# Independence Is Decided Per Layer Not Per Stack ## Claim Whether you own your software is not one decision but one per layer, and a stack can be self-hosted where it differentiates you and rented where it does not. <!-- atomicity: 1 idea ✓ --> ## Scope & conditions - Holds where layers have genuinely separable data and interfaces, so a rented layer can be exited independently of the rest of the stack. - Breaks where the rented layer holds the system of record — here, customer and order data sits in BigCommerce — in which case the per-layer framing understates the lock-in, because the layer you rented is the one you cannot leave. - Applies whether the motivation is cost, capability, or (as in this case) offloading operational complexity you don't want to own. ## Evidence WP Engine's Commerce Connect keeps WordPress for content, brand, and SEO while moving payments, shipping, tax, orders, and customers to BigCommerce — a deliberate split rather than a full re-platform. The product's own pitch treats this as obviously correct without naming it as a lock-in tradeoff: migration in is wizard-driven; what leaving the commerce layer costs is never addressed. CVE-2026-65643 supplies the security version of the same layer rule. A WordPress site can be patched and hardened while its shared-hosting tenant boundary fails in cPanel, a control plane the site owner does not operate. ## Links - [[Product & Startup Instincts MOC]] — ownership as a series of separate decisions rather than one posture - [[3. Resources/Permanent Notes/Building on a Platform You Don't Control Is Renting Not Owning]] — because contrasts; that note treats "renting" as uniformly a risk, this one argues renting a specific layer can be a coherent, decided position rather than a drift - [[3. Resources/Permanent Notes/The Hard Part of Replacing SaaS Is Recovering Your Data]] — because extends; names exactly what "the rented layer is the one you cannot leave" costs in practice - [[3. Resources/Permanent Notes/Lock-In Removes the Discipline That Would Punish Decay]] — because applies-in; a rented commerce layer only stays honest for as long as switching it is actually possible - [[3. Resources/MOCs/Open Source Governance MOC]] — because applies-in; per-layer independence is the right unit of analysis for evaluating hybrid open-source-plus-SaaS ecosystem choices - [[3. Resources/MOCs/WordPress Hosting Security Stack MOC]] — applies-in: the MOC lists this note under Synthesis additions (2026-08-28 — cPanel tenant isolation). ## Tests / falsifiers - Model migration-out cost for this hybrid architecture versus a single-vendor SaaS platform and versus staying on WooCommerce; if exit cost is comparable across all three, the per-layer framing adds no decision-relevant information. - Track whether WP Engine or Commerce ever publishes data-portability terms for the commerce layer; their absence at launch is itself evidence the tradeoff wasn't decided, only priced into the pitch. ## Maintenance - Last tended: 2026-08-16 - Next review: 2026-11-16 ## Source - [[3. Resources/Readwise/Articles/WP Engine Partners With BigCommerce Parent Commerce to Launch Ecommerce Platform for High-Growth Merchants]] — reports the content/commerce split, the wizard-driven migration pitch, and the absence of any migration-out discussion. - [[3. Resources/Literature Notes/Critical cPanel Flaw Could Let One Hosting Customer Take Root Control - Literature Note]] — hosting-layer isolation failure that site-level controls cannot repair.