The 4 Pillars of Analytics for Construction Teams

Analytics is often talked about as if it’s a single thing: dashboards, KPIs, graphs. On a UK construction site, though, analytics only becomes useful when it helps you make better decisions—earlier—about labour, materials, programme, quality and risk.

So, what are the 4 pillars of analytics? In practice, they’re four levels of maturity that build on each other:

1. Descriptive analytics – What happened?
2. Diagnostic analytics – Why did it happen?
3. Predictive analytics – What is likely to happen next?
4. Prescriptive analytics – What should we do about it?

In this post we’ll break each pillar down in plain terms, show how it applies to construction analytics, and give real site examples—plus how to implement each pillar using SiteSamurai without turning your project team into data scientists.

---

## Pillar 1: Descriptive analytics (What happened?) Descriptive analytics is the foundation. It summarises what has already occurred using clean, consistent data.

• “I think we’re behind” and
• “We’re 6 days behind on Level 3 drylining, with 42% of rooms signed off.”

Construction analytics examples (descriptive)

- Daily progress: completed tasks vs planned tasks
- Labour and plant: hours booked vs hours planned
- Quality: number of snags raised/closed, average close-out time
- H&S: near misses, inspections completed, actions overdue
- Commercial: variations logged, RFIs raised, potential compensation events

Site example: refurbishment project in Manchester

On a city-centre refurbishment, the site team were relying on a whiteboard and weekly lookahead notes. The PM felt the project was “drifting”. Once daily updates were captured consistently, the descriptive view showed:
- Ceiling grid installation was consistently under target by 15–20% per day
- Snag close-out was slowing as areas were handed over

That clarity didn’t fix anything by itself—but it stopped arguments and created a single version of the truth.

How to do it in SiteSamurai

Practical steps:
- Standardise your daily records (progress, labour, delays, photos)
- Use structured fields (trade, location, task, status) so reporting is reliable
- Build simple views: planned vs actual, open vs closed, by subcontractor, by zone

Descriptive analytics should be quick to produce and easy to trust. If it takes hours of spreadsheet work, it won’t happen consistently.

---

## Pillar 2: Diagnostic analytics (Why did it happen?) Diagnostic analytics goes one step further: it explains the drivers behind the numbers.

• constraints (access, permits, drawings)
• sequence clashes (M&E vs ceilings vs fire stopping)
• rework (failed inspections, incorrect installs)
• supply chain issues (late deliveries, substitutions)
• productivity blockers (waiting time, incomplete areas)

Construction analytics examples (diagnostic)

- Correlating delays with RFIs or design changes
- Linking snag rates to specific subcontractors, plots, or work packages
- Identifying which locations repeatedly fail inspections
- Analysing how often a task is blocked by “area not ready”

Site example: new-build residential in Birmingham

A site was missing weekly targets on first-fix M&E. Descriptive analytics showed the shortfall; diagnostic analytics revealed the pattern:
- Most lost time occurred in cores and risers
- The reason codes were consistently “access not available” and “permit not issued”

That pointed to a process issue (permit workflow and access coordination), not an “M&E productivity” problem. The fix was to tighten the permit turnaround and coordinate access windows by floor.

How to do it in SiteSamurai

To get diagnostic value, you need consistent categorisation:
- Log delay reasons (e.g., design, access, materials, labour, weather)
- Track RFI status and link it to affected locations/tasks
- Use inspection outcomes and failure reasons (e.g., tolerances, fire stopping detail, incomplete works)
- Attach photo evidence to avoid debate later

The goal isn’t to blame—it’s to remove friction from delivery.

---

## Pillar 3: Predictive analytics (What is likely to happen next?) Predictive analytics uses patterns in existing data to forecast future outcomes.

• production rates
• backlog size
• recurring constraint types
• trend lines (snag closure rate, inspection pass rate)

Construction analytics examples (predictive)

- Forecasting date of completion for a work package based on current output
- Predicting handover readiness by tracking inspections and snags per area
- Identifying subcontractor performance risk (e.g., declining productivity trend)
- Predicting cost risk where variation volume is rising faster than approvals

Site example: school extension in Leeds

A contractor tracked weekly door set installations. The programme assumed 25 door sets/week. Actual output averaged 18/week for three consecutive weeks.

• resequencing (prioritising corridors for fire strategy sign-off)
• bringing in an additional fixing gang
• locking down delivery slots to avoid missing ironmongery

Without prediction, the risk would have surfaced too late—during commissioning and handover.

How to do it in SiteSamurai

You can create practical predictive signals by:
- Tracking planned vs actual productivity rates per trade
- Monitoring leading indicators (RFIs ageing, inspections booked vs completed, constraint log volume)
- Setting thresholds for alerts (e.g., “if snag close-out rate < snag creation rate for 2 weeks, flag risk”)

Predictive analytics turns reporting into a forward-looking management tool.

---

## Pillar 4: Prescriptive analytics (What should we do about it?) Prescriptive analytics recommends actions. It’s the pillar that turns insight into decisions—what to change, who should do it, and by when.

• prioritised action lists
• scenario planning (option A vs option B)
• resource reallocation recommendations
• automated workflows and escalations

Construction analytics examples (prescriptive)

- If concrete pours are slipping, recommend resequencing and increasing formwork turnaround
- If inspection failures spike in a zone, trigger targeted toolbox talks and hold points
- If RFIs are ageing beyond SLA, escalate to design manager and notify affected work packages
- If a subcontractor falls below productivity threshold, trigger a recovery meeting and revised method statement

Site example: commercial fit-out in London

A fit-out project had repeated ceiling inspection failures due to fire stopping details around services penetrations.

• Identify the top failure reason and locations
• Recommend a corrective action: a dedicated fire stopping supervisor for two weeks, plus a revised inspection checklist
• Implement a hold point: ceiling close-up only after photo-verified fire stopping sign-off

The result was fewer re-inspections, less rework, and a smoother path to practical completion.

How to do it in SiteSamurai

Prescriptive analytics becomes achievable when you combine insights with workflows:
- Convert flagged risks into assigned actions with owners and due dates
- Use standard operating checklists for recurring issues (e.g., pre-pour checks, first-fix quality gates)
- Automate escalations when actions are overdue or thresholds are breached
- Produce a weekly “Top 5 risks” view linked directly to evidence and tasks

This is where construction analytics pays for itself—by preventing avoidable delays and rework.

---

## How the 4 pillars work together on a live site Think of the pillars as a ladder:

• Descriptive tells you the facts (progress, defects, delays).
• Diagnostic tells you the causes (constraints, rework drivers, coordination issues).
• Predictive tells you what will happen if nothing changes (forecast slippage, handover risk).
• Prescriptive tells you what to do next (actions, resequencing, resource changes).

A common mistake is trying to jump straight to predictive dashboards without sorting data capture. If your daily records and categorisation are inconsistent, prediction becomes guesswork.

---

## Practical implementation plan for construction teams (without the fuss) If you want to build your construction analytics capability in SiteSamurai, use this staged approach:

1. Weeks 1–2: Nail descriptive

1. Weeks 3–4: Add diagnostic codes

1. Month 2: Introduce prediction

1. Month 3: Make it prescriptive

---

## Final takeaway: the 4 pillars of analytics, in construction terms If you remember nothing else, remember this:

• Descriptive: What did we build this week?
• Diagnostic: What stopped us building more?
• Predictive: Where will we be in four weeks at this rate?
• Prescriptive: What do we change on Monday to protect programme and margin?

With SiteSamurai, you can capture the right site data once and use it repeatedly—first for reporting, then for root cause analysis, then for forecasting, and finally for action-driven management.

If you’d like, share your project type (fit-out, civils, resi, education, healthcare) and I’ll outline a set of practical KPIs and dashboards mapped to these four pillars.