14 capability spines. One rubric. Same questions on every platform.
Every commerce platform and ERP in Celeste's database is scored against the same 14 spines, broken into specific evaluation questions, sourced from vendor docs, public implementations, and Acro's engagement history.
13,305 evaluations on record. No vendor influence. No pay-to-play tier. No curated shortlist hiding behind a recommendation. The same lens applied to every platform, then weighted by what your business actually does.
The 14 spines.
Each spine breaks down into the questions Celeste scores against. The list under each spine is what gets evaluated.
Catalog architecture
How the catalog models products, variants, configurables, and merchandising logic.
- Configurable / bundle / kit handling
- Variant attribute model and inheritance
- Catalog size ceiling and faceted-search performance
- Personalized catalog per customer / account / segment
- Multi-language / multi-currency catalog parity
B2B pricing logic
Customer-specific pricing - where most B2B deals live or die.
- Contract pricing per account
- Tiered + volume + threshold rules
- Quote-driven pricing and rep overrides
- Promotion / discount stacking and conflicts
- Pricing source of truth (storefront vs ERP)
Order routing
How orders move from intake to fulfillment.
- Multi-warehouse selection at checkout
- Split-ship and partial-fulfillment logic
- Drop-ship and supplier-direct paths
- Backorder, pre-order, and ATP behaviour
- Returns flow (RMA, restock, credit, replace)
ERP integration
Native vs middleware, and where each model breaks.
- Coverage of the native connector (if any)
- Real-time vs batch sync windows
- Customer / pricing / inventory directionality
- Custom-field and metadata round-trip
- Failure handling and observability
Customer accounts
Roles, hierarchies, approvals, buyer-side controls.
- Account hierarchies and entity rollups
- Role-based permissions and purchase rules
- Buyer-side approval workflows
- Multi-buyer carts and shared lists
- Self-service profile, payment, address management
Payments and credit
Terms, credit holds, payment-on-account, multi-currency.
- Net-terms and credit-hold logic
- Partial payment / split tender
- Multi-currency and FX handling
- Stored payment, AR application, deposits
- Tokenisation, PCI scope, fraud signals
Quote and proposal
Quote-to-order, rep-assisted carts, sales-pricing overrides.
- Quote lifecycle (draft, sent, expired, won)
- Rep-assisted cart and impersonation
- Approval thresholds on quotes
- Conversion of quote to order without reprice
- Proposal / document generation
Channel coverage
EDI, marketplaces, partner portals, distributor catalogs.
- EDI document set (850, 855, 856, 810, 846)
- Marketplace listing automation
- Partner / distributor portals
- Punchout (Ariba, Coupa, SAP)
- Headless API surface for owned channels
Inventory accuracy
Real-time vs cached availability, lot / serial tracking, regulatory metadata.
- Cache freshness window for storefront ATP
- Lot and serial tracking through fulfillment
- Regulated metadata (UDI, FDA, hazmat)
- Reservation logic during checkout
- Cross-channel inventory reconciliation
Tax and compliance
Tax engine, exemption certificates, regulated-industry workflows.
- Tax engine integration (Avalara, Vertex, native)
- Resale / exemption certificate workflow
- Cross-border duty + landed cost
- Industry compliance (FDA, FSMA, HIPAA, ITAR, FedRAMP)
- Audit trail and tax-period close
Customization model
What you can change without breaking upgrades. The configure-vs-code line.
- Configuration depth before custom code is required
- Theme / template safety on upgrades
- Extensibility API and event hooks
- Sandboxing and multi-environment support
- Upgrade cadence and breaking-change history
Operations and tooling
Admin UX, search, reporting, automation, observability.
- Admin UX for ops, marketing, CS
- Search and merchandising controls
- Workflow automation and triggers
- Reporting depth and export
- Observability: logs, traces, error alerts
Performance and scale
Cache strategy, search infrastructure, headless boundaries, catalog scale ceilings.
- Catalog scale ceiling under load
- Page render strategy (SSR, ISR, edge)
- Search infrastructure (native, Algolia, Elastic)
- Headless boundary and CDN coverage
- Concurrent-user benchmark behaviour
Total cost trajectory
Licence + implementation + run cost across years two and three.
- Licence model and revenue-share tiers
- Typical implementation cost band
- Year 2-3 cost drivers (apps, ops, custom)
- Hidden costs (transaction fees, addons)
- Cost trajectory under your growth shape
Five scoring levels. The same five everywhere.
Every sub-question gets one of five letters. The letter names the architectural fit, not just whether the capability exists.
The platform handles this out of the box without configuration tricks. Real native, not "available via app".
BigCommerce B2B Edition has native account hierarchies. Score: Y.
Configurable to do this, but the configuration is non-trivial OR carries a known operational cost (latency, edge case, etc).
Shopify Plus does contract pricing via discount logic + scripts. Score: B.
Functionality lives in a third-party app. Quality varies and you take on the app vendor as a dependency.
WooCommerce quote-to-order via a plugin. Score: A.
You can build it, but it is custom development on top of the platform. Counts as scope on day one.
Magento custom EDI for non-standard transaction set. Score: C.
The platform cannot do this without violating its architecture. You will fight the platform forever.
Shopify multi-entity legal-entity rollups in one storefront. Score: N.
One spine, one platform: B2B pricing on Shopify Plus.
Here is how the B2B pricing spine gets scored on Shopify Plus (with B2B catalog enabled). Each sub-question gets a letter and a one-line rationale.
Your report shows the spine read per platform, weighted by what matters most for your business. That weighting is where 13,305 evaluations becomes a recommendation rather than a comparison grid.
Where the 13,305 evaluations come from.
Vendor documentation
Public developer docs, partner portals, certification programs, and pricing pages. Cited per platform.
Public implementations
Storefront crawls, BuiltWith data, integration registries, and capability audits across real customer sites running each platform.
Acro engagement history
Anonymized data from 100+ Acro Commerce engagements. The corner cases vendors do not document - and the ones Acro had to engineer around.
Same scoring. Different lens per business.
A B2B distributor with contract pricing and an ERP-of-record gets a different weighting than a manufacturer running 1,200 SKUs with quote-driven sales gets a different weighting than a CPG brand running DTC. The spines are the same; the importance Celeste assigns to each spine shifts based on the way you actually operate.
That is why two businesses on the same shortlist of platforms can get different #1 recommendations from Celeste. Same evidence, weighted to the reality of the buyer, not a generic ICP.
Acro's architects use the same scoring engine in paid Discovery and Strategy engagements.
Celeste is free for the diagnostic because architectural clarity should precede a six-figure platform decision. If the report points you somewhere Acro cannot help, that is fine. Make the right call. If it points you somewhere Acro is built to deliver, we are here.
Run Celeste on your URL