We are a demand forecasting and inventory consultancy and training provider, built on more than two decades of forecasting research, teaching, and open-source development. Everything we know is published in the open — two books, seven packages, and a decade of articles. What we sell is the skill of applying it.
What OpenForecast is
Three things, tightly connected. A consultancy that builds forecasting and inventory models around how businesses actually operate — with particular strength in the hard cases, like intermittent demand in spare parts and retail. A training provider that teaches practitioners how to forecast and manage inventory properly, whether in their ERP system, Excel, R, or Python. And the home of a family of open-source forecasting packages — smooth, greybox and others — grounded in peer-reviewed research and used by analysts worldwide.
The packages are the engine, the consulting and training are the application, and this site is where all of it lives in the open.
How we got here
In the 1990, Robert Fildes founded the Centre for Marketing Analytics and Forecasting at Lancaster University with the specific idea in mind: forecasting research should not be detached from practice and should inform real business decisions. Talk to companies, understand what they actually face, and let that shape the work. That philosophy is what made CMAF distinctive, and it is the direct ancestor of how OpenForecast operates today.
Nikos worked at the centre from 2009 to 2019, running engagements and consultancy alongside his research. Ivan joined in 2013 and started writing about forecasting shortly after — the blog that eventually became this site. The software followed in 2016, built because the methods needed implementations before anyone could use them. Then the projects: work with companies, and code development for software vendors. ADAM was written in 2020, summarising a decade of research into exponential smoothing conducted with businesses rather than about them.
In 2026, Ivan left academia to do this work full time. OpenForecast is the same approach, without the university around it: methods developed in the open, tested against real problems, and applied by the people who built them.
Who we are
Ivan Svetunkov
Ivan is the founder and the director of OpenForecast and the creator of the smooth and greybox packages. He holds a PhD in Management Science from Lancaster University, where he spent more than a decade researching and teaching forecasting and statistics, latterly as Senior Lecturer and Director of the Centre for Marketing Analytics and Forecasting. He is the author of the ADAM framework and the monograph behind it, and his research spans state-space models, intermittent demand, and supply chain forecasting.
Nikolaos Kourentzes
Nikos is one of the most cited researchers in business forecasting, formerly Professor in predictive analytics at the University of Skövde and at Lancaster University, and now he is the head of research at Indicio Technologies, a Swedish forecasting software company, and an external forecasting expert for the International Monetary Fund. He actively researches in the areas of model selection and uncertainty, forecast combination, hierarchical forecasting, and the behavioural side of the forecasting process. He is a co-author of the Principles of Business Forecasting textbook, and applies these methods across supply chains, tourism, and healthcare.
Advisory Board
Robert Fildes
Robert is Emeritus Professor of Management Science at Lancaster University and founder of the Centre for Marketing Analytics and Forecasting. He is a founding editor of the International Journal of Forecasting, a co-founder and and a past president of the International Institute of Forecasters, and co-author of Principles of Business Forecasting. His research on forecasting practice in organisations — how forecasts are actually produced, adjusted, and used — shaped how the field understands the gap between method and application. He advises OpenForecast.
How we work
The same principle runs through the research, the teaching, and the consulting: show your working. The methods we use are published, the software is open source, and the models we build are ones your team can understand, explain, and run without us. We would rather demonstrate exactly how something works than ask you to take it on trust — which is why the books and the packages are free, and always will be.
It follows that we will tell you when we are not the right answer. If the diagnosis is that your problem is a process problem rather than a modelling one, or that your team needs training rather than a consultant, we will say so.
We are based in the UK and work with clients internationally. Consulting is conducted remotely, with on-site work where it helps; training runs online, or at your premises anywhere we can reasonably travel.
Where to find us. Company news, course announcements and new posts go out on LinkedIn. Ivan posts about forecasting, statistics and the packages on Bluesky.
Want to work with us?
Read about our consulting and training, explore the packages, or get in touch.
Prefer to write? Send us a message