FAQ
How many KPIs can you model?¶
Depending on the needs of the customer, we either do a big model with multiple KPIs if they are exposed to the same drivers. In that case we normally scale up to 100+ KPIs within one model.
If the KPIs can be considered independent or can be grouped, we easily scale to thousands of models in parallel, make it ideal for situations like demand forecasting.
How many models can you handle?¶
We have spent a lot of effort perfecting our Bayesian Inference engine such it can utilize a methodology called Variational Inference. This makes is possible for us to scale to thosands of models in parallel.
How much data do I need?¶
The amount of data needed depends on what the model should be used for. But since we utilize Bayesian inference we can work with as little as a month’s worth of data and at the same time can scale up into decades worth of data. Of course more data will improve the quality and precision of the models produced.
How many input variables can you handle?¶
Our Artificial Intelligence engine easily scales up to thousands of input variables. At the same time, it enables built-in model sanity and transparency from a business perspective. This enables explainable and robust models which your stakeholders can trust.
What data do I need?¶
The data needed is of course dependent on what you want to model. We support a wide range of data sources such as
- Sales
- Marketing
- Branding
- Visits
- Distribution
- Macroeconomic factors
- Events
But in order to get started we only need the data on the commercial indicators you are interested in, then any additional data you have will improve the insights that can be extracted from the model.