Two operations with similar vehicle counts can rationally choose opposite planning models. The mistake is not choosing wrong — it is choosing by default, usually copying whatever the software's demo did.
The two models
Fixed (master) routes assign customers to a route and day pattern that repeats. Monday's route 7 visits broadly the same customers every Monday. Optimisation happens periodically, not daily.
Dynamic routing re-plans from the current order set each planning cycle. Customers may fall on different routes and different sequences from one day to the next.
Most real operations sit somewhere between: a fixed backbone with a dynamic overlay, or fixed territories with dynamic sequencing inside them.
Take the last 12 weeks of orders. What percentage of customers appear in the same week-day slot every week?
What each model buys you
| Dimension | Fixed routes | Dynamic routing |
|---|---|---|
| Raw efficiency | Lower — plan cannot react | Higher — matches actual demand |
| Service consistency | High — same driver, same time | Lower unless constrained |
| Driver familiarity | High | Low to moderate |
| Planner workload | Low daily, high periodic | Moderate daily |
| Settlement and accounting | Simple, stable | More complex |
| Absorbing volume spikes | Poor | Good |
| Data quality demands | Moderate | High |
| Change management | Easy | Harder |
When fixed routes are the right answer
- Demand is stable and predictable. Vending, bottled water, linen, most pest control, many waste rounds.
- The driver has a commercial role. If the driver sells, merchandises, collects payment or holds the customer relationship, continuity is worth real money.
- Customers expect the same time each visit. A café expecting delivery before the 08:00 rush cares more about consistency than about your route efficiency.
- Settlement is per route. Route accounting operations reconcile stock and cash per route; stable routes make variance analysis meaningful.
- Your data is not good enough for daily planning. Honest self-assessment: dynamic planning with poor geocodes and fictional service times produces worse results than a stable route a driver knows.
When dynamic routing wins
- Order sets change materially day to day. E-commerce delivery, courier, wholesale where the customer orders when they need something.
- Volumes are volatile. Seasonal peaks are absorbed by adding vehicles into a re-planned network rather than by overloading fixed rounds.
- Time windows dominate. Booked slots are hard to fit into fixed structures.
- You have same-day or add-on work. Dynamic re-planning during the day is only possible if the base plan is dynamic.
- Geographic coverage is sparse. With low stop density, the efficiency gap between fixed and dynamic widens sharply.
The hybrid patterns that usually win
Fixed territory, dynamic sequence. Customers belong permanently to a territory (preserving relationships and settlement) but the daily order is optimised. Low risk, captures most of the sequencing gain, very common in DSD.
Fixed backbone, dynamic overflow. Core customers on master routes; overflow, new customers and one-off work planned dynamically onto a flexible fleet. Handles growth without destabilising established rounds.
Dynamic with continuity constraints. Full daily optimisation, but with a penalty for changing which driver serves a customer. The engine keeps continuity where cheap and breaks it where the cost is justified. Requires a product that supports soft continuity constraints — a specific thing to ask about in evaluation.
Periodic re-optimisation. Fixed routes rebuilt on a schedule (quarterly, or after any 10% shift in volume). Cheap to run, and captures most of the value if your demand drifts slowly.
The transition problem
Moving from fixed to dynamic is a change management project, not a software project. Expect:
- Driver resistance, strongest from the longest-serving drivers with the best local knowledge. Their objections are usually specific and often correct; capture them as constraints.
- Customer complaints about a new face or a different time. Communicate before, not after.
- A short-term efficiency dip as the organisation learns and the data improves. Budget six to twelve weeks.
- Exposed data problems. Fixed routes hide bad data because the driver compensates from memory. Dynamic planning surfaces every wrong geocode immediately.
That last effect catches people out. It is not the new system failing; it is the old data being seen for the first time.
Frequently asked questions
Can we run different models in different depots?
Yes, and it is often correct. A dense urban depot with volatile e-commerce volume and a rural depot serving stable trade customers have genuinely different problems. Consistency of approach across a network is an aesthetic preference, not an operational requirement.
Does dynamic routing mean drivers never see the same customers?
Only if you configure it that way. Continuity constraints, territory boundaries and driver-preference weightings let you keep most relationships intact while still gaining flexibility. Ask vendors how they express this before assuming it is impossible.
How much efficiency do fixed routes give up?
Typically 5–15% in operations with variable demand, and close to nothing where demand is genuinely stable. Measure it in your own data by re-optimising four historical weeks unconstrained and comparing to what you actually ran.
What about same-day orders?
Same-day work effectively requires dynamic capability, at least for a portion of the fleet. A common structure is a fixed base fleet plus a smaller flexible fleet that absorbs same-day and urgent work.
Which model is easier to settle and audit?
Fixed routes, clearly. Stock, cash and performance variance are all easier to interpret when the comparison unit is stable. If your business runs sell-from-truck settlement, weigh this heavily — see our route accounting guide.
