The average Magento RFP is a copy-paste artifact that asks 60 questions and surfaces almost nothing useful about agency capability. The questions are too generic, the response format invites marketing copy, and the evaluation criteria are not specific enough to distinguish a strong team from a polished sales pitch. The right RFP is shorter, sharper, and built around questions that agencies cannot answer without revealing how they actually work.
Bemeir’s Magento team has both responded to hundreds of RFPs and helped clients build them for replatform decisions. The framework below is the version that consistently produces useful evaluations rather than equivalent-looking responses from every bidder.
What to leave out
The categories of questions that produce no useful signal:
Company background questions. Years in business, team size, office locations, founding story. This information is on the agency’s website. Including it in the RFP just gives them a place to copy marketing copy and pads the response length.
Generic capability checkboxes. “Do you offer Magento development?” Yes, every agency you sent the RFP to does. The yes-or-no format invites no useful evidence.
Open-ended methodology questions. “Describe your approach to Magento development.” This produces a page of agile-flavored marketing language that all sounds alike. The same answer arrives from every agency.
Reference lists. Agencies provide references they have curated. The references will say nice things. The signal-to-noise ratio is near zero unless you call the references with your own list of pointed questions, in which case the agency-provided list is less useful than your own research.
Pricing for unspecified scope. Asking for blended hourly rates or general budget ranges before scope is defined produces meaningless numbers that the agency will revise once they know what you actually want.
What belongs in the RFP
The questions below are designed to require evidence and to be answerable in 1-2 paragraphs each. The format is intentionally compact; verbose responses are themselves a signal.
A scoped technical problem to solve. Give bidders a specific, real problem from your store: “Our checkout has a 4.2s LCP on mobile. Describe the first three diagnostic steps you would take and what tools you would use.” This surfaces practitioner depth in a way that capability checkboxes cannot. The agencies that respond with “we would run an audit” without specifics are different from the ones who name PageSpeed Insights, Sentry, server-side timing breakdowns, and the specific Magento profiler tools they would reach for.
Named team and senior availability. “List by name the engineers who would be assigned to our account, their tenure with your agency, and their hours per week available for our project.” This forces the agency to commit to specific staffing rather than the bait-and-switch where senior engineers sell the work and juniors deliver it.
Code samples or write-ups of real work. “Share an architecture decision document, code sample, or post-mortem from a comparable engagement, with client name redacted if necessary.” Agencies that have real practitioner depth have artifacts they can share. Agencies that do not will fall back on marketing case studies.
Patch and upgrade cadence. “What is your standard cadence for applying Adobe Commerce security patches on managed clients, and how is it tracked?” The expected answer references the Adobe security bulletin cycle and includes specifics about staging, testing, and production rollout. Vague answers signal a vague practice.
Performance baseline approach. “How do you establish a Core Web Vitals baseline on a new engagement, and how do you track it over time?” Real practitioners will name the specific tools (Lighthouse CI, CrUX, real-user monitoring), the cadence, and the dashboard they share with clients.
Incident response. “Describe your last production incident on a client store. What was the root cause, the time to detect, the time to mitigate, and what changed after?” This question separates agencies that operate maturely from those that hope nothing goes wrong.
Hyvä, headless, and frontend posture. “What is your default frontend recommendation for a new Adobe Commerce build today, and under what conditions does that recommendation change?” The answer reveals whether the agency has a thoughtful position on Hyvä, PWA Studio, headless, and Luma, or whether they recommend whichever they have the most installed base of.
Off-ramp terms. “What does it look like operationally if we end the engagement after the build? How are credentials, code, and operational knowledge handed off?” Strong agencies have an articulated off-ramp because they have run it before. Weak agencies have not thought about it.
RFP structure that produces useful responses
| Section | Word limit | Purpose |
|---|---|---|
| Scoped technical problem response | 400 words | Tests practitioner depth on a concrete problem |
| Named team and availability | 200 words | Forces explicit staffing commitment |
| Code or document sample | Attached artifact | Shows what real output looks like |
| Patch/upgrade cadence | 150 words | Reveals operational discipline |
| Performance baseline approach | 150 words | Reveals measurement discipline |
| Incident response example | 300 words | Reveals maturity under stress |
| Frontend posture | 200 words | Reveals technical opinions vs. order-taking |
| Off-ramp terms | 200 words | Reveals confidence and operational readiness |
| Pricing (post-scoping) | After scope alignment call | Pricing comes after scope, not before |
The word limits matter. Without them, agencies produce 40-page responses that are impossible to evaluate side by side.
How to evaluate the responses
A useful scoring rubric weights evidence over assertions:
- 2 points for each answer that includes a specific tool, technique, name, or number
- 1 point for each answer that is on-topic but generic
- 0 points for marketing language without substance
- Negative 1 point for any answer that contradicts a separate answer in the same response (consistency check)
Across the eight categories, a response can score 0-16. Strong agencies cluster at 12-14. Weak agencies cluster at 4-6. The middle band is where judgment calls happen.
The supplemental call that decides the bid
The RFP responses narrow the field to two or three candidates. The decision is made on a 60-minute call with each finalist that goes deep on one of their answers. Pick the answer where you have the most domain doubt and pressure-test it.
If they claimed a 1.2s LCP on a recent Hyvä build, ask exactly which optimizations did the work, in what order, and how they verified each step. If they described a B2B integration with NetSuite, ask which records sync in which direction, what idempotency posture is used, and what happens when NetSuite is down. The agency that can go three levels deep in technical conversation is the one that will deliver. The agency that retreats to generalities is the one whose RFP response was a sales document.
Bemeir recommends this format to clients running their own RFPs because it is the version that has consistently produced strong agency selections across mid-market and enterprise tiers. The Adobe Commerce ecosystem has too many agencies for capability checkboxes to differentiate; the RFP is only useful if it forces the agencies to show their work.
What this leaves out and why
This framework deliberately excludes ESG questionnaires, indemnification clauses, data-handling policies, and similar compliance content. Those belong in the master services agreement, not the RFP. Mixing them into the capability evaluation makes both worse: the compliance questions are answered by lawyers, the capability questions are answered by salespeople, and neither set produces useful signal.
The structure also deliberately delays pricing until after a scoping call. Asking for pricing without scope produces estimates that have no relationship to what the engagement will actually cost. Better to align on scope first, then ask the finalists to price the same scope. The numbers you receive are then comparable.
A good Magento RFP is short, evidence-driven, and designed to separate practitioners from polished sales teams. The agencies you want to work with will appreciate the directness. The agencies that struggle with it are the ones the format is helping you screen out.
