“Our vision statement of contemporary marketing mix modelling is machine learning supported, granular, automated, and experiment-calibrated analysis, delivering insights faster and on a continuous basis. Robyn is our step towards this vision,” we heard from Igor Skokan, Marketing Science Director at Facebook.
The participants could see our approach to working on Robyn in practice from the example of a project with a commercial client. It was also an opportunity to exchange experiences with other specialists from around the world on this urgent topic.
Innovation Lab showed how Robyn is being used in practice and how we are approaching marketing mix modelling (MMM). If you already know that this is something for you and you could not be at the event, sign up for the demo.
We presented Predicty, a tool we are developing based on Robyn. Predicty is a solution for allocating budgets and having an impact with eCommerce components on overall sales. Thanks to the commercial collaboration we started a few months ago, we were able to discuss it not only in theory but also from a practical point of view.
What were we talking about? Below is a summary of the speeches, and if you want to see those speeches for yourself, the full recording is still available at this link.
Agata Czapla, Growth Manager for Innovation Lab, introduced the topic and our approach to Robyn. The Facebook team created Robyn to democratize and enable data analytics despite the upcoming changes. We view Robyn as our back end and built on it to develop a tool that can be easily used by non-technical people. On top of that, based on our experience in eCommerce, we’re successively adding regressors that are components of eShops because we know how much they affect revenue.
Artur Wala, Head of Innovation Lab, talked about the process, applications, and the tool. He also talked about the individual components and functions of the solution, such as budget and channel allocation. Based on his experience working with digital projects, he emphasized how important it is for us that the tool is comprehensive and accessible to data specialists and others who need that information. So, by simply moving the sliders in Predicty, users can immediately see which campaign is worth investing in, which campaign should be closed, and which channel needs attention.
The presentation was followed by a Q&A session where Kamil Janik, the Technical Lead, shared his knowledge and experience in the field. The participants were interested in the use of individual regressors or the possibility of using Predicty in different industries. It became clear how important and urgent this topic is, but also that our tool can be useful for many companies in their daily budget and advertising decisions.
“This fresh perspective on MMM with eCommerce specific from Divante is very interesting and inspiring” Gabriel Matwiejczyk, Marketing Science Parter in Facebook, said.
After the summit, we read encouraging words and thoughts on social media like this Twitter comment (thanks!). This only spurs us on and gives us extra motivation. We plan to start the next revolution along the lines of the GPT-3, and the thought that we’re the first to develop this type of tool only added extra wind in our sails.
For those who are interested in the topic and could not attend the summit, the whole event was recorded so you can watch it at any time.
If Predicty sounds like a solution that would work for your company, let's talk. You can also sign up for the newsletter which will provide you with some knowledge about budget allocation and MMM. See you at the next summit!
Published October 19, 2021