Composable Regret

Composable Regret is the phenomenon where retailers, brands, and enterprises that adopted composable / MACH Architecture / headless commerce architectures report that the complexity, cost, and operational overhead exceeded expectations — sometimes reversing course back toward packaged or hybrid platforms. The term entered mainstream DXP and commerce industry vocabulary in late 2024–2025, driven primarily by VTEX's public exit from the MACH Alliance in March 2025 and subsequent trade-press synthesis.

---

Background

Composable Commerce promised that assembling best-of-breed services via APIs — a Packaged Business Capabilities (PBCs) model — would give retailers flexibility, speed, and long-term resilience. By 2023–2024 a counter-narrative emerged: for many organisations, the architectural freedom produced integration complexity, hidden total cost of ownership (TCO), and operational burden that exceeded the benefits. VTEX Co-CEO Mariano Gomide de Faria first named these failure modes publicly in a LinkedIn post and subsequent blog post titled "The MACH Mirage" (VTEX, published 2025-03-25), and announced VTEX was temporarily suspending its MACH Alliance membership. (vtex.com/en-us/blog/mach-mirage, as-of 2025-03-25)

---

Named failure modes

VTEX's "MACH Mirage" post identifies four structural failure modes (as-of 2025-03-25):

• Integration Puzzle — API mediation layers create fragile cascade failures; each new service adds integration maintenance burden. "The 'best-of-breed' narrative conveniently ignores the crippling reality of managing multiple vendor relationships, each with its own licensing fees, support contracts, and infrastructure requirements." (VTEX blog, 2025-03-25)
• Hidden Cost Epidemic — licensing, development, and maintenance costs compound into a "financial black hole." VTEX characterised implementing highly customised composable solutions as requiring organisations to "build a software company within their organisation." (VTEX blog, 2025-03-25)
• Operational Quagmire — business users navigate disconnected tools; marketing teams reportedly wait weeks for developer deploys. Umbraco CTO Filip Bech-Larsen stated: "Engineering backlogs multiplied, meaning that marketing teams have sometimes had to wait weeks for changes." (CMS Critic, 2025-11-19)
• Security and Privacy Nightmare — data replicated across dozens of vendors multiplies PCI and privacy audit scope. Bech-Larsen noted that juggling multiple vendors "increases risk in regulated industries." (CMS Critic, 2025-11-19)

An additional root cause identified by practitioners: organisational unreadiness. Umbraco's "Composable Regrets: Lessons From the First Wave" blog argues: "Most composable projects do not fail because of APIs or services. They fail because data is fragmented, inconsistent, or poorly understood." Enterprises became "de facto system integrators, stitching APIs together, debugging data flows, and juggling multiple vendor relationships." (Umbraco blog, approx. 2025)

---

Key voices

• Mariano Gomide de Faria (VTEX Co-CEO): "What began as a technical liberation movement has instead led us down a treacherous path paved with hidden costs, operational nightmares, unfulfilled promises, and financial ruin." (via RETHINK Retail, 2025-05-29, reporting his LinkedIn post)
• Filip Bech-Larsen (Umbraco CTO): "We've heard about the 'composable regret' people are talking about, and all the promises of composable are not really happening." (CMS Critic, 2025-11-19)
• Beth Moriarty (VTEX) at Shoptalk Spring 2025: "A lot of brands are basically failing in their approach to do a fully composable best-in-breed deployment." (RETHINK Retail YouTube, 2025-04)
• Jenna Flateman Posner (CDO, SNIPES): "MACH promised freedom from monoliths, but for many, it delivered complexity, rising costs, and integration pain." (RETHINK Retail, 2025-05-29)
• Dilip Keshu (agency board member): "I felt MACH took itself too seriously… MACH had its day in the sun, but new trends always follow." (RETHINK Retail, 2025-05-29)
• Michael von Bodungen (VTEX General Manager NA) at Shoptalk 2025: "Retailers are really focused on how they can make their tech stack simpler to reduce the total cost of ownership, but still get strong value out of the platforms they run." (RETHINK Retail YouTube, 2025-04)

---

VTEX's proposed alternative: Pragmatic Composability

VTEX simultaneously launched a self-assessment tool ("Composability Extremism Index": mkt.vtex.com/composability-extremism-index) and proposed "Pragmatic Composability" — pre-integrated out-of-the-box capabilities for core functions, with modular best-of-breed only where there is measurable ROI. Conditions VTEX set for re-engaging with the MACH Alliance: lean/simple architecture; OOTB fully connected ecosystem; no or minimal middleware; defending retailer interests over provider interests. (VTEX blog, 2025-03-25; Retail Technology Innovation Hub, 2025-03-26)

VTEX CEO on the direction: "For VTEX, the future is not in giving brands a Lego box of disconnected tools, but in delivering a commerce system that feels more like a service than software." (Retail Tech Podcast, 2025-05)

VTEX reframed MACH principles as "table stakes for modern web application development" no longer deserving a dedicated alliance. (Forbes, 2025-03-27; VTEX blog, 2025-03-25)

---

MACH Alliance response

The MACH Alliance did not directly rebut the "composable regret" narrative. Instead, its 2026 Enterprise Technology Report (surveying 600 enterprise decision-makers across 7 global markets) pivoted to an AI/agentic framing: "78% of organisations with fully implemented, scaled MACH technology report clear evidence of ROI on AI investments, compared to 13% of organisations in early planning stages of MACH." (MACH Alliance ETR 2026, as-of 2026-02, PRIMARY)

The Alliance closed 2025 as its largest-ever member cohort (100+ organisations) and characterised 2026 as the "agentic commerce" era, making no public reference to "composable regret" in its Insights Hub content stream as of mid-2026. (BusinessWire, 2025-12-16)

MACH Alliance President Casper Aagaard Rasmussen publicly requested an open conversation with VTEX following the exit, and introduced a "Best Platform Transformation" award category at the 2025 Composable Conference, acknowledging that the path to composability is incremental and varied. (RETHINK Retail, 2025-05-29)

---

Platform vendor TCO counter-claims

Pro-composable TCO data from first-party vendor sources (treat as interested parties):

• Salling Group (Danish retail): 75% TCO reduction claimed (commercetools blog, date unconfirmed)
• APG & Co. (Australian fashion): 35% reduction in technology operating costs (commercetools blog, date unconfirmed)
• Contentful / Bang & Olufsen: +60% CVR, +27% AOV (Contentful, 2024-11, per Packaged Business Capabilities (PBCs) run 2026-08-14)

---

Orchestration as proposed synthesis

A third camp — distinct from "composable failed" (VTEX) and "composable is fine" (MACH Alliance) — proposes orchestration layers as the path through the regret problem:

• Umbraco announced for 2026 a SaaS ingestion endpoint plus GraphQL Delivery API composing content from PIM, DAM, ERM, CRM, and CMS into a unified output, citing Forrester's finding that the market is shifting toward "wider adoption of GraphQL APIs" and "turnkey connectors." (CMS Critic, 2025-11-19)
• 64labs (Salesforce Commerce Cloud SI) argues that SFCC has "quietly pulled ahead" by becoming "de-composable" — offering headless-capable storefronts without requiring full disaggregation. (64labs, 2025-11-06)
• RETHINK Retail describes the 2025-onward direction as "hybrid architectures — integrated core with modular extensibility at the edges." (RETHINK Retail, 2025-05-29)

Incisiv CSO Girijesh Agarwal and VTEX's Michael von Bodungen (Shoptalk 2025) framed the enterprise reality: the average company manages over 300 SaaS applications, and 84% of organisations cite managing cloud spend as their top cloud challenge (as-of 2025-04). (RETHINK Retail YouTube, 2025-04 — reported as session context, not a direct speaker quote)

---

Adoption statistics context

• MACH Alliance 2025 annual research (conflict of interest — Alliance member ecosystem): 87% "widely implemented" MACH; 91% increased MACH infrastructure in prior year; 61% of tech stack projected MACH-based by start of 2026. (as-of 2025 survey)
• Incisiv survey: over 43% of retailers say their CMS won't meet future needs; 42% say the same about their product catalog system. (as-of 2025; cited in RETHINK Retail, 2025-05-29)

---

Key terms

Term	Meaning
Composable Regret	Practitioner term for disillusionment post-composable adoption
Composable Extremism	VTEX's own label for dogmatic full best-of-breed MACH deployment
Pragmatic Composability	VTEX's alternative: OOTB core + selective modular extension
Integration Debt	Hidden ongoing cost of maintaining API connections across a composable stack
De-composable	64labs' term for platforms (e.g. SFCC) that offer headless optionality without requiring full disaggregation
Orchestration Layer	Middleware or API composition service that presents a composable stack as a unified output

---

What practitioners report

Beth Moriarty (VTEX, Shoptalk Spring 2025): "a lot of brands are basically failing in their approach to do a fully composable best in breed deployment" — enterprise brands need a hybrid rather than a binary choice. (RETHINK Retail YouTube, 2025-04)

64labs (Salesforce Commerce Cloud SI, 2025-11-06): "Most retailers don't have a MACH problem. They have an ROI problem, a velocity problem, and a bandwidth problem." Root causes attributed to: microservices needing specialist teams, licensing and integration cost overruns, operational burden on business users, security complexity, and architectural idealism over business outcomes.

Uniform.dev (composable personalization vendor): "A simple schema change in the PIM cascades through the commerce platform, affects search functionality, and hinders the personalization engine." (Uniform.dev blog, date unconfirmed)

---

Next frontier

• Pragmatic Composability — VTEX's alternative framework; named as the successor concept throughout this run; no dedicated page
• VTEX — entity page; 4 concept-page inbound refs; exited MACH Alliance March 2025; no page
• Strangler Pattern — migration methodology; 2+ inbound refs; no page; how teams exit composable without full replatform
• Digital Experience Platform (DXP) — Gartner 2025 MQ context; composability as mandatory criterion; no page