Headless Commerce for Agencies: Complete 2026 Guide

Headless commerce is one of the most consequential decisions a mid-market brand will make in the next 24 months, and one of the most consequential offerings an ecommerce agency will add to its menu. Done well, a headless engagement is a $150K to $500K build that turns into a $20K-per-month retainer for three to five years. Done poorly, it is the project that costs you a senior engineer, a key client, and a piece of your reputation. The difference is almost never about technology. It is about scoping, pricing, staffing, and saying no to the wrong deals. This headless commerce agency guide is for owners and ops leads who already know how to ship Next.js or Hydrogen and now need to understand the business mechanics: how to scope a headless build, how to price it for margin, what retainer structure pays for itself, what team you actually need, and which RFPs to walk away from.

Key Takeaways:

• Headless commerce builds typically land between $120K and $500K in 2026, with retainers from $10K to $35K per month.
• The most profitable architecture in 2026 is Next.js (or Astro/Remix) on Vercel with Shopify or BigCommerce as the commerce engine.
• Scope creep on headless engagements is almost always content-system-related, not commerce-engine-related.
• Healthy headless practices carry 50 to 65 percent of revenue through retainers within 24 months.
• The single biggest predictor of profitability is whether the client has an in-house front-end engineer who can co-own the codebase post-launch.

When Headless Actually Pays (And When It Doesn't)

Before discussing pricing or team structure, the qualifying question for every headless RFP is whether the buyer has a real reason to go headless beyond CMO enthusiasm. According to recent Gartner research on composable commerce and the MACH Alliance 2025 industry survey, the brands that succeed on headless share five traits:

1. They publish to more than one surface (site plus mobile app, in-store kiosk, marketplace, or social storefront).
2. Their conversion rate is measurably affected by site speed; the team can quote LCP and INP without prompting.
3. They have a content strategy that exceeds what a SaaS theme can support (long-form editorial, custom landing patterns, programmatic SEO).
4. They have, or are willing to hire, an in-house senior front-end engineer.
5. Their roadmap justifies the ongoing cost of two stacks (front-end and commerce engine) instead of one.

If a client cannot articulate at least three of these, headless is the wrong answer. The agency that signs the deal anyway will absorb the post-launch maintenance cost for years.

A reliable disqualifier matrix:

| Signal | Probably headless | Probably not headless | |---|---|---| | Annual GMV | $15M plus | Below $5M | | Surfaces published to | 2 plus | Web only | | In-house FE engineer | Yes | No | | Reason for going headless | Performance, omnichannel, content depth | Competitor has one | | Post-launch retainer budget | $10K plus per month | "We will figure it out" | | Project timeline | 16 to 30 weeks | "ASAP" |

The most common failure mode is a $7M apparel brand whose CMO read a Shopify Hydrogen case study, buys a $250K Next.js build, and then cannot maintain it. The agency ends up giving away 200 to 400 hours of post-launch support to protect the case study, which destroys the project's economics.

Architecture Choices That Pay in 2026

Three architecture patterns account for roughly 80 percent of profitable headless engagements:

1. Next.js + Shopify (Plus or standard). The default 2026 stack. Shopify Storefront API handles commerce; Next.js on Vercel handles front-end; Sanity, Contentful, or Storyblok handles content; Algolia handles search. Predictable, well-documented, large engineer pool.
2. Next.js + BigCommerce. The same shape, but better suited to B2B and multi-channel. BigCommerce's API rate limits are higher and B2B Edition pairs cleanly with custom front-ends.
3. Shopify Hydrogen + Oxygen. Shopify-native headless. Best when the brand wants Shopify to own both ends of the stack and the agency can commit to the Hydrogen specialty.

Less common but viable: Astro + Shopify (for content-heavy brands with simple commerce), Remix + Shopify (for teams that prefer the Remix mental model), and composable stacks (Commercetools, fabric, Elastic Path) for true enterprise builds above $50M GMV.

What does not pay in 2026:

• Custom-built commerce engines. Almost never. The savings on licensing rarely offset the cost of ongoing engineering.
• Headless WordPress / WooCommerce. The maintenance burden compounds; better to recommend a SaaS migration as part of the engagement.
• Multiple front-end frameworks in one codebase. A "let's use Vue for the PDP and React for the cart" architecture is a future support nightmare.

Document the architecture choice in the SOW with a one-page rationale. If the client wants something off this list, scope a discovery sprint before signing the build.

Pricing a Headless Build

Realistic 2026 pricing for headless engagements:

| Scope | Budget | Timeline | Team | |---|---|---|---| | Headless rebuild of existing Shopify store | $120K to $220K | 14 to 22 weeks | 1 senior FE, 1 mid FE, 1 PM, 1 designer | | Next.js + BigCommerce with custom CMS | $180K to $320K | 18 to 26 weeks | Above + 1 BE / integrations engineer | | Composable build (multi-region, multi-channel) | $300K to $500K plus | 24 to 36 weeks | 5 to 7 person team plus client co-team |

A scope-estimation model that works:

• Base front-end (PLP, PDP, cart, checkout shell). 280 to 420 hours.
• Content system integration (CMS schemas, preview, draft mode). 80 to 160 hours.
• Search (Algolia, Typesense, or Klevu). 60 to 120 hours.
• Personalization layer. 40 to 100 hours.
• Analytics, GA4, server-side tracking. 40 to 80 hours.
• Performance budget enforcement. 30 to 60 hours.
• Accessibility (WCAG 2.2 AA). 60 to 120 hours, often more.
• DevOps and observability (Vercel, Datadog, Sentry). 30 to 60 hours.
• Migration (redirects, SEO, image pipeline). 60 to 160 hours.

Total a reasonable Tier 1 headless build at 800 to 1,100 billable hours, blended at $150 to $220 per hour. That is your floor. Anything below that price means you are eating the difference, which most agencies discover three weeks before launch.

For the actual SOW math, see our project pricing calculator and scope of work generator.

Retainer Structure: The Reason To Take the Deal

A headless build without a retainer is a one-time revenue event that ages badly. The codebase needs maintenance, the dependencies need upgrades, and the brand will want new templates and features. Three retainer SKUs work:

| Retainer tier | Monthly fee | Hours / scope | Best for | Renewal rate | |---|---|---|---|---| | Headless care plan | $4K to $9K | Dependency upgrades, monitoring, minor edits, monthly Core Web Vitals review | Post-launch $10M to $25M brands | 80 to 90 percent | | Growth retainer | $10K to $20K | Above plus 30 to 60 hours of dev, CRO, A/B testing | $20M to $60M brands | 70 to 85 percent | | Embedded team | $20K to $35K plus | 2 to 4 FTE equivalents, weekly priority slot | $50M plus brands with active roadmap | 60 to 75 percent |

Two operating rules protect retainer margin:

1. Never absorb dependency upgrades into a flat fee. Major Next.js, React, and CMS upgrades are 40 to 120 hour events. Quote them as planned overage, billed separately.
2. Define "support" tightly. Bug fixes within the warranty window are free; new features always trigger a change order. Without this discipline, every retainer becomes an unbounded scope.

Healthy headless agencies build their recurring revenue mix so retainers cover at least 75 percent of senior front-end payroll within 18 months of the practice's first launch.

Team Structure for a Headless Practice

The most consequential staffing decision in a headless practice is whether to centralize front-end senior talent or distribute it across pods. Both work; the wrong choice for your size kills the practice. A workable 12-person headless agency:

• 1 partner / commercial lead
• 1 head of delivery
• 2 senior front-end engineers (React, Next.js, Hydrogen)
• 2 mid-level front-end engineers
• 1 senior full-stack lead (back-end and integrations)
• 1 integrations engineer (Node, Python)
• 1 senior PM (running 1 to 2 active builds, 4 to 6 retainers)
• 1 mid-level PM (running retainers)
• 1 designer / UX lead
• 1 QA + accessibility specialist

This shape supports roughly $2.5M to $3.5M in revenue at 50 to 60 percent gross margin with a retainer-heavy mix. The most common mistake is hiring a third front-end engineer before hiring an integrations specialist; on headless engagements, the integration surface is what creates scope creep, and an integrations specialist closes more deals than they cost.

For staffing ratios, anchor on the team utilization calculator and run quarterly capacity reviews against retainer commitments.

Common Scope-Creep Traps

Headless engagements have a small number of scope-creep traps that account for almost every margin-destroying build the author has reviewed:

1. The CMS schema rewrite. The client decides the CMS schemas need to be "redesigned" three weeks before launch. Always lock CMS schemas at week four.
2. Personalization and segmentation. "Can we just show different hero banners by source?" sounds free; it is 40 to 80 hours.
3. Server-side tracking. GA4, Meta CAPI, server-side GTM. Almost always quoted thin and almost always 60 to 120 hours.
4. Internationalization. Adding a second locale post-SOW is roughly 80 to 160 hours including content workflows.
5. Image pipeline replacement. Migrating from Shopify CDN to Cloudinary or imgix mid-build is a 40 to 100 hour task.
6. Pre-launch content surprises. Clients add 50+ new PDPs in the last two weeks. Lock the catalog at week six.
7. Accessibility audits. Mid-build accessibility findings can add 80 to 200 hours of remediation if the team wasn't building to WCAG 2.2 AA from week one.

The cleanest defense is a one-page change-order policy referenced in every SOW. Any request not listed in the SOW gets a 24-hour estimate and the client signs before work starts. See our SOW vs. MSA guidance for the actual clauses.

Discovery Sprint: The Pre-Engagement That Pays

Almost every profitable headless engagement starts with a paid discovery sprint, not a free RFP response. A workable sprint:

• Length: 2 to 4 weeks.
• Fee: $12K to $30K, applied against the build if the client proceeds.
• Outputs: Architecture diagram, hour-level estimate, integration inventory, performance budget, accessibility baseline, content-system schema, retainer proposal.

Three reasons discovery sprints pay:

1. They qualify the buyer. Clients unwilling to pay for discovery are almost never willing to pay properly for the build.
2. They de-risk pricing. A real estimate built from real discovery is 60 to 80 percent more accurate than an RFP response written in two days.
3. They generate revenue without commitment. The agency makes money on every discovery sprint even when the client does not proceed with the build.

Agencies that introduce paid discovery typically see their build close rate triple within six months, because the only buyers who pay for discovery are the ones who are actually buying.

Profitability Benchmarks

| Metric | Healthy range | Warning sign | |---|---|---| | Project gross margin | 40 to 55 percent | Below 30 percent | | Retainer gross margin | 55 to 70 percent | Below 45 percent | | Average retainer per client | $9K to $18K per month | Below $6K | | Retainer client LTV | $200K to $500K | Below $120K | | Annual revenue per FTE | $200K to $280K | Below $160K | | Senior FE utilization | 60 to 75 percent | Below 50 or above 85 percent | | Discovery-to-build conversion | 60 to 80 percent | Below 40 percent |

The most informative number on this list is senior front-end utilization. Below 50 percent and you are over-staffed; above 85 percent and you are about to lose a senior engineer to burnout or to a competitor. Tracking utilization weekly through a time tracking system for agencies is non-negotiable.

Anonymized Scenario: A 10-Person Headless Agency in Brooklyn

A Brooklyn-based headless agency we have benchmarked grew from $1.6M to $4.2M in annual revenue between 2023 and 2025 through four operational changes:

• Introduced a paid discovery sprint at $18K. Discovery-to-build conversion moved from 32 percent (free RFP) to 71 percent (paid discovery). Build close rate effectively doubled.
• Standardized on Next.js + Shopify with Sanity for content and Algolia for search. Refused to bid composable enterprise builds for the first 18 months. Internal velocity rose 40 percent because no one was learning new tools mid-project.
• Restructured headless care plans into three SKUs ($5K, $11K, $22K per month). Stopped quoting bespoke retainers. Average retainer grew 44 percent in a year.
• Built a public-facing case study process: every shipped build became a case study within 30 days. Inbound pipeline from organic search and partner referrals tripled.

P&L outcome: retainer share moved from 24 percent to 56 percent, EBITDA went from 9 percent to 21 percent, and average client LTV nearly tripled. The owner reported that the most consequential change was introducing paid discovery; everything else followed.

Productized Service Lines Worth Building

Six productized headless offerings that consistently sell and run profitably:

1. Headless Discovery Sprint ($15K to $25K, 3 to 4 weeks): Architecture, estimate, integration inventory.
2. PLP/PDP Migration to Next.js ($35K to $70K): Migrate the highest-traffic templates only, leave checkout on Shopify.
3. Performance and Core Web Vitals Sprint ($10K to $20K): LCP, INP, CLS focus across an existing headless build.
4. Accessibility Remediation Sprint ($12K to $30K): WCAG 2.2 AA remediation with documentation.
5. CMS Schema Rebuild ($15K to $35K): Redesign Sanity, Contentful, or Storyblok schemas for performance and editorial UX.
6. Personalization Pilot ($20K to $45K): Server-side personalization on hero, PLP, and PDP with measurement infrastructure.

Productized services drive agency profitability because they collapse the discovery cycle, predict utilization, and let junior PMs run them. Pair them with a productized service catalog so buyers can self-select.

Tooling and Operations Stack

A workable 2026 stack for a headless agency:

• Code + deploy: GitHub, Vercel, Netlify, Cloudflare Pages.
• Observability: Sentry, Datadog, Vercel Analytics, SpeedCurve.
• CMS: Sanity, Contentful, Storyblok, Hygraph.
• Commerce: Shopify, BigCommerce, Commercetools (enterprise only).
• Agency ops: Time tracking, retainer billing, and capacity planning bound together in an agency management platform. The analytics dashboard is non-negotiable once you have more than five active retainers.
• Client comms: A central client portal so requests do not live in email.

The single most consequential tooling investment for a headless agency is the system that ties sold scope to delivered hours per retainer and per project. Without it, you will quietly absorb 50 to 100 hours per build by week 16.

When To Walk Away from a Headless RFP

A short list of disqualifiers:

• The brand is below $5M GMV and headless is being chosen for status, not need.
• The client has no in-house front-end engineer and no plans to hire one.
• The decision-maker will not attend any discovery call.
• The brand wants a fixed bid before any technical discovery.
• The CMO insists on a specific framework you don't use.
• The procurement team will not commit to a 12-month retainer post-launch.

Saying no is a profitability strategy. Most headless agencies could lift EBITDA five to nine percentage points by declining the bottom 20 percent of their pipeline. The healthier the pipeline, the easier this decision becomes, which is why a steady agency lead generation engine is a profitability lever as much as it is a growth lever.

FAQ

How long does a typical headless build take?

A standard Tier 1 headless build (Next.js + Shopify) runs 14 to 22 weeks from kickoff to launch with another four to six weeks of post-launch stabilization. Tier 2 builds with custom CMS and additional integrations extend to 18 to 26 weeks. Composable enterprise builds regularly run 24 to 36 weeks.

What gross margin should a headless agency target?

A healthy blended gross margin is 50 to 60 percent across builds and retainers, with retainers above 60 percent and projects at 40 to 50 percent. Headless project margins are lower than templated builds because of the engineering complexity; the retainer is where the margin compounds.

Should we offer Hydrogen vs. Next.js?

Both are viable. Hydrogen is Shopify-native, slightly faster to deploy on Shopify-only stacks, and tightly coupled to Shopify's roadmap. Next.js is more flexible, has a vastly larger engineer pool, and pairs with any commerce engine. Most independent agencies should standardize on Next.js + Shopify or Next.js + BigCommerce; agencies pursuing Shopify Hydrogen Partner status may choose Hydrogen.

How big should the team be on a headless build?

A Tier 1 headless build needs a minimum of one senior front-end engineer, one mid-level engineer, one PM, and a designer. Tier 2 adds an integrations engineer and a back-end developer. Smaller teams take longer, ship buggier, and burn more senior time on stabilization.

How do we sell headless to a sceptical client?

Don't. The buyers worth selling to already understand the upside and just want a credible partner. Sell to those buyers and walk away from the rest; the sceptical buyer becomes the unprofitable client.

Closing

Headless commerce is one of the highest-LTV engagements in agency commerce, but only for the agencies that productize their discovery, lock their architecture stack, and refuse to sign builds without a retainer. Everyone else is a freelance team with a Vercel account.

If you are scaling a headless practice and want to see how AgencyPro helps you run capacity, retainers, and profit-per-client in one place, book a demo and we will walk through the metrics that move first.