Route optimization is one of those terms in logistics that has been used so broadly it has started to lose meaning. A TMS that consolidates loads onto fewer trucks calls itself route optimization. A carrier selection tool that picks the cheapest lane calls itself route optimization. A manual coordinator who calls three carriers and picks the middle price is, technically, also doing route optimization.
The word covers everything, so it tells you nothing. What actually tells you whether your routes are well-optimized is a small set of operational metrics that most desks either do not track or track inconsistently.
Metric 1: Lane cost per TEU, rolling 90-day average
Yes, this is the obvious one. But the reason most desks underuse it is that they calculate it in aggregate or per-invoice rather than per lane per carrier over time. A rolling 90-day average per carrier per lane is the version that actually tells you something.
Spot rate volatility on North Europe-to-Iberia land freight can swing 15 to 25 percent across quarters based on backhaul demand and driver availability (industry-observable range, not a specific citation). If you are only looking at monthly totals, you miss the carrier that has been consistently 12 percent above market on one lane for eight weeks. The rolling 90-day per-lane view shows it immediately.
Metric 2: Planned-to-actual transit time variance by lane
This metric is distinct from on-time delivery. On-time delivery measures whether the cargo arrived before the committed customer window. Planned-to-actual transit time variance measures whether your route planning assumptions are accurate at the lane level.
If your planning tool assumes 72 hours for a Spain-to-Netherlands road move and your actual average is 84 hours, you have a planning accuracy problem. That 12-hour gap is buffer time the desk is systematically not building in -- which means loads that arrive at port with 12 hours less free time than expected.
High variance on a specific lane often points to a carrier reliability issue, a routing assumption that is no longer current (a regular motorway corridor with new congestion patterns, a transhipment port with degraded schedule reliability), or a planning tool that has not been updated with current carrier performance data.
Metric 3: Reroute cost premium vs. planned route cost
When a disruption forces a reroute, what does the rerouted load cost compared to what the original load would have cost if it had run as planned? This premium is your disruption cost baseline.
A well-optimized desk with good carrier relationships and fast rerouting capability typically sees a 10 to 20 percent cost premium on rerouted loads. Desks that reroute slowly and under time pressure -- because they have to call through a mental list of contacts before finding capacity -- see premiums of 35 to 60 percent, because by the time they confirm a carrier they are in shortage territory and paying for urgency.
If you can bring your disruption-to-reroute time below 15 minutes, you almost always get a better rate because more carrier options are still genuinely available. Wait 90 minutes and the options have narrowed and the carriers know you are in trouble.
Metric 4: Empty leg rate across your carrier allocations
What fraction of your carrier-committed capacity is running empty or significantly underloaded? This metric only applies if you have some level of committed carrier relationships (rather than fully spot-buying), but for any desk managing contractual capacity it is critical.
Empty legs represent both a direct cost (you are often paying for capacity you are not filling) and an indirect cost (carriers that consistently run empty on your lanes will deprioritize you when you need flexible capacity during disruptions). The target range depends on your lane structure, but for mixed European road/sea freight operations, an empty leg rate above 18 to 22 percent of committed capacity typically indicates a load consolidation problem, not a carrier problem.
Metric 5: Time-to-first-alert on disruption events
This metric is often overlooked because desks conflate disruption monitoring with disruption response. Time-to-first-alert measures how quickly your monitoring setup flags an event after it becomes detectable in the data. Disruption response time measures what the desk does after the alert fires.
Poor time-to-first-alert means the desk finds out about a port slot closure 40 minutes after it happened because someone checked the terminal portal manually. Good time-to-first-alert means the desk gets an automated notification within three to five minutes of the event being registered in the port authority's system.
The value of fast alert time is not just that it gives you more response time. It also shifts the desk's mental model from reactive to anticipatory. A coordinator who has been getting three-minute alerts for three months starts reading the pattern: she knows that a vessel ETA shift of more than five hours on the Antwerp approach typically generates a slot cascade, and she starts looking at alternative carriers before the formal alert fires on the secondary loads.
Reading the five together
| Metric | What it tells you | What to watch for |
|---|---|---|
| Lane cost per TEU (90-day rolling) | Carrier rate accuracy | Persistent outlier carriers on specific lanes |
| Planned-to-actual transit variance | Planning assumption accuracy | Systematic underestimates on specific lanes |
| Reroute cost premium | Disruption response speed value | Premiums above 30% signal slow rerouting |
| Empty leg rate | Load consolidation efficiency | Above 18-22% suggests consolidation gaps |
| Time-to-first-alert | Detection sensitivity | Over 20 minutes = manual detection dependency |
The combination that indicates a genuinely well-optimized route network: lane cost close to market, low transit variance, reroute premiums under 20 percent, empty leg rate in a healthy range, and alerts firing within five minutes. Any single metric looking good while others are poor usually points to a specific process bottleneck rather than an overall route optimization problem.
One thing these metrics do not measure
They do not measure carrier relationship quality directly. A desk with strong long-term carrier relationships can sometimes absorb a disruption that these metrics would flag as a problem -- because a carrier calls to offer capacity before an alert fires, or because the desk has informal rate flexibility that does not show up in the per-TEU calculation. Qualitative factors matter.
But carrier relationships do not substitute for systematic measurement. The desks we have seen work best use both: strong relationships as a buffer and systematic metrics as the diagnostic tool that tells them which carrier relationships are actually delivering and which are coasting on historical goodwill.