Solutions

The last mile is not one problem

A grocery slot, a pharmacy cold-chain consignment and a pallet going to a loading dock share almost nothing except a vehicle. Kelvara ships configured templates for each — service rules, proof requirements, exception handling and reporting already in place.

E-commerce fulfilment

Volume that quadruples in November

Retail parcel flow is spiky, address quality is poor, and margins are thin enough that a second delivery attempt erases the profit on the order. Kelvara attacks all three: normalising addresses on intake, offering locker diversion at checkout, and planning to a cost ceiling rather than a fixed route template.

Checkout promise that reflects real remaining route capacity
Automatic locker diversion when a doorstep attempt is uneconomic
Peak-season surge planning with temporary depots and hired vans
Returns collected on the same route as outbound deliveries

Case study

Nordreta

Online homeware retailer · Vilnius, Riga, Tallinn · 1.8M parcels a year

Nordreta ran three regional couriers under separate contracts and had no single view of whether a parcel had actually arrived. Peak weeks pushed first-attempt success below 80%, and the customer-service team was reconciling three tracking portals by hand.

After moving planning and tracking onto Kelvara, Nordreta consolidated onto one dispatch board, added locker diversion for parcels under 8 kg, and started quoting two-hour windows in Vilnius. The customer-service load fell faster than the delivery cost did.

94.1%

first-attempt success, up from 79.6%

−17%

cost per parcel

−41%

"where is my order" contacts

Case study

Vilkas Foods

Online grocery · Kaunas and Klaipėda · 4,000 orders a day

Grocery is the hardest last mile in the Baltics: two-hour windows, three temperature zones in one tote stack, and a customer who is genuinely inconvenienced if you are late. Vilkas was hitting its slots 88% of the time and paying for it in credits.

Kelvara replanned their evening peak against learned service times per building type — the single biggest source of error in their old model, which assumed four minutes everywhere. Slot adherence moved without adding a vehicle.

97.4%

two-hour slot adherence

+2.3

drops per vehicle-hour

extra vans at peak

Grocery

Narrow slots, chilled goods, no second chances

Grocery planning is dominated by service time, not driving time. Kelvara models the walk from the kerb to the door — stairs, lifts, intercoms, tote returns — because that is where a two-hour slot is actually lost.

Multi-temperature tote allocation across chilled, ambient and frozen
Slot-availability API so the storefront never sells a window you cannot serve
Substitution and refusal handling captured at the door
Empty-tote collection scheduled back into the return leg

Pharma cold-chain

Custody you can hand to an inspector

Pharmaceutical distribution does not just need the parcel to arrive; it needs a defensible record that the parcel stayed within its temperature range and changed hands only between authorised people. Kelvara treats the evidence chain as the product.

Continuous temperature logging bound to the consignment record
Excursion alerts that open an exception before the van reaches the pharmacy
Named-recipient handover with credential checks and dual signature
Immutable audit export for GDP inspection and internal QA

Case study

Aurela Pharma

Pharmaceutical wholesaler · 620 pharmacies across Lithuania and Latvia

Aurela's cold-chain evidence lived in three places: data-logger exports, paper consignment notes and the driver's word. Assembling the record for a single quality investigation took a QA specialist most of a day.

Kelvara now binds logger telemetry to the consignment as it is recorded, so an excursion raises an exception in the depot while the goods are still recoverable. Investigations that took hours are answered from one screen.

100%

consignments with continuous temperature record

6 min

to assemble an audit pack, from ~5 hours

−63%

stock written off after excursions

Case study

Baltnest Logistics

Palletised distribution · Baltic states and north-east Poland

Baltnest's drivers were losing forty minutes a stop waiting for a free dock. Goods-in windows were agreed by phone, recorded in a spreadsheet, and routinely contradicted by what the planner had assumed.

Kelvara made the dock booking part of the plan: the window is confirmed before the route is committed, the receiver gets a reschedule link, and the driver's ETA updates the booking automatically when the day slips.

−38%

waiting time at goods-in

+11%

pallets per vehicle-day

2.1h

average driver time returned per shift

B2B pallet

The constraint is the dock, not the road

Palletised B2B delivery lives or dies on receiving-site constraints: dock availability, tail-lift requirements, forklift access and the hour the warehouse actually accepts goods. Kelvara plans against those constraints as first-class inputs.

Dock-slot booking confirmed before the route is committed
Vehicle-capability matching for tail-lift, pump truck and side loading
Consignment-level proof with pallet counts and damage photos
Cross-border legs with customs documentation attached to the job

Common ground

What every deployment gets

Historical replay

We plan your last quarter before you commit to anything, so the business case comes from your data.

Depot-by-depot rollout

One site runs live alongside the existing process until the numbers hold. Then the next.

Named implementation lead

One person in Kaunas who knows your network, from kickoff through to steady state.

Exportable data

Your operational history stays yours, available as scheduled warehouse exports at any time.

Which of these looks like your operation?

Tell us the sector and roughly how many stops a day you run. We will come back with a realistic view of what changes and what does not.

Talk to us See the platform