Many demand forecasts are never checked against the decisions they drive. We work with companies whose decisions depend on forecasts — distributors, manufacturers, spare parts operations, retailers — and start by understanding the existing process and measuring how accurate those forecasts actually are. Then we fix problems identified by the measurements: censored demand hidden by stockouts, error measures that reward the wrong decisions, arbitrary safety stock levels and more.
The problem we solve
Most demand planning runs on tools that quietly cost money. Typical forecasting engines extrapolate the past and break the moment a promotion, price change, or supply disruption enters the picture. But the forecast is rarely the only thing that is wrong. Your sales history might record zeroes because you were out of stock, not because nobody wanted to buy the product. The safety stock levels you define might rely on unreasonable assumptions and lead to stockouts for popular products and overstock for the less popular ones. We streamline the forecasting process, starting with a thorough analysis of the existing operations in your company, followed by data cleaning and processing, and only then moving to forecasting and inventory management.
We work through the whole analytics chain: a thorough analysis of how your operations actually run, then data cleaning and processing, then aligning the evaluation setup with the decisions the forecasts feed, and only then the forecasting and inventory models themselves.
What we do
Consulting
Most engagements start with a diagnostic: we map how forecasting and replenishment actually work in your company, measure how accurate the forecasts are, and report what the gaps are costing you. About a month of analysis and writing the report, from £5,000 depending on SKU count and the state of your data.
What follows depends on what the diagnostic finds — cleaning and restructuring the data, rebuilding the models, resetting safety stock rules, or training your team to do it themselves.
For companies running anywhere from a few hundred to several hundred thousand SKUs, usually with a mix of fast and slow movers.
Practitioner training
Courses in demand forecasting, inventory management, statistics, and analytics — built on a decade of executive training we developed and delivered at Lancaster University. Your planning team learns not just which buttons to press, but why the methods work.
Our next open course, Demand Forecasting Principles, runs online across four weeks in October 2026. We also run bespoke training for teams of six or more.
For planners and analysts who produce or review forecasts.
Why OpenForecast
Research-grade methods, in the open
Our work runs on the open-source forecasting packages we build and maintain — smooth, greybox and others — grounded in a published methodology (ADAM), benchmarked on international forecasting competition data, and used by analysts worldwide. You can inspect exactly what our methods do before you ever hire us.
The right method, not the fashionable one
We use machine learning where it earns its place, and statistical models where they do. Neural networks and gradient boosting beat conventional approaches on some problems and lose badly on others, particularly with short histories or sparse demand. Knowing which case you are in is the expertise.
Explainable models
You can see the logic, explain the numbers to finance, and stand behind the plan. Where a less interpretable model is genuinely better, we say so and explain what you gain and give up.
Academic rigour, practitioner focus
OpenForecast is led by Ivan Svetunkov and Nikolaos Kourentzes — forecasting researchers with decades of published work between them and years of experience making forecasting work in practice, not in slides. We wrote the methods, built the software, and teach the courses. More about us →
The tutors are true experts in statistical forecasting models. Their insights on the simplest aspects of the field were valuable, even for a seasoned practitioner like myself.
Everything we know, published
Our methods are not proprietary secrets. Two books, a decade of articles, and full documentation are freely available — you can read exactly how we work before you ever hire us. Start with the resources, explore the packages, or read the blog.
Recent from the blog
ISF2026: PTS Taxonomy of Multiple Source of Error State Space Models for Demand Forecasting
This time, at ISF2026, I presented the paper that I have worked on together with Juan Ramon Trapero and Diego Pedregal. The idea of the paper is to introduce a taxonomy of the models in the Multiple Sources of Error (MSOE) framework. In the Single Source of Errors one, there is ETS, in the MSOE, … Read more
stick function for the EDA in time series
You have probably seen my post about the STI classification of Hans Levenbach (this one). Well, I’ve decided to implement it, and it has landed in the greybox package for R/Python. What’s greybox? It is a package for statistical modelling focusing on forecasting and time series analysis. I created it back in 2018 to split … Read more
smooth in python: Non-normal distributions in ETS/ARIMA
So, you know quite well that the normal distribution is one of the most popular distributions in statistics. The reasons are manifold, including convenience for the academic community and the fact that it is taught in every single statistics course in the world. But what if we don’t want to be normal? There are situations … Read more
If your forecasts drive real inventory decisions, we should talk.
Book a free 30-minute intro call — no preparation needed. We will ask about your situation and tell you honestly whether and how we can help.
Prefer to write? Send us a message