The conventional advice on choosing a Magento agency focuses on what to look for in a good partner. The more useful exercise, especially for retailers who have been burned by a previous engagement, is to focus on what to walk away from. The agencies that produce the largest gaps between what was sold and what was delivered usually display warning signals during the sales process itself. The skill is recognizing the signals while they are still cheap to act on.
This article is a field guide to the patterns that should disqualify an agency from your shortlist regardless of how good the rest of their pitch looks. The list is drawn from rescue projects we have inherited from other agencies, where the post-mortem usually reveals that the warning signs were present from the first sales call but were rationalized away under deadline pressure.
Red Flag 1: The Estimate Has No Visible Derivation
A confident single number on the bottom of a proposal, presented without supporting hours by phase and role, is a negotiation, not an estimate. The agency arrived at the number by working backward from what they thought you would accept, not forward from what the work actually requires. The gap between the negotiated number and the actual cost will show up later as change orders, scope arguments, or quality compromises.
Healthy estimates are decomposable. The agency should be able to show you discovery hours, architecture hours, development hours by component, QA hours, deployment hours, and project management hours. Each line item should have documented assumptions that drove the estimate. Contingency should be named explicitly rather than hidden in inflated line items.
When you ask for the derivation and the agency says “we don’t usually share that level of detail” or “trust us, we’ve done this before,” walk away. The unwillingness to defend the estimate is the warning that the estimate cannot be defended.
Red Flag 2: Named Resources That Get Replaced After Signing
The proposal names a senior architect, a lead developer, and a project manager. The contract gets signed. The kickoff meeting introduces different people. This pattern is so common in agency engagements that experienced retailers now lock in named resources contractually with substitution requiring explicit approval.
Agencies that resist named-resource contractual language are telling you that the people who pitched are not the people who will do the work. The senior architect who wrote the proposal is on three other projects. The lead developer is fictitious or rotated. The actual team is a tier below what was sold.
The disqualifying signal is not that the agency cannot guarantee the original team forever – circumstances change. The disqualifying signal is that the agency will not commit in writing to the original team for the foreseeable future and will not commit to comparable substitutes if changes become necessary.
Red Flag 3: Generic Answers to Technical Questions
The discovery conversation should include specific technical questions about your codebase, integrations, and architecture. The agency’s answers should be specific in return. When you describe your custom checkout, a good agency will ask about your payment integrations, your tax calculation logic, and your fulfillment workflow integration. When you describe your B2B implementation, a good agency will ask about company hierarchies, approval chains, and ERP sync patterns.
Generic answers – “we’d handle that during discovery” or “every project is different so we’d need to scope it” – are the agency telling you that they do not have a strong pattern recognition for your problem space. Good agencies have seen your problem before and have opinions about it. Agencies that speak in generalities have either not seen the problem or have not learned from seeing it.
Red Flag 4: Hyvä, B2B, or Headless Depth That Cannot Be Substantiated
If the agency claims Hyvä expertise, ask for three recent migrations with module compatibility decisions, checkout customizations rebuilt, and CWV measurements before and after. If they claim Adobe Commerce B2B depth, ask for a recent ERP integration with company hierarchy design, shared catalog logic, and approval workflow configuration. If they claim headless or PWA Studio depth, ask which production sites they have shipped and how they handled SEO, personalization, and content management.
The agencies with real depth will give you specific stories with named modules, real numbers, and clear opinions about trade-offs. The agencies without depth will either refuse to share specifics (“we can’t disclose client details”) or speak in generalities that could apply to any project. This is the cleanest filter for separating actual specialists from generalist shops that have added Hyvä or B2B to their marketing copy without building actual practice depth.
Red Flag 5: References That Sound Coached
Agency references are positive by design – the agency picks references it expects to give favorable reviews. The valuable information is in the texture of the answers, not the headline sentiment. Coached references give you smooth narratives with no friction, generic positive adjectives, and reluctance to discuss specifics. Authentic references give you texture, discuss disagreements that were resolved constructively, and have specific opinions about strengths and weaknesses.
The diagnostic question that surfaces coaching most reliably is “what didn’t go well, and how was it handled?” Coached references will say “everything went great.” Authentic references will describe a specific issue, the agency’s response, and the resolution. The agency that consistently provides coached references is signaling that authentic references would not give favorable reviews.
Red Flag 6: Subcontracting Without Disclosure
A pattern that is increasingly common in mid-market Magento engagements is agencies that subcontract specialized work to offshore partners without disclosing the arrangement. The pitch is from a domestic boutique agency. The actual development is happening in another country at a different firm. The retailer thinks they are paying for the boutique and is actually paying a markup over offshore rates.
The pattern is not inherently disqualifying if it is disclosed, scoped, and quality-controlled. The disqualifying behavior is concealing the arrangement. Ask explicitly: “Who will be writing code in our repository? Are any of them employees of subcontractor firms? If so, who are those firms and what is their role in our project?” Agencies that answer transparently can be evaluated. Agencies that obfuscate the answer should be walked away from.
Red Flag 7: Documentation and Knowledge Transfer Are Out of Scope
Healthy Magento agencies include documentation and knowledge transfer as explicit project deliverables. Architecture decision records, integration documentation, deployment runbooks, and onboarding materials are part of what they ship. The knowledge transfer is structured around phase boundaries so your team builds context as the project progresses, rather than being dependent on the agency forever.
| Knowledge Transfer Signal | Healthy Pattern | Walk-Away Signal |
|---|---|---|
| ADR maintenance | ADRs committed alongside code | Decisions exist only in Slack |
| Integration docs | Auth flows, schemas, error handling | Tribal knowledge only |
| Runbook delivery | Updated runbook per release | No runbooks, “we handle ops” |
| Onboarding readiness | Your devs productive in days | Permanent dependency on agency |
| Phase boundary KT | Formal handoff sessions | Knowledge transfer is “out of scope” |
| Code ownership clarity | Your IP, your repo, your CI | Agency-owned tooling, lock-in |
When documentation and knowledge transfer are “out of scope” or “available as a separate engagement,” the agency is structurally building dependency. They benefit from your inability to switch. The relationship that starts as a partnership ends as a hostage situation.
Red Flag 8: A Single Pricing Model Forced on a Mismatched Project
Healthy agencies use the pricing model that fits the work. Fixed-price for well-defined scope. Time-and-materials for discovery-heavy work. Hybrid structures for engagements with a mix. Retainers for ongoing maintenance and support.
Unhealthy agencies push a single pricing model regardless of fit. The fixed-price agency that quotes fixed-price for a discovery-heavy replatform is setting up an engagement that will produce constant change-order friction. The T&M agency that resists fixed-price for well-defined component work is signaling they do not want accountability for delivery. The agency that pushes a retainer for project work is selling utilization rather than outcomes.
The walk-away signal is rigidity in pricing model. The agency that will not adapt their commercial structure to your project’s risk profile is going to be similarly rigid in execution.
Red Flag 9: No Clear Position on Platform Trade-Offs
Mid-market Magento decisions are increasingly comparative. Adobe Commerce vs Shopify Plus. Hyvä vs PWA Studio. Adobe Commerce B2B vs Salesforce Commerce Cloud B2B. The healthy agency has opinions about which option fits which retailer profile and is willing to recommend against their own services when that is the right call.
The unhealthy agency claims neutrality and will build whatever you want. That posture sounds client-friendly but is actually a warning signal. The agency that does not have opinions about platform fit is either inexperienced enough not to have developed them or is unwilling to risk losing the engagement by being honest. Either way, you are not getting the strategic input that distinguishes a real partner from a builder-for-hire.
Bemeir’s typical engagement starts with a candid platform conversation. If the retailer is better served by Shopify Plus than by Adobe Commerce, we say so – that recommendation has produced more long-term relationships than it has lost individual projects. The willingness to deliver an uncomfortable recommendation is a leading indicator of the partnership’s long-term value.
Red Flag 10: The Sales Process Itself Is Disorganized
The sales process is the agency’s best foot forward. If the sales process is disorganized – missed deadlines on proposal delivery, contradictory information from different people, unanswered emails, last-minute scope changes – the delivery process will be worse.
The diagnostic is to watch how the agency handles the sales process itself. Are deadlines met? Is communication consistent across the people you interact with? Are commitments honored? Does the agency push back on unrealistic timelines from your side rather than agreeing to everything to win the work?
An agency that cannot run a clean sales process is showing you what delivery will look like. Walk away while the cost of walking away is just the time invested in evaluation.
How to Use This List
The presence of one red flag is informative but not always disqualifying – every agency has imperfections. The presence of three or more in combination, however, is almost always disqualifying because the patterns compound during the engagement.
The retailers who run rigorous evaluation processes find that the shortlist of qualified Magento agencies is shorter than they expected. The depth required to handle enterprise Magento engagements, Hyvä migrations with real CWV outcomes, and Shopify Plus B2B development lives in a handful of boutique specialists. Bemeir is one of them; we acknowledge there are peers worth comparing us against. The comparison among genuine specialists is the right comparison to make rather than comparing specialists against generalist shops that lack the depth your project requires.
Further reading on agency evaluation discipline includes the Adobe Commerce Solutions Partner directory, the Hyvä partner program, and structured procurement frameworks from Gartner and Forrester.
