Building a Defensible Hard FM Maintenance Framework: The Practical Guide to Getting Beyond SFG20
Seven articles have laid out the problems: labour models built on fictional hours, asset registers nobody has verified, a subscription platform with acquisition risk, compliance dashboards that measure the wrong things, and technology making fixed-interval scheduling increasingly redundant on instrumented plant. This article does one thing: tells you what to do instead. A step-by-step framework, grounded in practice, that any Hard FM operator or estate director can start building from today.
This framework does not require a transformation programme, a new CAFM platform, or a seven-figure technology investment. It requires a decision to stop treating SFG20 as a complete answer and start using it as one component of a more intelligent maintenance model. The six steps below can be applied progressively, in order, on any estate of any size. Each step builds on the previous one. Each one delivers standalone value even if the next step is never taken.
A word on what this framework is not. It is not a rejection of SFG20. SFG20 remains the appropriate default maintenance specification for the majority of M&E asset classes on the majority of UK estates. The framework proposed here uses SFG20 as the baseline for everything the other layers do not cover. The argument of this series has always been that SFG20 is a very good servant and a poor master. This is the framework that makes it a servant.
The Six-Step Framework
Nothing in this framework works without accurate asset data. Before a criticality assessment, before technology integration, before a revised PPM schedule, the asset register must reflect the actual estate. Commission a physical survey. Tag every maintainable asset with a unique identifier. Record asset class, manufacturer, age, and condition. Hold the resulting register independently of the contractor's CAFM. This is the foundation. Everything else is built on it.
Map every asset on the verified register against its statutory maintenance obligation: L8 Legionella (ACoP L8), fixed wire testing (BS 7671 / IET Wiring Regulations), fire alarm maintenance (BS 5839), emergency lighting (BS 5266), lift thorough examination (LOLER), pressure systems (PSSR), F-gas (EC 517/2014), and for NHS estates, the full suite of Health Technical Memoranda. These tasks are non-negotiable. They are the floor. Record them separately in the CAFM with their own completion tracking, evidence requirements, and escalation paths. Do not allow them to be treated as equivalent to discretionary SFG20 tasks in reporting.
Not all assets carry the same consequence of failure. A chiller serving a critical care unit is not the same as a fan coil unit in a back-office meeting room. A standby generator on a data centre is not the same as a domestic hot water calorifier in a low-occupancy building. The criticality assessment assigns each asset class a tier based on two dimensions: consequence of failure (safety, statutory breach, business continuity, occupant comfort) and likelihood of failure (age, condition, maintenance history, operating load). The tier determines what happens to the SFG20 schedule for that asset class: increase frequency, decrease frequency, add condition monitoring, or move to a manufacturer-specified interval.
For the assets identified as Tier 1 in the criticality assessment — safety-critical, business-continuity-critical, or high-value primary plant — replace SFG20 frequency scheduling with manufacturer-specified intervals and, where BMS or IoT monitoring is available, condition-triggered maintenance. Track manufacturer specifications systematically in the asset register. Flag any asset where the manufacturer interval is less demanding than SFG20 (potential cost saving) and any where it is more demanding (potential compliance gap).
For the majority of the estate — the assets that are not safety-critical, not under active condition monitoring, and not covered by specific manufacturer intervals — SFG20 remains the appropriate maintenance specification. Apply it as written, with the access complexity multipliers and productive wrench time adjustments described in Article 2 of this series. Do not use SFG20 generic task durations for labour modelling without site-specific adjustment.
A defensible maintenance framework is not just a schedule. It is a documented evidence chain: what was done, by whom, on which asset, at what time, with what result, against which standard. Every PPM task needs a service record, not just a CAFM closure. Every statutory task needs the evidence the regulator will ask for. Every deferral needs a documented risk assessment. And the framework needs a live change management process: when plant is installed, replaced, or decommissioned, the register updates in real time.
What This Framework Looks Like in a Contract and a Tender
| Framework Component | How to Specify in Tender | How to Measure in Contract |
|---|---|---|
| Verified asset register | Require physical survey at mobilisation as a priced deliverable. Specify unique asset identifier scheme. Require client-held golden copy. | Asset register accuracy audit at months 3 and 12. KPI: verified count vs CAFM count within 2%. Penalty for non-delivery of golden copy by month 3. |
| Statutory separation | Require statutory tasks to be separately coded, tracked, and reported in CAFM. Require evidence records for each statutory task closure. | Monthly statutory compliance report separate from overall PPM report. 100% statutory completion KPI. Zero tolerance for evidence gaps. |
| Criticality assessment | Require contractor to deliver criticality assessment within 60 days of mobilisation. Specify two-dimension scoring: consequence and likelihood. | Criticality tiers included in asset register by month 3. Tier 1 assets on separate reporting track. Annual review of tiering sign-off. |
| Manufacturer intervals | Require contractor to record manufacturer maintenance specification against each asset in the register. Flag any divergence from SFG20. | Quarterly report on assets where manufacturer interval overrides SFG20. Evidence that warranted plant is maintained to warranty conditions. |
The gap between what SFG20-aligned means in a contract and what it means in practice is not a gap in the standard. It is a gap in the specification, the evidence requirement, and the willingness to enforce both.
- Art. 1 Is SFG20 Outdated? The Hard FM Baseline That Built an Industry and Why It Is Now Costing You Money
- Art. 2 SFG20 Labour Hours: Why Your PPM Pricing Is Wrong Before the Contract Starts
- Art. 3 Who Owns SFG20? BESA, Facilities-iQ, and the Hard FM Commercial Risk
- Art. 4 What SFG20 Compliance Actually Costs: Five Hard FM and TFM Perspectives
- Art. 5 The Asset Register Problem: Why the Foundation of Every Hard FM Contract Is Built on Data Nobody Has Verified
- Art. 6 Beyond SFG20: The Credible Alternatives for Hard FM Maintenance Strategy
- Art. 7 What Technology Does to the Case for SFG20: IoT, BMS Integration, Digital Twins and AI Maintenance Analytics
- Art. 8 How to Build a Defensible Hard FM Maintenance Framework Without SFG20 as the Anchor. (this article)
- Art. 9 SFG20 State of FM Report 2026: What the Data Actually Shows About SFG20 Compliance, Asset Registers and Hard FM Dependency.
Every step of the framework in this article is something Baachu has built for Hard FM operators and estate clients across the UK.
- Asset register verification: Physical survey, unique ID scheme, golden copy for the client.
- Criticality assessment: Two-dimension scoring applied to your estate. Tier 1 assets identified.
- Labour model validation: Site-specific productive wrench time model. Access complexity multipliers.
Contact us: hello@baachu.com · baachurain.com
Frequently Asked Questions
The answer is six steps applied in sequence. First, verify the asset register through a physical survey with unique identifiers. Second, separate statutory obligations from discretionary PPM. Third, apply a criticality assessment to tier the estate. Fourth, replace SFG20 frequencies with manufacturer intervals for high-criticality assets. Fifth, use SFG20 as the default for the rest. Sixth, build an evidence chain into the contract. It is a system where SFG20 is one component rather than the only answer.
It scores each asset class on consequence of failure (safety, business continuity) and likelihood of failure (age, condition). Tier 1 assets receive the most intensive treatment: manufacturer-specified or condition-triggered maintenance. Tier 2 and Tier 3 remain on SFG20 but with frequencies adjusted based on their scoring.
For FM and estates teams who also use BCIS for cost benchmarking, the BCIS Intelligence Series closes with the equivalent practical framework: how to stop using BCIS as a crutch and build a proper intelligence stack.