On-time measured against the wrong promise
Performance reported against the internal plan rather than the commitment given to the customer, so internal reporting is green while the customer escalates.
Logistics
Shipment-level cost to serve and service measures defined against the contractual commitment.
Logistics generates enormous quantities of event data and very little agreement about what it means. Three systems record a delivery; each has its own timestamp and its own idea of a successful outcome, and the customer has a fourth.
- Service performance figures the customer recognises and accepts
- Margin visible by customer, lane and service level
- Subcontracted volume measured on the same basis as own fleet
- Fuel, energy and emissions reporting produced from operational data rather than estimates
We build a shipment-grain model where each leg, event and exception is represented once, costs are allocated to the shipment rather than the depot, and service measures are defined the way the customer-facing commitment is written. That turns operational reporting into something that can be used in a commercial conversation.
We work with freight forwarders, hauliers and parcel carriers, third-party logistics providers, port and terminal operators, and the supply chain functions of retailers and manufacturers.
What goes wrong
Recurring failure modes in logistics
Cost allocated to the depot, not the shipment
Without shipment-level cost allocation, loss-making customers and lanes are invisible, and pricing is set on averages.
Event data with inconsistent timestamps
Scanner, telematics and TMS timestamps differ in time zone, precision and meaning. Reconciling them is unavoidable before any duration measure is trustworthy.
Subcontractor performance unmeasured
Significant volume moved by partners whose performance data arrives as spreadsheets, so network performance has a blind spot exactly where control is weakest.
Metrics
Metrics and their definition errors
Each metric with its definition and the error that most often distorts it. If our reading is wrong for your business, that disagreement is a productive starting point.
| Metric | Definition | The trap |
|---|---|---|
| On-time in full | Deliveries arriving complete within the committed window, over total deliveries. | Which promise counts (original, revised or customer-agreed) must be a labelled, separate measure for each basis. |
| Cost to serve | Fully allocated cost of fulfilling a shipment, customer or lane. | Allocation of empty running, waiting time and failed deliveries is where most of the real variance lives. |
| Fleet utilisation | Revenue-earning vehicle hours or loaded kilometres over available capacity. | Define availability against planned rather than calendar time, or maintenance looks like poor performance. |
| Pick productivity | Units or lines picked per labour hour, by area and shift. | Indirect labour and travel time must be included or excluded consistently; otherwise area comparison is meaningless. |
| Dwell time | Time between arrival and departure at a site or terminal. | The arrival definition (gate, yard or dock) changes the figure substantially and must be stated on the report. |
Use cases
What we typically build
Service performance reporting
OTIF and exception reporting at shipment grain with root-cause categorisation, plus customer-facing views that match the contractual commitment.
Cost-to-serve model
Allocation of linehaul, handling, last-mile, waiting and failure cost to shipment level, exposing margin by customer, lane and service level.
Fleet and telematics analytics
Utilisation, fuel and energy consumption, driver hours compliance and maintenance cost per kilometre, from telematics joined to the operational plan.
Warehouse productivity
Labour productivity by area and shift, dock-to-stock time, inventory accuracy, and the effect of slotting changes measured rather than assumed.
Network and capacity analysis
Flow, imbalance and seasonality analysis to inform depot footprint, trailer allocation and subcontractor mix decisions.
Customs and compliance reporting
Declaration volumes, holds and clearance times across borders, with the documentation trail needed for audit.
Integration
Source systems we connect to
- Transport management systems
- Warehouse management systems
- Telematics and tachograph data
- Handheld scanners and mobile apps
- EDI messages (ORDERS, DESADV, IFTSTA)
- Customs and declaration platforms
- Subcontractor portals and files
- Fuel cards and maintenance systems
Regulatory context
Regulatory context by market
- UK and EU drivers' hours and tachograph record-keeping; UK Operator Licence undertakings
- EU Mobility Package posting and cabotage record requirements
- Customs: UK CDS, EU import control systems, transit documentation retention
- Switzerland: LSVA heavy vehicle charge records and cross-border documentation
- Belgium: Mobility and kilometre charge reporting for heavy goods vehicles
- Sustainability: CSRD Scope 3 transport emissions, GLEC framework reporting
Questions
Logistics: common questions
By building a single event model with a canonical event type, source, confidence and normalised UTC timestamp, plus a documented precedence rule stating which source wins for each event type. Durations are then always calculated from canonical events, so two reports cannot disagree about transit time.
Yes. Where partners have no API, we build a controlled file or portal ingestion with schema validation and a quality gate, so bad submissions are rejected and visible rather than silently averaged in. Making partner data quality measurable is often the first improvement clients notice.
Operational tracking belongs in the operational system; analytics should be near-real-time at best and does not need to be a live feed. We typically land events every few minutes for operational dashboards and keep the analytical model at a lower frequency. Streaming the whole estate rarely survives a cost review.
From activity data (distance, mode, load weight, fuel or energy consumed) using an agreed emissions factor set, with the factor version stored against each calculation so figures can be restated when factors change. That versioning is what makes the number defensible under CSRD assurance.
Services that apply here
Next step
Analytics work in logistics
A 45-minute working session with an engineer rather than a sales call. You describe the problem; we set out what we would build, in what order, and what it is likely to cost.
Answered within one working day, by the engineer who would do the work