Most Magento agency RFPs measure the wrong things. Slide deck quality, client logo lists, and case-study word counts do not predict whether the agency will actually ship a stable Adobe Commerce site on time. The questions that do predict outcomes are technical, specific, and uncomfortable enough that weak agencies cannot fake their way through them. If you are an in-house Magento developer or a CTO running an agency search in 2026, here is the twelve-question screen we recommend.
These are questions Bemeir’s Magento team hears in serious procurement processes and asks of the engineers we hire. They are also the questions we would expect a buyer to ask us, and we have learned to answer them with specifics rather than generalities. The goal is to separate agencies that have actually shipped production Adobe Commerce work from agencies that have a clean website and good salespeople.
The questions
1. “Walk me through the last production incident your team handled on a client’s Magento store.”
What you are checking: whether the agency has a production support muscle, not just a project delivery muscle. Strong agencies will describe a specific incident with specific symptoms, the diagnostic steps they took, the fix, and the post-mortem changes. Weak agencies will deflect with “we have a 24/7 support team” without specifics.
Red flag: any answer that doesn’t include database queries, error logs, or a specific code change.
2. “Show me a Hyvä migration you shipped in the last 12 months and what the Core Web Vitals looked like before and after.”
What you are checking: real Hyvä experience, not certificate-level familiarity. The agency should be able to produce PageSpeed Insights screenshots or Chrome User Experience Report data from a real client (with client permission), showing the actual before/after Lighthouse and Core Web Vitals numbers.
Red flag: marketing language about “significant improvements” without the actual metrics.
3. “Tell me how your team manages Magento security patches across multiple clients.”
What you are checking: ongoing maintenance discipline. Strong agencies will describe a specific process: how they monitor Adobe Security Bulletins, how they evaluate severity, how they coordinate patch testing across clients, how they decide on emergency vs scheduled deployment, and what the SLA is per severity tier.
Red flag: “we patch when Adobe releases” without process detail, or “the client handles patching.”
4. “What is your code review process for production deploys?”
What you are checking: engineering discipline. Strong answers describe a multi-stage process: pull request templates, mandatory reviewer assignment, automated linting and static analysis, manual review, and a separate QA approval gate. Weak answers describe ad hoc review or “the senior developer signs off.”
Red flag: any answer suggesting code can hit production without review.
5. “Walk me through how you handle a third-party Magento extension that turns out to be poorly written.”
What you are checking: technical depth and pragmatism. The right answer involves a specific decision framework: assess whether the extension can be patched in place, fork it into a maintainable in-house version, replace it with a different extension, or rebuild the functionality natively. Agencies that have done this know which extensions to avoid by name.
Red flag: agencies that recommend extensions without naming any specific vendors they avoid.
6. “Show me how your team approaches Magento performance optimization beyond just enabling cache.”
What you are checking: depth in performance work. Strong answers cover full-page cache hit ratio analysis, database query profiling with the Magento Profiler or New Relic, image optimization pipelines, JavaScript bundle analysis, Varnish tuning, Redis configuration, third-party script governance, and Core Web Vitals field data monitoring through tools like the Chrome User Experience Report.
Red flag: any answer that stops at “we enable full-page cache and CDN.”
7. “What is the senior-to-junior ratio of the team that would actually work on our project?”
What you are checking: who will do the work, not who is in the proposal. Many agencies pitch with senior architects and deliver with junior developers. The ratio should be no worse than 1 senior to 2 juniors, and the senior should be billed at meaningful hours on the project, not just at kickoff and launch.
Red flag: vague answers about team composition, or any answer that puts a senior on 5% of project hours.
8. “How do you handle Magento upgrades, including the upgrade from current Magento version to the next major version?”
What you are checking: lifecycle thinking. Strong agencies have a specific upgrade methodology, can name the most common breaking changes between recent Magento versions, and have a published process for handling upgrades on existing clients. They know about Adobe’s release cadence and the Adobe Commerce release roadmap.
Red flag: agencies that have never done a major version upgrade for a client.
9. “Tell me about a project that went badly and what you learned from it.”
What you are checking: self-awareness and honesty. Strong agencies have a specific story with specific lessons that changed how they work. Weak agencies will either claim they never had a bad project or describe a bad project that was entirely the client’s fault.
Red flag: any answer that does not include something the agency would do differently next time.
10. “What is your approach to documentation for the work you ship?”
What you are checking: long-term partnership viability. Strong agencies produce three kinds of documentation: technical handover documentation for the client’s engineering team, runbook documentation for ongoing operations, and architecture documentation for future enhancement work. Weak agencies produce none of these and the client is locked in by inability to switch.
Red flag: documentation is “in the code” without anything written for humans.
11. “Walk me through your approach to ERP or CRM integration if our project requires it.”
What you are checking: integration capability, which is where most B2B Magento projects fail. Strong agencies can describe specific patterns (event-driven sync, polling integration, middleware architecture), have worked with specific ERPs (Microsoft Dynamics, SAP, NetSuite, Oracle EBS), and have opinions about idempotency, error recovery, and monitoring.
Red flag: any answer that treats integration as “we’ll figure it out during build.”
12. “Who owns the relationship with our team after launch?”
What you are checking: post-launch reality. Strong agencies have a named technical contact assigned at kickoff, with that contact remaining the same person through launch and into ongoing support. Weak agencies hand off from project team to support team at launch, and the support team has none of the project context.
Red flag: vague hand-off processes or “you’ll work with our account manager.”
How to actually run the screen
The twelve questions are most useful in two formats. Either a structured 90-minute technical screen with the proposed lead architect (not the sales team), or a written response requirement included in the RFP with each question requiring a specific, named example.
The agencies that respond strongly to this screen self-select. We have seen procurement processes where the screen alone eliminated five of seven shortlisted agencies before contract negotiation. The cost of running the screen is small relative to the cost of picking a weak agency.
The corollary: how to recognize a strong fit
Beyond the twelve questions, four signals correlate strongly with successful Magento agency relationships in our experience.
Specificity in their answers. Strong agencies talk about specific clients (with permission), specific incidents, specific code changes, specific numbers. Weak agencies talk in generalities.
Willingness to push back. The right agency will tell you when your scope is wrong, when your timeline is unrealistic, or when your platform choice is suboptimal. Agencies that say yes to everything are agencies that will under-deliver later.
A working code sample or technical assessment as part of the evaluation. Many strong agencies will offer a paid technical audit or a small scoped engagement before the main contract, as a way to demonstrate fit before either side commits.
Direct access to the technical lead. If you cannot get the proposed lead engineer on a call before signing, you will not have access to them after signing either. This single signal predicts a lot about how the relationship will run.
What we would tell a CTO running this screen
The honest framing: most Magento agencies cannot answer eight or more of these twelve questions with specifics. That is not because the work is unusually hard; it is because most agencies have built their pitch around presentation quality rather than engineering depth. The screen filters for the small share of agencies that have actually shipped production Adobe Commerce work at a level you would want to inherit.
Bemeir has built our practice around being one of those agencies, and we expect this kind of scrutiny from buyers we want to work with. When prospective clients run a strong technical screen, the conversation that follows is more grounded, the expectations on both sides are clearer, and the resulting engagement has a much better chance of being a five-year partnership rather than a 90-day disappointment. The questions are not adversarial. They are the conversation that needs to happen before signing, and the agencies that welcome them are the agencies worth signing with.
If you are running a Magento agency search and want a benchmark for what good answers look like across the twelve questions, our team is happy to walk through ours. The goal of the conversation is not selling the engagement; the goal is helping you ask harder questions of every agency you are evaluating, including us. The CTOs who do this end up with better outcomes regardless of which agency they pick.
