Platform

How does delivERP work?

From data to decisions and back. The platform collects data on orders and resources, forecasts, makes dispatch decisions and passes them to drivers and pickup points, and the results come back as data.

Our team: analyses, reports and improvements

  1. Data

    • orders with requirements: time windows, weight, volume, vehicle type
    • drivers and vehicles: availability for planning, actual presence during the day
    • every update and GPS positions from the fleet
  2. Forecasts

    • demand in every hour
    • travel times from our own navigation
    • service time at the address and order readiness
  3. Decisions

    • how many drivers and vehicles for each day and hour
    • who takes which order and in what sequence
    • re-planning after every change
  4. Execution

    • driver app: yours or ours
    • tablet at the pickup point
    • your systems via integrations
  5. Oversight and results

    • live view of decisions and manual control
    • exceptions that need a person
    • reports and data for settlement

Actual trips, stops and results come back as data, so the system learns on your fleet.

Without a team’s work, even a good tool delivers only part of its potential. That is why delivERP is part of a service rather than a licence: configuration, rollout and continuous fine-tuning to your fleet are our job (how we differ from a software vendor).

Know-how

Why can our know-how not be copied quickly?

Because the result does not come from one algorithm but from hundreds of modules and improvements refined in real operations since 2014. The knowledge behind them is not in any source that people or AI tools could learn it from.

delivERP consists of hundreds of cooperating modules: analyses and forecasts, dispatch decisions and interfaces for dispatchers, drivers and pickup points, because logistics has, and will long have, many touchpoints with people. The algorithms themselves are mostly known from the literature. What determines the result is what happens around them:

  • parameterisation tailored to a specific fleet and location,
  • feeding them the right data and separating valuable data from noise and interference,
  • understanding how the data depends on people’s behaviour and how people’s behaviour depends on the system’s decisions.

This knowledge is not in textbooks, documentation, publications or on the internet. It is practical knowledge gained over over 12 years of running our own fleet and working with clients’ fleets: with senders, warehouses, recipients, couriers and vehicles. Much of it has never been written down, because it comes from thousands of everyday situations.

AI tools make it faster to write code today. But they do not know what has never been written down, they do not see what a person notices in operations, and they do not shorten the time it takes to test a solution in a fleet’s day-to-day work.

For years we have invested above all in the fundamentals rather than in what you see on screen. Some screens look much as they did years ago, while underneath, in the decisions, forecasts and data, almost everything has changed. Many systems grow the other way round: under pressure to grow fast they gain features and impressive screens, while weak fundamentals stay in place, often unnoticed, because none of the key decision-makers works at that level day to day.

Refining a single detail still often takes us many months, and there are thousands of such details. Not because coming up with a solution takes that long. We test each one in real operations straight away, and there it meets the people who are everywhere in logistics: the sender, the recipient, the courier, the coordinator who watches the system at work and influences its decisions, and those who make business decisions. Only then does it become clear what does not work or does not deliver the expected results, and fine-tuning begins. That time cannot be sped up.

Example

How does the system know how long a trip will take?

From our own navigation system, which learns from real journeys, above all on the routes your fleet drives.

We use public OpenStreetMap maps. We developed the remaining components ourselves: the engine that calculates and evaluates routes, and the traffic model, which accounts for traffic levels by time of day, day of the week and other factors, as well as the current situation on the road. The model learns from GPS data from many sources, above all from the fleet covered by the deployment.

The system precisely determines when a drive ends and a stop begins, and when the stop ends. This allows it to measure travel time and service time at the address separately, and dispatch needs both.

It finds the optimal route not only in terms of kilometres and time but also road tolls. In the last mile it finds shortcuts that standard systems miss, such as a route through a car park or a gate instead of two junctions with traffic lights. This can save up to 10 minutes on a single trip.

Why do we not use off-the-shelf navigation systems?

Because their limitations stem from the very design of their algorithms. Off-the-shelf mapping and routing systems are designed for navigation, not for logistics optimisation: they do not expose some of the details we need, or they calculate them in a way that does not meet the needs of dispatch. The dependencies between navigation and dispatch become visible only when both are developed together. We have done so for years.

Integrations

Which systems does delivERP integrate with?

With the systems you already use. We do not replace them: we adapt to your current processes and change them only where this is necessary to achieve savings.

We integrate with:

  • transport management systems (TMS),
  • ordering systems and online shops,
  • point-of-sale systems (POS),
  • order handover screens at pickup points (e.g. KDS),
  • multi-channel order aggregators,
  • ERP and warehouse management systems (WMS), e.g. for instructions to move consignments between staging areas,
  • reporting and BI tools.

We already have ready-made integrations with many popular systems and develop new ones ourselves. Elements that your systems lack and that can be added from outside (such as the driver app, the dispatch panel or reports) come from our ready-made solutions. If we do not have one and see savings potential in it, we develop it.

When does a system have to be replaced?

Only if it cannot be integrated or does not offer the flexibility needed for the changes that deliver the savings. In that case replacing it is the only option.

Data

Where do we process data?

Mainly on our own servers in the European Union. delivERP runs as a service, with nothing to install at the client, and we sign a data processing agreement with every client.

About 90% of our infrastructure runs on servers we manage ourselves; we use public cloud services only selectively. This gives us full control over where and how data is processed. Our own infrastructure also makes integration with clients’ systems easier, lowers costs and improves the platform’s performance.

Let’s check whether your fleet has potential

Start with a conversation and historical data on your fleet and orders. The analysis is free. If we do not see potential, we will tell you straight away.