SFG20 Labour Hours: Why Your PPM Pricing Is Wrong Before the Contract Starts
SFG20 task durations are estimates. Every experienced FM estimator knows this and builds workarounds in private. This article makes that private knowledge public, explains the commercial damage it causes, and sets out what a defensible labour model actually looks like.
The Problem in One Paragraph
SFG20 Facilities-iQ displays estimated durations against each maintenance task. BESA's own published FAQ describes these as 'generic service timings' that 'could vary dependent on the size, age and/or location of the equipment or building.' That sentence is doing a great deal of work. What it means in practice is that the times shown are a starting assumption, not a cost model. In a competitive bid environment, that starting assumption frequently becomes the pricing model — because there is no time, no asset data, and no commercial incentive to replace it with something more accurate. The contract is won on the assumption. The margin is lost in the delivery.
What Productive Wrench Time Actually Is, and Why It Is Not 1,800 Hours
A standard working year for a full-time field engineer is approximately 1,800 hours, net of annual leave and bank holidays. This is the number that appears in many labour models. It is not the number of hours that engineer actually spends with tools in hand, doing the work that SFG20 task durations describe.
The gap between 1,800 hours and real productive time is composed of the following categories, none of which appear in SFG20 schedule timings.
| Time category | Description | Typical daily allowance |
|---|---|---|
| Travel between sites or plant rooms | Driving, walking, navigating multi-floor buildings, waiting for lifts and escorts | 45 to 90 minutes |
| Permit-to-work administration | Raising, clearing and closing PTW forms, isolation verification, waiting for permit authorisation | 30 to 90 minutes on permit-required tasks |
| Access preparation | Moving furniture, removing ceiling tiles, setting up access equipment, repositioning scissor lifts | 15 to 45 minutes per access event |
| Tool and material logistics | Collecting parts from stores, waiting for parts deliveries, returning tools | 20 to 40 minutes per day |
| Handover and documentation | Completing job sheets, uploading evidence to CAFM, briefing the next shift | 20 to 30 minutes per day |
| Unplanned reactive demand | Reactive callouts pulling engineers off PPM schedules, reactive repair support | Unpredictable: budget 15 to 25% of shift capacity on reactive-heavy contracts |
| Competence and induction | Site inductions, refresher training, toolbox talks, H&S briefings | Variable: higher in first three months of a new contract |
When these categories are applied, the realistic productive wrench-time figure for a field engineer on a complex FM estate sits between 1,300 and 1,400 hours per year. The industry broadly accepts this range, though few publish it. The gap between 1,800 theoretical and 1,350 realistic is around 25 percent. On a contract where labour is 60 to 70 percent of total cost, that gap is not a rounding error. It is a structural pricing risk that is either absorbed in the contingency pot or discovered when the contract goes into margin decline in year two.
The contractor who prices 1,800 hours and delivers 1,350 is not cutting corners. The estate simply takes longer than the model assumed. That is a data failure, not a delivery failure.
The Five Layers SFG20 Task Times Do Not Include
SFG20 task durations describe the time required to perform the maintenance task itself, in ideal conditions, on an asset of standard specification. Real sites do not offer ideal conditions. The following five layers add time and cost to every task on a real estate, and none of them appear in the SFG20 baseline.
Layer 1: Access Complexity
Access conditions vary enormously across estates and even within a single building. A fan coil unit in an open-plan office ceiling void accessible by step ladder is not the same asset as a fan coil unit in a sealed ceiling above a server room, requiring a scissor lift, dust sheets, and a formal access request. SFG20 assigns the same task duration to both. Estimators who do not apply an access multiplier to their labour model are assuming that every asset on the estate sits in an open-plan office. Almost none of them do.
| Direct access, no permit | Task time × 1.0 (SFG20 baseline) |
| Ceiling void, step ladder required | Task time × 1.3 |
| Ceiling void, scissor lift required | Task time × 1.6 to 1.8 |
| Roof plant, height safety equipment | Task time × 1.5 to 1.7 |
| Confined space entry | Task time × 2.0 to 3.0 |
| Restricted hours (out of hours only) | Task time × 1.4 plus overtime premium |
| Infection control zone (healthcare) | Task time × 1.5 to 2.0 plus escort cost |
These are indicative multipliers based on FM contracting experience. Site-specific conditions must be assessed on every estate at tender stage.
Layer 2: Permit-to-Work
Permit-to-work systems are mandatory on many estates, particularly in healthcare, education, industrial sites, data centres, and any building with complex isolation requirements. A permit is not just a form. It involves raising the permit, having it authorised by the client's Authorised Person, verifying isolations, performing the task, and closing the permit with a clearance check. On a busy site with a single AP covering multiple contractors, permit wait times can run to 40 to 60 minutes per task. On a healthcare estate with infection control procedures attached to the permit, longer.
An FM estimator pricing a hospital contract using SFG20 task durations without a permit allowance is omitting what could be the largest single time component on many of the tasks in scope. The standard does not alert them to this. The tender document usually does not either, unless the client's FM team has been thoughtful about scope.
Layer 3: Asset Age and Condition
SFG20 schedules are designed for assets maintained to the correct standard throughout their life. An AHU that has been serviced every year since installation, with filters changed at specified intervals and bearings replaced on schedule, will take approximately the time SFG20 suggests to service. An AHU that is ten years overdue for a bearing change, has a seized access panel, and has bird ingress debris in the filter rack will take considerably longer. On public sector estates, NHS trusts, and local authority property portfolios, the latter is not uncommon. Backlog maintenance is a national problem. SFG20 task times assume an estate in good order.
Layer 4: Asset Register Accuracy
The labour model is only as good as the asset register it is built on. A register that shows 200 fan coil units when the estate has 240, or that lists assets as functional when they have been removed during a fit-out, produces a labour model that is wrong at source. This is not a hypothetical problem. Asset registers on FM contracts, particularly those transferring between operators, are routinely incomplete, inaccurate, or structured in ways that do not align with SFG20 task codes. Experienced estimators build a gap assumption into their models. Less experienced ones do not.
Layer 5: Reactive Demand Interaction
PPM engineers on most FM contracts are also the first-line response for reactive demand. When a reactive call comes in, the PPM schedule slips. When reactive demand is higher than the comprehensive threshold model assumed, PPM completion rates fall. When PPM completion rates fall, the client notices. When the client notices, there are performance deductions. None of this is visible in the SFG20 task duration. The interaction between planned and reactive workload is a contract design problem that the standard was never intended to solve. But estimators who price a PPM labour model in isolation from the reactive demand profile are building a model that will be tested to destruction within the first quarter.
How This Breaks Differently Across Five Contract Types
The gap between SFG20 timings and real delivery cost is not uniform. It is largest on the contracts where it is hardest to identify at tender stage and hardest to recover once the contract has started.
NHS and Healthcare Estates
Infection control requirements add escort time to almost every task in clinical areas. Permit-to-work systems are mandatory and often complex. Out-of-hours restrictions mean many tasks can only be performed during shutdowns, compressing large volumes of work into short windows with overtime exposure. SFG20 task times on a healthcare contract without site-specific access allowances are not a pricing model. They are the first draft of a loss-making contract.
Local Authority and Education
The challenge is geographic spread combined with an asset register that is almost always incomplete. A contract covering 200 school buildings across a county will have assets that nobody has inspected recently, access arrangements that differ by site, and buildings extended, partitioned, and modified without those changes being reflected in any central register. There is no such thing as a uniform estate.
PFI and Long-Term Service Contracts
SFG20 is updated regularly, and those updates can change task scope and frequency in ways that affect the labour model. A labour model built on SFG20 timings at contract inception may be materially incorrect within five years as assets age, as legislation changes, and as the estate configuration evolves. PFI contractors managing this risk through a fixed labour model are carrying a liability that grows with every year of the contract.
Commercial Office Portfolios
On a multi-tenanted building, access restrictions driven by occupier requirements, out-of-hours working rules, and estate management protocols add time to tasks SFG20 assumes are unrestricted. On a Grade A trophy asset where evidence photography is required for every task and engineer presentation standards mean re-attending if dress code is not met, task times increase in ways the standard does not anticipate.
Industrial and Critical Infrastructure
Data centres, utilities, and manufacturing facilities have the highest access complexity and highest consequence of error. Confined space entry, gas detection requirements, electrical isolation procedures under CDM, and lone worker restrictions can multiply the effective task time by a factor of two or three compared to the SFG20 baseline. An estimator using SFG20 durations without a site-specific access review is pricing the wrong contract.
What Experienced Estimators Actually Do: The Workarounds Nobody Documents
Here is what makes the SFG20 labour loading problem genuinely damaging. The gap between the standard and reality is not unknown. Every experienced FM estimator knows it exists. They compensate for it. They just do not document the compensation in a way that the client can see, the procurement team can audit, or the operations team can use when they take over the contract.
- Build a shadow hours model alongside the SFG20-derived task list, applying site-specific access multipliers and permit allowances based on a reading of the ITT documentation.
- Apply a productive time factor of 70 to 75 percent to the theoretical annual hours, reducing 1,800 hours to between 1,260 and 1,350 productive hours per engineer.
- Add a reactive demand overlay, estimating the proportion of engineer capacity that will be consumed by reactive callouts below the comprehensive threshold.
- Build a gap assumption for the asset register, adding a percentage contingency for assets that will be found on site that are not in the register provided at tender.
- Price the risk pot separately, covering unknown access conditions, permit complexity, and asset condition on assets not visited during the tender site survey.
None of these steps appear in the pricing submission. The client sees an SFG20-aligned price. They do not see the layers of judgement that produced it. When the estimator gets those judgements wrong, the contract loses money. When they get them right, the client is paying for expertise that is entirely invisible.
This invisibility is commercially damaging in both directions. The client cannot distinguish between a contractor who has built a rigorous site-specific model and one who has applied generic timings and hoped for the best. The procurement process treats both submissions as equivalent. The one who wins is frequently the one who has been most optimistic, not the one who has been most accurate.
How the Client Gets Damaged Too
The SFG20 labour loading problem is usually framed as a contractor problem. It is also a client problem, and in some respects it is a worse one.
When a contractor under-prices a contract because their labour model is based on SFG20 generic timings rather than a site-specific analysis, one of three things happens within the first eighteen months.
The contractor delivers to specification and loses margin. Maintenance visits are rushed to hit the numbers. Evidence quality drops. Asset condition deteriorates faster than the regime should allow. Within two to three years, the client has a backlog problem that was caused by a pricing model built on insufficient data.
The contractor begins to manage the shortfall by reducing task scope, deferring non-statutory visits, and prioritising the tasks that generate evidence over the ones that generate asset condition improvement. The client's CAFM shows green. The estate is not green.
There is a third outcome, which is that the contractor renegotiates. Variation claims, scope change requests, additional works quotations, and contract review meetings become the mechanism for recovering the pricing gap. The client ends up paying more than the original tender price for a service that was never properly scoped to begin with. The root cause is a labour model built on a standard that was not designed to be a cost model.
What a Defensible Labour Model Actually Looks Like
The alternative to SFG20-derived pricing is not guesswork. It is a structured methodology that uses SFG20 as the task frequency baseline and replaces the generic time component with site-specific inputs.
- Asset register verification: Identify the gap between the client-provided register and the likely actual estate. Apply a gap factor of 10 to 20 percent on most public sector contracts. Flag any asset classes where the register appears particularly weak.
- Task mapping from SFG20: Use SFG20 Red and Pink task codes as the statutory and business-critical baseline. Add Amber tasks where criticality assessment or client specification requires it. Do not add Green tasks by default — ask why each Green task is in scope.
- Access classification: Classify each major asset class by access complexity: direct, ceiling void, roof, confined space, restricted hours, or specialist zone. Apply an access multiplier to the SFG20 task duration for each class.
- Permit-to-work allowance: Identify which tasks require PTW on this estate. Add a permit administration allowance of 30 to 60 minutes per PTW task. Model the impact of permit queue times on peak demand periods.
- Productive time factor: Apply a productive time factor of 70 to 75 percent to theoretical annual hours. Adjust upward for a simple single-site contract. Adjust downward for a dispersed multi-site or high-permit estate.
- Reactive demand overlay: Estimate reactive demand as a proportion of engineer capacity. Benchmark against available incumbent data, or use contract intelligence to reference comparable contracts in the same sector and geography.
- Build the explicit risk pot: Document the assumptions that are uncertain. Price the risk of those assumptions being wrong. Do not hide it in contingency — show it as a named line item.
This methodology does not produce a perfect model. What it produces is a documented, auditable set of assumptions that can be reviewed, challenged, and defended. When the contract starts and the assumptions are tested against reality, the team knows exactly where the model was wrong and can manage the gap rather than discover it.
The Baachu Approach: Labour Loading Validation and Bid Pricing Support
Baachu's commercial consulting and bid support practice has worked on over 100 FM bids, from single-site contracts to framework portfolios worth over £1 billion. The labour loading problem described in this article is present on almost every contract we have worked on. The degree of damage it causes depends on how early in the bid process it is identified.
On engagements where we are brought in at tender stage, we build a site-specific labour model that replaces SFG20 generic timings with access-adjusted, permit-allowed, productive-time-factored hours. We use Baachu Rain's contract intelligence data to benchmark reactive demand profiles against comparable contracts in the same sector and geography. We identify the asset register gaps and price them explicitly rather than absorbing them into contingency.
On engagements where we are brought in after contract award — on a contract that is already losing money — the conversation is different. The labour model is wrong and the contract is live. The remedies are limited: scope negotiation, TUPE headcount review, task bundling, and route density optimisation. These can recover some of the gap. They cannot recover a contract that was mispriced by 15 to 20 percent before mobilisation began.
The clearest thing Baachu can offer on this issue is the same thing this article offers: make the invisible visible. The labour loading assumptions that every estimator makes in private should be in the model, documented, and defensible. If the model is wrong, knowing exactly where it is wrong is the difference between a managed problem and an unmanaged one.
How Baachu Can Help on Your Next Hard FM Bid
- Labour loading validation: we review your SFG20-derived hours model and apply site-specific access, permit, productive time and reactive demand adjustments.
- Asset register gap analysis: we assess the completeness of the client-provided register and build an explicit gap assumption into the pricing model.
- Reactive demand benchmarking: using data from 11,000+ UK FM contracts, we benchmark your reactive allowance against comparable contracts in your sector.
- Risk pot structuring: we identify the assumptions that carry the most pricing risk and structure an explicit, named risk provision rather than a generic contingency.
Contact: hello@baachu.com · Market intelligence: baachurain.com
Frequently Asked Questions
Productive wrench time is the proportion of a field engineer's working day actually spent performing maintenance tasks — tools in hand, on the asset. It excludes travel, permit administration, access preparation, tool logistics, documentation, and reactive demand. On a complex FM estate, productive wrench time typically falls between 1,300 and 1,400 hours per year, against a theoretical 1,800-hour working year.
BESA's own FAQ describes them as 'generic service timings' that vary by site, age, and equipment. They describe the task itself under ideal conditions — no access difficulty, no permit requirement, no asset condition issues. They do not include travel, permit administration, access preparation, or reactive demand. Using them directly as a pricing input without site-specific adjustments systematically underestimates the real cost of delivery.
A permit-to-work (PTW) allowance is additional time built into the labour model to cover the raising, authorisation, clearance, and closure of PTW documentation for tasks that require it. On complex estates, permit wait times can run to 40 to 60 minutes per task. A standard allowance of 30 to 60 minutes per PTW task should be included in the labour model for any estate with mandatory permit systems, and this time does not appear anywhere in SFG20 task durations.
The best benchmark is incumbent contract data — if available — showing the actual reactive demand volume and response profile. Where this is not available, contract intelligence from comparable estates in the same sector and geography provides the next best reference. Baachu Rain tracks 11,000+ UK FM contracts and can provide reactive demand benchmarks by sector, estate type, and contract structure to inform the reactive overlay in a bid labour model.
- 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 (this article)
- Art. 3 Who Owns SFG20, Who Profits from It, and What Happens If Ownership Changes?
- Art. 4 What Does SFG20 Compliance Actually Cost? Five Hard FM perspectives with specifics on each.
- Art. 5 Why Is the Hard FM Asset Register Always Wrong? TUPE, retendering, SFG20 mapping failures, and why 30 years of outsourcing has produced chaos, not continuity.
- Art. 6 Is There a Credible Alternative to SFG20? ISO 41001, condition-based maintenance, and the hybrid model.
- Art. 7 What Does Technology Do 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.
- Art. 9 SFG20 State of FM Report 2026: What the Data Actually Shows About SFG20 Compliance, Asset Registers and Hard FM Dependency.
Pricing a Hard FM bid and need a labour model that survives contact with the actual estate? → Talk to us at Baachu
Next: Article 3 · Compliance Is Not Strategy: Why Estates Need More Than a Schedule Library
Read Article 3 →
For construction cost benchmarking, the same question — how accurate is the standard data? — is examined in the BCIS series.