Dynamic replanning in DIF: managing 12-hour mandatory delivery, urgencies and controlled substances without blowing up the route

In Italian pharmaceutical intermediate distribution there is a service rule that orients everything else, and which those coming from adjacent sectors — express courier, food distribution, e-commerce last mile — struggle to frame: the intermediate distributor has 12 working hours from the pharmacy’s request to deliver the drug. It’s a regulatory constraint (Italian decree 219/2006 on “continuous and appropriate” service) and not a commercial service level. It’s the binding constraint. There is no premium customer paying more to have it in 6 hours: everyone has the right to 12 hours.

A note for non-Italian readers: the 12-hour rule is a specificity of the Italian pharmaceutical wholesale system, but the underlying operational dynamics (binding service constraints, urgency-driven intraday events, regulated substances handling) are observable in all mature European pharma supply chains with country-specific differences.

ADF — the Italian Pharmaceutical Distributors Association, 35 companies serving 19,000 Italian pharmacies — declares in its association data that the real delivery average of its members is approximately 3 hours. Four times below the legal obligation. It means the system works well beyond the standard, and that the competitive advantage between distributors is not “making the 12 hours” but “reducing the 3-hour average when it really counts”.

To this add two operational specifics that change everything compared to generalist transport. First, on controlled substances the Italian Ministry of Health imposes a clear rule: no partial deliveries. If the stock isn’t complete for that request, only what’s available departs and the rest is lost — it cannot be completed in a subsequent passage. Second, pharmacy urgencies can activate at any moment of the day: a life-saving drug requested at 14:30 must depart at 14:35.

All this clashes with a fact: the day’s fixed route was planned the evening before. Perfect Monday 8:00 AM planning — even with the best optimisation software available — gets disrupted by 5-20 unexpected events during the day. The difference between a distributor absorbing these events without degrading service and one paying their hidden cost is the discipline of dynamic intraday replanning: continuous recalculation of the route while the route is happening.

This article addresses how dynamic replanning works in DIF, what typical events activate it, how to decide what to redo and what to leave, and why — increasingly — it’s a capability of planning systems, not just of the human dispatcher.

The regulatory constraint: 12 hours, urgencies, controlled substances

Italian decree 219/2006 is clear: the authorised pharmaceutical wholesaler must provide “continuous and appropriate service” to pharmacies in its territory, with delivery within 12 working hours of the request. For molecules unavailable in the regional network, the deadline is 48 hours with manufacturer supply. Management violations (un-updated registers, declared-vs-actual discrepancies) are penalised €500-1,500 per episode — figures not individually dramatic but which, summed across an AIFA (Italian Medicines Agency) audit or a local health authority inspection on 200-300 served pharmacies, can accumulate significantly.

On this basis, Italian intermediate distributors — pushed by competition among them and by pharmacist expectations — have progressively shortened average delivery times to arrive at the 3 hours declared by ADF. It’s an implicit service promise: the pharmacy orders and expects the goods to arrive within half a day.

On controlled substances (Table II, Table III, some psychotropics) the rule is even more stringent. Regulations on the entry/exit register of controlled substances and the lot-by-lot traceability obligation require that deliveries be complete or not made: the distributor can deliver only what’s in stock at the moment of the request, and missing items cannot be recovered in a second trip. For the pharmacy it means that, if 3 morphine vials are needed for a terminal patient and the distributor has only 2, the pharmacy must try to complete the order with another wholesaler; there is no “one will come tomorrow from the same”. For the distributor, it means that every controlled-substances order must be validated on actual stock before departing, and that controlled-substances delivery tour planning is structurally more rigid than ordinary one.

To these two constraints add urgency management: the life-saving drug (chemotherapy agents, septic-shock antibiotics, antidotes, some biologics with time-critical administration) requested by a pharmacy intraday or by a hospital must be delivered in times often well below one hour. It’s not “preferred”: it’s necessary.

Static planning vs dynamic replanning

The traditional DIF planning model, inherited from 1990s systems and still prevalent in mid-market today, is static. It works like this: the evening before, the dispatcher (or the optimisation system, but in nightly batch) generates the next day’s routes based on orders received within a certain cut-off. Vans leave in the morning with the drop sequence already fixed. From the moment they depart, route modifications are manual, based on radio or phone calls between dispatcher and driver.

On sector benchmarks, this model produces 5-20 disruptive events per day for every 100 served pharmacies. It means a standard route undergoes, on average, between 1 and 4 deviations from the original plan during the day. Manual management of these deviations:

• Requires time from the dispatcher (calls, quick decisions, communication to the driver)
• Generates sub-optimal choices (the closest driver is not always the one with least impact on the rest of the route)
• Produces informal communication (radio, WhatsApp, calls) not tracked, which doesn’t end up in the system
• Creates operational friction when hot decisions conflict with other subsequent events

Dynamic replanning is the opposite model: the system maintains in real time the route state (where drivers are, what they’ve already delivered, how much load is left, how much time available), continuously receives new events (urgencies, order modifications, stock unavailability), and proposes (or automatically executes within predefined constraints) the optimal reallocation. The dispatcher becomes supervisor of a system reacting much faster than they could manually.

It’s the pattern the B2B logistics sector is progressively codifying as logistics control tower: the human dispatcher stays in the control room but supervises a unified system that sees the entire route in real time — vehicle positions, completed drops, incoming urgencies, active time constraints — and proposes or applies optimal reallocations. The difference from the pre-digital model is not the introduction of automatic control replacing the dispatcher, it’s the introduction of a unified vantage point the dispatcher can govern.

On DIF realities that have made this jump — and among Optivo clients it’s the 2025-2026 emerging pattern — the operational effect is significant: dispatcher time freed by 30-50% on event management, percentage of deliveries within 12 hours generally above 99%, average urgency reaction time halved.

The three typical events triggering intraday replanning

Not everything changes equally during the day. Three categories of disruptive events cover 90-95% of cases in DIF.

Category 1 — Pharmacy urgency (off-route delivery)

The pharmacy calls (or sends electronic order) for a drug not foreseen in the day’s route. Typically it’s a life-saving drug (chemotherapy, shock-dose antibiotic, antidote, biologic), or a product of absolute necessity for a specific patient that the pharmacist must provide intraday.

The question to the system: which driver can deliver with least impact? Not always the geographically closest — it can be the one who has already completed the critical part of their route and has slot capacity, or the one already passing near the depot for reload. The system must simultaneously look at: distance, residual capacity, transit time, impact on the candidate driver’s original route’s subsequent drops.

Category 2 — Manufacturer stockout (substitution or cancellation)

The DIF warehouse signals that a drug foreseen for delivery is not in stock at van preparation time (the manufacturer delayed supply, or it was consumed by earlier urgencies). That delivery’s planning must be updated: the drug doesn’t depart, and the pharmacy must be informed so it can search elsewhere or wait. If it’s a controlled substance, the “no partial” rule applies: only what’s available of the rest of the order departs, missing controlled substances skip.

The system must update in real time the van’s manifest, generate communication to the pharmacy, and — if the delivery was critical — flag it to the dispatcher for follow-up.

Category 3 — Pharmacy order modification within window

The pharmacy, after seeing the morning situation, asks to add or remove references from the order already in preparation. It’s a routine operation that however — if not managed with an integrated system — ends up generating informal communications among pharmacy, phone operator, warehouse and dispatcher with high error rates.

The system must handle the modification in real time, update the van manifest if it hasn’t departed yet, or reject the modification if the van is already en route with the drop imminent.

How replanning works: the operational logic

When a disruptive event arrives, the system must decide in a few seconds. The operational logic we see working on DIF clients articulates in four steps.

Step 1 — Event classification. Is it a pharmacy urgency (category 1)? A stockout (category 2)? An order modification (category 3)? Each category has different handling logic.

Step 2 — Optimal candidate identification. For category 1 (urgency): which driver can deliver with minimum impact on the overall route? The system evaluates in real time GPS position of drivers, residual capacity of vans, orders still in queue, maximum depot return time. For category 2 (stockout): what communication is needed? For category 3 (modification): is it still manageable, and with what minimum latency to the drop?

Step 3 — Overall impact calculation. Reallocating an urgency to driver A means delaying their subsequent 4 drops by X minutes each. The system evaluates: does this delay trigger the 12-hour threshold on one of the drops? Does it create criticality on tight time windows (hospitals, clinics with opening hours)? The trade-off is always between the new urgency and existing service.

Step 4 — Decision and communication. If the system can decide autonomously within constraints predefined by the dispatcher, it acts: updates the driver’s manifest, sends the update to their app, communicates with the requesting pharmacy. If the decision is ambiguous or exceeds constraints, it proposes 1-2 alternatives to the dispatcher with respective impacts, and the dispatcher decides manually with the data in front.

The leap over the traditional manual model is qualitative. The dispatcher doesn’t manage each event individually with phone calls: they manage system exceptions. The number of events manageable without degrading service grows 3-5 times; communication error risk reduces significantly.

Effect on the overall route: the hidden cost of disruptive events

Every disruptive event has a hidden cost on the overall route that the static model typically doesn’t measure, and that the dynamic model makes visible and optimisable. On a typical route of 25 drops in urban area, a single urgency reallocation:

• Typically adds 4-8 km and 12-20 minutes of driving to the driver who takes it on
• Costs €4-8 in fuel and ~€5-10 in driver time
• Pushes the original route’s subsequent drops forward by 4-8 minutes each
• On tight time windows, can force an additional re-routing

On 5-20 events per day for 100 served pharmacies, the cumulative effect is significant: typically 80-200 extra km per day and 4-10 additional person-hours compared to a route perfectly executing morning planning. On annual basis, for a wholesaler with 200-300 served pharmacies, that’s €20-50,000 of additional operating costs that static planning doesn’t see and doesn’t optimise.

The dynamic system reduces this hidden cost by 30-50% on sector benchmarks, because every reallocation is done based on the truly optimal candidate (not the closest on the map) and because impacts on subsequent drops are compensated in real time by redistributing residual load among more drivers.

The strategic value: beyond operational efficiency

There is a strategic dimension to dynamic replanning beyond operational savings. It’s the service promise capability.

A wholesaler operating with dynamic replanning can confidently promise sub-threshold service levels (e.g. 2-3 hour average delivery, sub-hour urgencies) because they have the tools to maintain the promise even when the day gets complicated. A wholesaler operating only with static planning must maintain wider safety margins (oversize capacity) or accept occasional risk of degradation on tough days.

On the commercial ground, this translates into two concrete advantages. First, in negotiations with demanding pharmacies — chains, hospitals, local health authorities — the ability to promise tight SLAs and maintain them is a sales differentiator. Second, on retention of historical pharmacies, service consistency (not a peak of excellence, but quality constancy) is what keeps the customer.

Replanning as technological capability: what’s needed

Three technological components that — if absent — make dynamic replanning impossible to do well.

Real-time GPS position of all vehicles, with accuracy below 30 metres and latency below 15 seconds. It’s the base. Without current position, every reallocation decision is “blind”. Optivo clients who have OptivoTrack enabled have this base data.

Bidirectional driver app. The driver must be able to receive real-time updates of their own manifest (drop added, drop removed, sequence modified) and confirm the state of each drop (in-load, delivered, exception). Without bidirectional app, every change requires a phone call.

Optimisation engine running in near-real time. It’s not enough to have data and the app: a system is needed that, given the event, calculates the optimal candidate in 5-15 seconds. On new-generation engines (including OptivoRoute’s), this is the standard performance. On legacy systems, recalculation takes minutes or anyway requires manual intervention.

All three components are integrable separately with existing systems — the best-of-breed architecture pattern we discussed in detail after Transpotec 2026. For a DIF wholesaler, the first step is typically the introduction of the optimisation engine with dynamic replanning, on top of telematics and a driver app that may already be in-house.

KPIs of dynamic replanning

Frequently asked questions

Does dynamic replanning replace the dispatcher?

No, it frees them. The dispatcher remains indispensable for exceptional decisions (those exceeding the system’s predefined constraints), for pharmacy communication, for non-standardised situation management. What changes is the share of their time dedicated to routing micro-decisions that can be automated: typically goes from 60-70% of the day to 20-30%. Free time serves value-added activities: relationship with most critical customers, medium-long optimisations, quality management.

Does it work for wholesalers with few vehicles (10-15)?

Yes. The value of dynamic replanning grows with fleet size because there are more possible candidates for each reallocation. But even on 10-15 vehicles the advantage is significant, because on fleets of this size the dispatcher is typically unique and already at 100% load — freeing 30-50% of their time also frees client management capacity.

How much does integrating a dynamic replanning system cost?

It depends on telematics and driver app already present. If the distributor already has decent telematics with real-time GPS data and a POD app, adding the optimisation engine with dynamic replanning typically pays its annual cost in 4-8 months on km and dispatcher time savings alone. If telematics and app integration is also needed, payback extends to 9-12 months.

How are controlled substances handled in the replanning system?

With rigid pre-coded rules. Morning controlled-substances stock is a locked variable: the system can’t partially “cut” a controlled-substances order to free capacity. Controlled-substances reallocations among drivers during the day are possible only for complete drops that remain complete. It’s a rule the replanning system knows and respects automatically.

Does the system decide autonomously or does the dispatcher always need to intervene?

Configurable. Common practice is: autonomous decision for events within predefined constraints (e.g. acceptable delay <15 minutes on subsequent drops, extra distance <8 km, no impact on critical time windows), proposal to dispatcher for events exceeding constraints. Configurations are customisable per area, per customer, per event category.

In summary

In Italian pharmaceutical intermediate distribution, the 12-hour legal constraint, ADF’s 3-hour average promise, intraday pharmacy urgencies and the “no partial” rule on controlled substances compose a service system that morning static planning is no longer enough to sustain.

Dynamic replanning — continuous route recalculation while the route is happening — is the operational capability that allows absorbing 5-20 disruptive events per day without degrading service and without blowing up dispatcher time. On sector benchmarks, the dynamic model reduces additional km due to events by 30-50%, frees 30-50% of dispatcher time, brings the percentage of deliveries within 12 hours stably above 99.5%.

The economic impact is twofold. Operationally, on a 200-300 pharmacy wholesaler that’s €20-50,000 of annual operating costs that become visible and optimisable. Strategically, it’s the capability that allows sustaining sub-threshold service promises (3-hour averages, sub-hour urgencies) which become commercial differentiators.

If you want to understand how dynamic replanning integrates with your current telematics and driver app infrastructure, and how much dispatcher time you recover on your fleet size, talk to our team. Three months of real operational data are enough to build the impact projection.