Adobe Commerce Partner: How DWAO Builds Stores That Hold Up on Their Worst Day

There's a particular silence that falls over a company whose checkout is down during their biggest promotion of the year. We've been the team called into that silence — sometimes to fix our own predictions, more often to rescue an estate built by someone who treated peak season as a marketing event instead of an engineering deadline. Either way, the lesson transfers: commerce platforms are judged on their worst hour, not their average, and everything about choosing an Adobe Commerce partner flows from that sentence.

Adobe Commerce — the platform the industry still half-calls Magento — is deeply capable and famously tolerant of bad decisions. It will let a partner customize anything, integrate everything and defer any hard question until it arrives with interest. This guide covers what a Commerce partner should actually own, the traps written into the platform's flexibility, the replatforming realities, and how we run Commerce work at DWAO.

What an Adobe Commerce Partner Actually Owns

The role, taken seriously, spans five territories:

Commerce architecture. Catalog structure, pricing and promotion logic, checkout flow, B2B company hierarchies where they apply — designed against your merchandising reality, with the discipline to prefer configuration over code. Attribute decisions made casually in week three ripple through search, integrations and performance for years.

Integration engineering. A store is the visible tip of a system: ERP for products and pricing, OMS for orders, payments, tax, shipping, PIM, marketing. Integration is routinely half a Commerce project's genuine effort and nearly all of its schedule risk — third parties have their own calendars. Partners who leave integration for "phase two" are scheduling your crisis.

Performance and peak readiness. Load testing against realistic traffic models (your peak, not an average day), caching architecture that actually hits, CDN strategy, and a rehearsed peak-season runbook: freeze calendars, failover plans, on-call rotations. This is annual discipline, not launch-week theater.

Security operations. Commerce platforms hold payment-adjacent data and attract attention accordingly. Adobe ships security patches on a cadence; applying them promptly with regression cover is table stakes. An estate with patch debt is a breach with a countdown timer — we've audited stores where this line alone justified the engagement.

Conversion evolution. The reason the store exists: checkout friction hunted, search relevance tuned, page speed defended, and an optimization backlog that treats revenue-per-session as the product. Stores that stop evolving decay commercially long before they decay technically.

A Commerce partner covering all five is operating a revenue platform. One covering two is building you a liability with a beautiful theme.

The Customization Tax (and Why Restraint Is the Rarest Skill)

Adobe Commerce's extension ecosystem and code-level openness are genuine strengths — and the platform's most reliable failure mode. The pattern is always the same: each customization individually reasonable, the sum unmaintainable. Sixty modules, three doing overlapping jobs badly, checkout logic from 2019 nobody dares touch, and every upgrade quoted like a rebuild.

Our position, which occasionally costs us billable hours and reliably earns us renewals: configuration first, extensions vetted like dependencies you're marrying, custom code only where your business genuinely differs — and argued in writing against its ten-year cost. Most "requirements" for custom code are habits imported from the old platform. A partner who says yes to everything is not agreeable; they're compounding your future maintenance bill at your expense.

The same restraint applies to the build-versus-buy question on extensions: a marketplace module you'd heavily modify is usually worse than clean custom code, and clean custom code is usually worse than adapting the process to configuration. That hierarchy — process, configuration, extension, custom — is the sort of thing to probe when evaluating partners. Ask where their last three projects landed on it.

Replatforming to Adobe Commerce Without the Classic Wounds

A large share of Commerce engagements are replatforms — from Magento 1 remnants, from legacy carts, from platforms that stopped fitting. Three wounds account for most replatform scar tissue, and all three are preventable:

The SEO faceplant. Organic rankings are an asset built over years and donated to competitors in one careless cutover. URL mapping, redirect chains, metadata parity and structured-data continuity are engineering deliverables with acceptance criteria — validated by crawl before launch, not diagnosed by traffic collapse after. We treat pre/post organic parity as a launch gate.

Data migrated, trust not. Catalog, customers and orders move with reconciliation counts at every stage — and history decisions made explicitly. (Do seven-year-old orders belong in the new store, or in an archive? That's a business decision to make deliberately, not a migration script default.)

The heroic cutover. One big weekend, no rollback, prayers. The alternative is boring and correct: staged migration, a rehearsed cutover in a low-traffic window, rollback paths tested, and the old platform kept warm until the first clean close. Boring launches are the ones nobody remembers — which is the goal.

How DWAO Runs Commerce Engagements

Builds and replatforms. Implementation runs integration-contracts-first: ERP/OMS/payment interfaces drafted and stubbed in the opening weeks, so the riskiest work gets the most calendar. Catalog architecture is workshopped with merchandisers, checkout customization faces a standing burden of proof, and load testing begins mid-build against your peak model. Launch is rehearsed, gated on SEO parity, and followed by hypercare sized to real risk.

Ongoing operations. Managed Commerce means patch currency under SLA with regression cover, performance watched against budgets with alerts before customers notice, extension estate reviewed at every platform update, and a conversion backlog shipped in sprints. Peak readiness is an annual program with your name on the calendar: load tests against forecast, failover rehearsals, freeze schedule agreed with marketing before November argues about it.

Support, triaged by revenue. A checkout error is all-hands now; a CMS block typo queues. Our severity definitions are written in commerce terms — order impact, payment impact, catalog impact — because generic IT severity language fails exactly when it matters. Post-incident, fixes land with monitoring that would have caught the issue earlier; over a year, support becomes an early-warning system rather than a fire brigade.

The structural facts. Adobe Gold Partner. Certified Commerce practitioners named in the contract — you interview them, they stay. Delivery teams across five countries for follow-the-sun capacity during builds and extended coverage during peaks, with your working hours as the anchor. And the restraint doctrine above, in writing, on every project.

B2B Commerce: The Same Platform, a Different Game

Adobe Commerce's B2B capabilities — company accounts, shared catalogs, negotiated pricing, requisition lists, quote workflows — turn the platform into something structurally different from a consumer store, and partner experience doesn't automatically transfer across that line. Worth understanding before you evaluate anyone, including us:

The data model deepens. B2C thinks in customers; B2B thinks in companies containing buyers with roles, approval chains and credit terms. Catalog and pricing become per-account realities — the same SKU legitimately has as many prices as you have negotiated contracts, and the architecture has to serve that without melting search or cache.

The integration center of gravity moves. In B2C, integration risk concentrates in payments and fulfillment. In B2B it moves upstream into ERP: contract pricing lives there, credit limits live there, punch-out and EDI relationships live there. A B2B commerce project is substantially an ERP integration project wearing a storefront — plan and staff it accordingly.

The buyer experience bar is different, not lower. B2B buyers tolerate density but not friction: fast reorder from history, CSV upload to cart, requisition workflows that mirror their procurement reality. "Consumer-grade UX" is the slogan; procurement-grade efficiency is the actual requirement.

Our B2B engagements staff ERP-fluent integration engineers alongside commerce specialists from day one, and our discovery includes your customers' purchasing teams where clients allow it — because the highest-converting B2B features we've shipped came from watching a buyer struggle through a reorder, not from a requirements workshop. Ask any prospective partner for their B2B-specific references; the data model differences are large enough that B2C-only evidence answers a different exam.

The First 90 Days on an Inherited Store

Since much of our Commerce work begins with an estate someone else built, here's the arc we run — useful as a preview of working with us, and as a template for auditing any takeover proposal:

Days 1–20: the risk-ranked audit. Patch currency first (this finding alone can reorder everything), then customization sprawl mapped module by module, integration robustness probed at the failure modes (what happens when the ERP is slow? when the payment provider times out?), performance baselined against your real peak history, and checkout walked like a paranoid customer. Every finding lands with a verdict — fix now, schedule, watch, leave alone — and a revenue-exposure rank. No fifty-page report; a ranked backlog with owners.

Days 20–60: stabilization. Security debt cleared under regression cover. The two or three highest-exposure findings fixed — typically a checkout edge case, an integration failure mode, a cache misconfiguration quietly costing speed. Monitoring installed where the audit found blind spots, because the next incident should announce itself to us before it announces itself to customers.

Days 60–90: the operating rhythm. Patch cadence under SLA. The conversion backlog groomed and shipping in sprints. Peak readiness scheduled against your calendar — load tests, failover rehearsals, the freeze conversation with marketing while it's still theoretical. And the honest capacity conversation: what your team owns, what we carry, what the store actually needs from a managed service versus targeted support.

Ninety days in, an inherited store should be measurably safer, observably faster, and boringly predictable. Boring, in commerce operations, is the luxury good.

Budgeting Commerce Honestly

What actually moves cost, before any vendor conversation: catalog and brand complexity (SKU volume, multi-store, B2B hierarchies), integration count (the biggest single swing — each system adds design, build and test), migration scope (clean catalogs move fast; neglected ones bill archaeology), customization appetite (the tax discussed above), and peak profile (a store with 10× seasonal spikes needs engineering a flat-traffic store doesn't).

Adobe's licensing is GMV-indexed and negotiated — our Commerce pricing guidance covers those mechanics, including the forecasting discipline that prevents both true-up surprises and prepaid optimism. For delivery, a scoping conversation against your drivers produces a genuine range, free; and if the honest answer is that your requirements don't need Adobe Commerce, we'll say that too — a store on the wrong platform is a client we'd eventually lose anyway.

The Store as a Data Asset (Commerce Doesn't End at the Order)

One more territory separates complete Commerce partners from storefront builders: what happens to the data your store generates. An Adobe Commerce estate produces the richest behavioral stream most companies own — browse paths, search queries, cart compositions, purchase histories — and in a distressing number of implementations, it just… accumulates.

The connected version looks different, and it's where Adobe's stack argument gets concrete for commerce:

Search and merchandising intelligence. Site search terms are your customers dictating merchandising strategy in their own words. Zero-result searches are a product-gap report; high-exit searches are a relevance bug report. We wire search analytics into the optimization backlog as standard — it's the cheapest conversion research that exists.

Analytics that survive the funnel. Commerce feeding Adobe Analytics properly — product-level, promotion-level, funnel-staged — turns "revenue is down" conversations into "PDP-to-cart fell for this category after that template change" conversations. The integration is a scoped deliverable in our builds, with parity checks, because commerce data that almost matches finance data convinces nobody of anything.

Audiences that leave the store. Purchase and browse behavior powering Target experiences and Journey Optimizer messaging — the browse-abandonment journey, the replenishment reminder, the loyalty-tier experience — is where the store stops being a channel and becomes the behavioral heart of the marketing estate. It requires the identity spine to be designed, which is why our commerce discovery asks CDP questions that surprise procurement.

None of this obligates a big-bang stack program; it obligates a partner who builds the store with these doors unlocked rather than bricked over. Retrofitting event tracking and identity into a store built data-blind costs multiples of designing it in — one more decision that gets made, visibly or invisibly, in the first month of your implementation.

The Worst-Hour Test

Every evaluation framework in this market eventually reduces to one question, so we'll leave you with ours: "Tell us about your worst peak-season hour, and what changed because of it."

Partners with real commerce scars answer specifically — the incident, the diagnosis under pressure, the runbook line that exists because of it. Partners without the scars haven't operated at stakes, and your Black Friday shouldn't be their education. We're happy to share ours — the incidents, the runbook lines they wrote, and the boring, rehearsed peaks that followed. Ask us; a senior Commerce consultant will respond within one business day.