Restaurant facilities management RFP: what to ask, what the answers actually mean

The mandate from the CFO lands on a Tuesday: consolidate the 11-vendor maintenance roster across the 18-location footprint, target 15 percent reduction in annual facilities spend, have something signed by end of Q3. Fourteen days to issue, sixty days to close. Whatever you publish is going to be read by someone on the corporate team who has never set foot in any of your kitchens.

This is the situation almost every multi-unit restaurant operator's facilities-services RFP gets written in. The CFO has a target, the procurement business partner has a template they used for office supplies, and the operations director has a list of vendor problems that triggered the consolidation mandate in the first place. The RFP that comes out the other end usually does one of two things: it asks every vendor to describe themselves and rewards the one with the best marketing department, or it asks every vendor to bid against an undefined scope and rewards the one with the most aggressive sales engineer. Neither of those picks the vendor that's actually going to ship the work for the next 36 months.

The framework below is the one a competent restaurant facilities RFP runs on: six categories worth asking about, the exact question wording inside each one, what a good answer looks like, what a red-flag answer looks like, a weighted scoring rubric, a reference-check script, and a 60-day evaluation timeline. The framework is honest. It applies to any vendor evaluation, not just a Boh evaluation. Run a Boh RFP through it (Boh, which manages back-of-house repairs, maintenance, and compliance for Southern California restaurants) and you'll find the same gaps and the same strengths a competitor RFP would surface.

Why most restaurant facilities RFPs produce the wrong vendor

Most restaurant facilities RFPs underperform for three reasons. First, they ask vendors to describe themselves in narrative form, which is a marketing-department evaluation, not an operational evaluation. Every vendor will claim same-day response, multi-trade coverage, and compliance-first culture. The narrative does not distinguish. Second, they treat all six categories as equally important, which is procurement-template thinking. The right weighting is the weighting that reflects your last vendor failure. If your incumbent's compliance documentation gap cost you a health-code event, you should be weighting insurance and references higher than the procurement template recommends. Third, they skip or compress the reference-check step, which is the single highest-signal step in the entire process and the only one that interrogates the vendor's actual customer relationships rather than the vendor's marketing claims.

The reframe is structural: the RFP should be designed so that a competent vendor sounds different from an incompetent vendor on the page. If the question is "describe your SLA," every vendor sounds the same. If the question is "publish your median emergency dispatch arrival time across the last 200 events, separated by trade and submarket, with the credit schedule that applies when arrival exceeds the median by more than 50 percent," the competent vendor produces a chart and the incompetent vendor produces a paragraph about culture. Structure the questions so the gap is visible.

The six categories worth asking about

Each category below contains five questions. Pick the ones that matter to your operation; not every question applies to every buyer. The questions are written as you'd send them to a vendor in the RFP document. The "what good looks like" and "red flags" notes are for the scoring side.

Category 1: Coverage and capabilities

1. List the specific trades you cover under a single accountable contract (e.g., hood cleaning, refrigeration, hot line, HVAC, ice machine, fire-suppression, plumbing, grease-trap, electrical). For each trade, name the credential held by the technician who performs the work. For hood cleaning, ask whether the contractor is trained, qualified, and certified, acceptable to the authority having jurisdiction (the standard set by NFPA 96), with the IKECA certification the recognized industry credential. Ask whether the technician is W-2 or a vetted sub-contractor.
2. Define your service area by submarket density, not regional boundaries. What is your median tech-to-site drive time across your three largest customer submarkets? Where are the submarkets you do not currently service?
3. Describe your in-house bench depth per trade. For refrigeration specifically, how many EPA Section 608 certified technicians (refrigerant-handling certification under 40 CFR Part 82, Subpart F) do you carry on the bench? How is shift coverage organized after 6pm and on weekends?
4. What is your maximum simultaneous-event capacity in our footprint? If a regional weather event triggered 40 simultaneous refrigeration calls across our locations on a Saturday night, how would dispatch sequence the work and what would the median arrival time look like?
5. What work do you explicitly NOT do, that is, what scope are we definitely going to need a separate vendor for? (The vendors who answer this honestly are the ones worth shortlisting.)

What good looks like: specific trade-by-trade credential disclosures, named submarket coverage with drive-time data, explicit out-of-scope items, willingness to disclose bench depth in numbers.

Red flags: "full-service coverage" without specifics, refusal to name out-of-scope work, vague language about "our network of trusted partners" without disclosing how the network is vetted or who carries the warranty.

Category 2: SLA structure

1. Define your SLA tiers with specific time commitments per priority class (P1 emergency / P2 same-day / P3 next-day / P4 scheduled). For each tier: dispatch acknowledgment time, on-site arrival time, and resolution time. Publish actual achieved median performance against each tier across the last 12 months, not the marketing claim.
2. Describe your service-level credit schedule. What is the credit per missed SLA event? Is it a flat dollar amount, a percentage of monthly fee, or a free service visit? Is there a cap on total monthly credits? Are there event categories where credits do not apply?
3. Describe your escalation matrix with named role bands (not individuals, those change). At what point does a stuck work order escalate from dispatch coordinator to operations manager to executive? What is the escalation-response SLA at each step?
4. What is your first-trip-fix rate on the most common failure modes (refrigeration compressor, hood-suppression inspection failure, ice machine production fault, hot line ignition)? What percentage of dispatches require a return visit, and what triggers the return?
5. Define your weather, force-majeure, and access exceptions. When can SLA performance be paused, and what counts as restaurant-side access failure (locked, no key, no contact reachable) vs. vendor-side performance failure?

What good looks like: published historical median performance per tier, named credit-per-event amounts, escalation matrix with role bands and response times, transparent exception conditions.

Red flags: "rapid response" without numbers, "credits handled case by case" (meaning never), no published exception conditions (meaning they can claim anything is an exception), refusal to publish historical SLA performance (meaning they don't track it or it's worse than the marketing claim).

Category 3: Technology and integration

1. Describe your work-order intake methods. Do you support web form, email, phone (and what hours), IVR (automated phone-tree dispatch), and API/EDI from a customer-side CMMS? Name the specific CMMS platforms you integrate with today by customer (Corrigo, ServiceChannel, FacilityForce, MaintainX, Limble, eMaint, Fiix).
2. What does post-visit documentation include? Photo documentation per visit, signed technician notes, parts list, time-on-site, before/after photos for compliance-relevant work (hood cleaning, fire-suppression inspection, grease-trap pumping)? In what format do you deliver it, and how long do you retain it?
3. Describe your customer-facing dashboard capabilities. What KPIs are visible to the customer in real time? Cost-per-location, work-order aging, SLA performance by site and by trade, PM compliance by location, vendor scorecards by sub-trade?
4. What is your NTE (not-to-exceed, the pre-approved spend ceiling for repair work without further authorization) approval workflow? What is the default NTE threshold per work order? How is over-NTE work approved, and what is the approval-response SLA on the operator side?
5. Describe your API and reporting export capabilities. What data can be extracted (work-order detail, asset history, cost summaries, photo doc) and in what format (CSV, JSON, scheduled SFTP delivery)?

What good looks like: named CMMS integrations with live customer references, photo-documentation as standard, real-time dashboard with cost and SLA visibility, transparent NTE workflow.

Red flags: custom-integration-only language with no named platforms, photo doc available "on request" rather than as standard, dashboard described in screenshots but not made available for trial during the RFP, NTE workflow described as "case-by-case" (meaning operator surprise on the invoice).

Category 4: Insurance and risk transfer

1. Publish your general-liability coverage limits, per occurrence and aggregate. Workers' compensation. Auto liability if technicians drive to site. Professional liability if you offer compliance advisory work. Excess umbrella if applicable.
2. Are you willing to name corporate, franchisee, and landlord entities as additional insured on a per-customer basis? With primary-and-non-contributory wording? With waiver of subrogation? With 30-day notice of cancellation? Publish your standard COI (certificate of insurance) language and any deviation from the customer's required language.
3. Describe your indemnification language in your standard MSA. Mutual or one-way? Caps on liability? Carve-outs for gross negligence and willful misconduct? Survival of indemnification post-termination?
4. Describe your sub-contractor insurance flow-down. If a vetted sub-contractor causes property damage at a customer site, whose policy is primary? Do you require the sub to carry equivalent coverage and to name the customer as additional insured? Is the sub-contractor flow-down audited?
5. Describe your claims history over the last three years at the corporate level. Have you had any general-liability claims paid in excess of $50,000? Any auto claims? Any sub-contractor-caused incidents that flowed to your policy?

What good looks like: publishable coverage limits ($2M / $5M / $10M typical bands), willingness to name corporate AND franchisee AND landlord as additional insured, primary-and-non-contributory standard, transparent claims history with no surprises.

Red flags: refusal to share aggregate limits, hesitation on additional-insured endorsement for multiple entities, indemnification language that's one-way in the vendor's favor, refusal to disclose claims history.

Category 5: Financials and stability

1. What is your company size, annual revenue band, full-time employee headcount, technician headcount specifically, geographic footprint, customer count?
2. What is your customer concentration? What percentage of revenue comes from your top customer, top three, and top ten? Are any of those customers under multi-year exclusivity?
3. What is your key-person risk? Are there one or two individuals whose departure would noticeably affect operational continuity? What's the succession plan?
4. What is your funding posture? Privately held, founder-owned, private-equity-backed (with what investment vintage), public? If PE-backed, when is the typical exit timeline relative to our contract term?
5. What is your operational continuity plan in the event of a cyber-incident, key-person event, or natural disaster affecting your primary dispatch center?

What good looks like: publishable size-band, no single customer above 25 percent of revenue, named successor for any key person, transparency about funding posture and exit timeline.

Red flags: evasive on revenue band ("we're privately held and don't disclose"), top customer above 40 percent of revenue (concentration risk to the operator), PE-backed with exit timeline shorter than the contract term, no documented continuity plan.

Category 6: References and transition

1. Provide three reference customers currently using your service: one similar in size to ours, one larger, one in a similar geography. For each reference, provide the operations contact (not the sales contact) with email and phone, and the duration of the customer relationship.
2. Provide your standard MSA as an attachment to the RFP response, not as a deliverable post-award. We will not negotiate term sheets we haven't read.
3. Describe your transition-in methodology. For a 15-location operator switching from a self-managed vendor roster to your network, what is the standard 60-90 day transition plan? What data do you need from the incumbent? What is the parallel-run period? What is the cutover risk?
4. Describe your transition-out methodology. If this contract is not renewed, what is the standard 60-day off-ramp? What customer data do you return (asset register, work-order history, photo doc archive), in what format, and how quickly? What is the customer's intellectual-property posture on the data you've collected on their behalf?
5. What does a customer lose by switching to your network, that is, what capability or relationship are they giving up that the incumbent or a peer competitor would retain? (The vendors who can articulate this are the honest ones.)

What good looks like: references with operations-side contact info, MSA attached, transition-in and transition-out as documented playbooks, customer-IP-retention posture that returns data to the customer at termination.

Red flags: sales-only references, MSA "to be developed during onboarding," vague transition language, vendor retains customer data after termination, refusal to answer the "what does a customer lose" question (vendors who can't name a tradeoff don't understand their own positioning).

The weighted scoring rubric

Two scoring tools you need: a per-question rubric and a category-weight template.

The per-question rubric is the same across categories: 0 = no answer or refusal; 1 = vague narrative without specifics; 2 = some specifics, gaps where it matters; 3 = solid specifics across the board, no published evidence; 4 = published evidence (numbers, schedules, documented playbooks); 5 = published evidence plus willingness to be audited against it.

The category-weight template depends on which gap is most painful in your operation. Three starting templates:

Category	Cost-conscious (technical side)	Compliance-first	Risk-averse enterprise
Coverage and capabilities	25%	20%	20%
SLA structure	30%	20%	15%
Technology and integration	15%	10%	15%
Insurance and risk transfer	10%	20%	20%
Financials and stability	10%	10%	15%
References and transition	10%	20%	15%

The cost-conscious column weights SLA and coverage heavily within the technical evaluation because the work-getting-done question dominates the spend question. The compliance-first operator (regulated, franchise-heavy, or insurance-driven) over-indexes on insurance and references because a documentation gap is the failure mode that triggers consolidation. The risk-averse enterprise spreads weights more evenly because their failure modes are more diffuse and the procurement function is mature enough to model all six categories simultaneously.

Pricing-architecture is a separate column. The six categories above scope the technical side of the evaluation. Pricing-architecture (T&M vs. flat-fee, annual escalator caps, parts markup ceilings, NTE flow, after-hours premium multipliers, trip-charge structure) is best scored as a parallel commercial column rather than folded into a technical category. Most multi-unit operators run a 70/30 or 60/40 technical/commercial split. A future expansion of this framework will cover the pricing-architecture question set in the depth the technical categories get here; for now, treat the table above as the technical scorecard and build your own commercial scorecard alongside.

Score each vendor across the 30 technical questions, sum the weighted score to a 100-point total on the technical side, and use the total (combined with your separate commercial score at whatever split you've chosen) as a shortlist filter. Pick the top three. Do not pick based on score alone; the score is a filter, not a decision.

The reference-check script: 10 questions to ask current customers

The reference check is the highest-signal step in the entire process and the step procurement teams routinely skip or compress. Plan 30 minutes per reference; pick three references per shortlisted vendor; ask the following 10 questions in this order:

1. How long have you been a customer? Anything shorter than 18 months is too new to be a useful reference. Anything past five years is potentially captive (switching cost may exceed dissatisfaction).
2. How did the transition-in go? What broke that you weren't expecting? How long did it take to reach steady-state?
3. What was your worst SLA miss in the last 12 months? What happened, how did the vendor handle it, what credit (if any) did you receive?
4. How does invoice reconciliation work for your accounting team? How many invoice disputes per month, on average? How quickly do disputes resolve?
5. What does the dashboard tell you, and what does it not tell you? Are there reports you've had to build externally because the vendor's reporting doesn't cover them?
6. How is your relationship with the dispatch team vs. the sales team? (The gap between the two is diagnostic.)
7. What's the worst dispatch outcome you've had? Walk me through what happened.
8. If you were running this RFP again, what would you ask that you didn't ask the first time? (This is the most-undervalued question; it surfaces the gap the customer has learned to live with.)
9. Would you renew if the price went up 8 percent? The answer separates necessary-relationships from sticky-but-resented-relationships.
10. Can you give me the name of another operator like us who uses this vendor, not on the reference list?

Question 10 is the most important question in the script. The reference customers are pre-vetted by the vendor; they're the vendor's friendliest accounts. The second-degree reference (the operator named by the original reference) is unselected and almost always more candid. If a reference cannot or will not name a peer, that's signal: the customer base may be smaller or less networked than the vendor implies.

The 60-day evaluation timeline

1. Week 1: Define the buy. Scope confirmation with the operator-side stakeholders (CFO, ops director, regional facilities manager). Write the RFP document. Identify the vendor list (target five).
2. Week 2: Issue the RFP. Send to five vendors with identical materials and a published Q&A deadline. Hold a single optional clarifying-questions call open to all bidders; never hold separate calls with individual vendors during the open window.
3. Week 3: Vendor response window opens. Field clarifying questions in writing, republish answers to all bidders simultaneously.
4. Week 4: Vendor response window closes. Receive responses; begin scoring against the rubric. Identify the top three.
5. Week 5: Shortlist and demos. Two-hour demo per shortlisted vendor with operations-side stakeholders present (not just procurement). Demo agenda is set by the buyer, not the vendor, and includes a walkthrough of the dashboard, a worked SLA-miss scenario, and a question-and-answer block.
6. Week 6: Reference checks. Three references per shortlisted vendor. Run the 10-question script. Identify the second-degree references and call at least two. This is the step most RFPs skip; it is also the step that most reliably surfaces the gap between marketing claim and operational reality.
7. Week 7: Final selection and MSA negotiation. Single vendor named. MSA negotiation begins from the standard template each vendor attached to their RFP response. Negotiate term, pricing escalators, SLA credits, additional-insured endorsements, and transition-in/out language. Plan for a 7–10 business day MSA cycle.
8. Week 8: Contract signed. Transition-in playbook activates per the vendor's documented methodology. Incumbent off-boarding begins in parallel with new-vendor onboarding.

The reference check is the only step in the entire process that interrogates the vendor's customer relationships instead of the vendor's marketing claims. Run it.

When NOT to run an RFP

Not every restaurant facilities buy belongs in an RFP. Four operator profiles where the RFP overhead exceeds the value:

Single-location operators. The economics of an RFP (60 days of internal time, vendor coordination overhead, scoring rubric application) assume a buy large enough to justify the work. A single-location operator's facilities maintenance buy is typically too small to support the procurement overhead. Better: shop two or three regional networks directly, ask the most important five questions across them in 30-minute conversations, and pick the one that answers with specifics.

Sub-$2M revenue restaurant groups. The threshold is rough, but operators below roughly $2M annual revenue typically don't have a procurement function and don't have the operational bandwidth to run an RFP without it consuming the operations director's attention for two months. The bid savings rarely justify the opportunity cost. Direct vendor shopping is usually more efficient.

Operators with no internal facilities function. If there's no FM-side stakeholder to interpret the vendor responses, score them against the rubric, and run the reference checks, the RFP defaults to picking the vendor with the best brochure, which is exactly the failure mode the RFP was supposed to prevent. Hire facilities expertise first (even as a fractional consultant), run the RFP second.

Operators already on a working managed-network relationship. If the incumbent is performing against SLA, the documentation is clean, the invoice reconciliation is working, and there's no specific failure mode driving the consolidation, the RFP is usually a competitive-tension exercise rather than a vendor change. There are cheaper ways to apply competitive tension: a benchmark conversation with two peer operators, a quick three-vendor pricing comparison on a single trade, or a planned MSA renegotiation around the contract anniversary.

For everyone else, multi-unit operators above roughly 8 locations, with an internal facilities function, with a specific failure mode that triggered the consolidation mandate, the RFP framework above is the right shape.

Monday-morning action, no pitch attached

If your operation is in the consolidation-mandate window, CFO-led, multi-unit, with a target percentage on annual facilities spend, the question to settle on Monday is which of the three weighting templates above fits your last vendor failure. Cost-conscious, compliance-first, or risk-averse enterprise. That choice does more to determine the RFP outcome than the bidder list does. The RFP that scores against the right weighting picks the right vendor; the RFP that scores against the procurement-template weighting picks the vendor with the best brochure. Pick the weighting first. Write the questions second. Run the references before the demo.

For multi-unit restaurant operators in Southern California weighing a Boh consolidation against a national platform or a CMMS-plus-network hybrid, the Maintenance Coverage page covers the Group-tier model; the Services page covers the trade scope and published SLA structure; and the Get In Touch page is the path to scope a comparative conversation that maps Boh's response into the six categories above. The framework is the framework whether Boh is the answer or not.