix call with 5 elements. That is 4 billable events (or 5 if counting the matrix elements separately) across 4 SKUs.

Three Cost Traps That Inflate Your Bill

1. Route Matrix element multiplication

Route Matrix does not bill per request. It bills per element: origins multiplied by destinations. A logistics platform calculating ETAs from 10 warehouses to 500 delivery addresses generates 5,000 elements in one API call. At $5/1K elements, that single call costs $25. Run it hourly for real-time updates and the monthly bill for Route Matrix alone reaches $18,000.

2. Place Details tier confusion

Place Details exists at three tiers: Essentials ($5/1K), Pro ($17/1K), and Enterprise ($20/1K). The tier is determined by which fields you request, not which plan you are on. A session terminated with formattedAddress and geometry triggers Essentials. Adding displayName, types, or addressComponents moves the request to Pro. Requesting photos, rating, or currentOpeningHours pushes it to Enterprise. Without explicit field masks, a simple address resolution call can cost 3-4x more than expected - and since Place Details terminates the autocomplete session, a single mis-configured field list multiplies across every user search on the site.

3. Autocomplete billing switches silently when sessions break

Autocomplete pricing looks free at first glance - sessions terminated by Place Details Pro or Enterprise make all keystrokes free (SKU "Autocomplete Session Usage" = $0). But the session billing is conditional:

• Session abandoned (user closes the tab before a Place Details call): every keystroke retroactively falls back to Autocomplete Requests at $2.83/1K. A 15-keystroke abandonment costs $0.042.
• Session terminated by Place Details Essentials: only autocomplete requests from the 13th onward are free; the first 12 are billed at $2.83/1K.
• Session token reused across requests: Google invalidates the session and bills all autocomplete as per-request.

At 100K sessions/month with a 20% abandonment rate and roughly 10 autocomplete requests per abandoned session, these fallbacks add about $565/month on top of the "free" autocomplete line (20,000 abandoned × 10 requests × $2.83/1K). Monitoring session completion rate is the only way to see it in the billing report.

Beyond Price: What Else Changes When You Switch

Cost is the catalyst, but three operational factors determine whether a migration delivers lasting value:

Data privacy and security. Google routes requests from its Maps and Places APIs through US-based infrastructure regardless of the calling region. For EU retailers processing customer addresses, this creates a data transfer obligation under GDPR Article 44. The EU Digital Markets Act (Article 6.5) has recognized Google's self-preferencing practices. Woosmap is EU-hosted with data processed in European infrastructure. Mapbox routes through US-based AWS. HERE offers multi-cloud deployment including EU regions.

End-user data usage rights. Major location providers include clauses in their terms of service that reserve the right to use end-user query data to improve their products, train models, or feed their broader advertising ecosystem. Google Maps Platform and Mapbox each reserve such rights in their current terms; HERE's provisions vary by product line. The specifics differ across providers and product tiers, and legal review of the data-processing addenda is a standard step before committing to any maps provider. For retailers handling sensitive customer addresses (healthcare, financial services, or markets with strict consent regimes), this clause often matters more than the per-1K rate. Woosmap's Terms of Service do not claim rights to reuse customer request data for model training or advertising, and its privacy commitments prohibit selling client personal data or running automated decision-making (profiling); any reuse beyond service delivery would require explicit contractual authorisation.

Terms of service restrictions. Google's TOS restrict caching geocoding results and prohibit displaying them on non-Google maps. If your architecture requires caching addresses for performance or displaying results on a custom map, this creates a hard technical constraint. Most alternatives, including Woosmap, impose no such restrictions. For a detailed comparison of how Azure Maps and other providers handle data residency, see our platform-by-platform guides.

Billing predictability. Google's current SKU structure - per-element Route Matrix, tier-dependent Place Details, and autocomplete billing that flips between free and per-request depending on session outcome - creates variance that is hard to forecast month over month. The March 2025 restructuring added subscription options (Starter, Essentials, Pro) but also removed the pooled credit that had absorbed usage spikes across APIs. Budgeting against the new model requires tracking session completion rates, field-mask hygiene, and Route Matrix element counts separately. Woosmap's pricing is SKU-based as well, but each SKU has a single rate by volume tier with no mid-session tier shifts and no abandoned-session penalty - a request that completes and a request that fails are billed identically.

How to Evaluate Your Current Google Maps Spend

Reading list prices is useful. Looking at your actual bill is decisive. Five steps to turn a monthly Google invoice into an evaluable migration case:

• Export your billing report from Google Cloud Console. Filter by SKU to see which APIs drive the most cost. Most teams discover that 60-80% of spend comes from two or three SKUs.
• Map your user journeys. Count the API calls per user session for your top three workflows. The store locator math above is one pattern - checkout address validation is another (Autocomplete + Address Validation + Geocoding = 3 APIs per order).
• Calculate your effective rate. Divide total monthly spend by total API calls. If your effective rate exceeds $4/1K requests, you are likely paying a premium that alternatives can undercut.
• Check your field masks and session completion rate. If you use Places API, verify that your implementation requests only the fields it needs - photo or atmosphere fields trigger Pro or Enterprise pricing ($17-$20/1K instead of $5/1K). Then look at how many autocomplete sessions complete with a Place Details call: every abandoned session falls back to per-request autocomplete billing.
• Model the migration as a discrete project, not a leap of faith. A typical e-commerce front-end migration from Google Places plus Geocoding to a compatible provider takes 2-4 weeks of engineering time, dominated by address-format normalization and field-mask rewiring - not by the map renderer swap. Against annual savings of $30K-$120K at growth-to-enterprise scale (see the scenarios above), payback on a $50K engineering investment lands at roughly 5 months at enterprise scale ($50K ÷ $10K monthly savings ≈ 5) and stretches to around 20 months at growth scale ($50K ÷ $2.5K monthly savings ≈ 20). Below 10K sessions per month, the economics rarely justify the migration on cost alone - the cases that do are usually driven by data residency, TOS constraints, or billing predictability.

Cette analyse a été rédigée par Jean-Thomas Rouzin, PDG de Woosmap. Jean-Thomas dirige une plateforme européenne de renseignements géolocalisés au service de 220+ clients d'entreprise dans la vente au détail, la logistique et les voyages, traitant plus de 27 milliards de requêtes API par an avec un SLA de 99,9% sur le plan Enterprise.

Pour la liste complète des alternatives à Google Maps évaluées par notre équipe, consultez notre guide de comparaison détaillé.

Visitez woosmap.com* pour explorer la plateforme.*