Time Series
Time Series models have enormous application areas all over business.
Time series models are capable of providing decision support for a business by constructing a model that accounts for as many drivers as possible as well as the inherent business dynamics of a specific business. This means that we can with one model provide insights and recommendations regarding
- Sales Predictions
- Customer Churn
- Customer Acquisition Cost
- Impact of competitors
- Budget planning
- Pricing Dynamics
- Impact of branding
Methodology
How to make time series models robust and actionable
These types of models have a series of challenges; they need to have business sanity, need to operate in a low data environment where the number of data points may be less than the number of input variables, be able to handle nonlinear dynamics, be transparent in the model design i.e. not a “black box”.
Standard tools in data science and machine learning are unsuitable for this, as they tend to be data-hungry, prone to overfitting, and unable to express business dynamics. So to overcome this we use Bayesian inference methodology, which allows us to express business knowledge as priors and thus reduces the need for data to train the model without overfitting.
Some of the key benefits of this approach are:
Standard tools in data science and machine learning are unsuitable for this, as they tend to be data-hungry, prone to overfitting, and unable to express business dynamics. So to overcome this we use Bayesian inference methodology, which allows us to express business knowledge as priors and thus reduces the need for data to train the model without overfitting.
Some of the key benefits of this approach are:
- Accounts for business dynamics
- Automatically estimates proper uncertainty
- Robust decision support
- Faster generalization
- Works well with high dimensional data with few data points
- Pool information across variables in a hierarchical fashion, reducing further the need for data
- Avoids overfitting
(Normal data science on top, what we do on the bottom)
Insights
How our models can provide insights in to your business
When we start working with a client we do a scoping session to determine what the goal of the model is and what data is available to build the model with. So insights the model can provide will be determined based on that, the obvious thing is if we don't have a set of data ex. media we can not provide any insights into media. Sometimes the data is limited to just sales and price, then the model usually focuses on sales forecasting and the impact of external factors such as macroeconomics.
But that being said, assuming we have the relevant data available to use in the model. Here is an example of the insights we can provide
But that being said, assuming we have the relevant data available to use in the model. Here is an example of the insights we can provide
- What factors drive our sales?
- What was the ROI of our latest campaign?
- How much revenue do I get per 1000 EUR invested in media?
- How many sales are we getting out of our SEO program?
- What is the synergy between offline and online media?
- How do price and media interact?
- How does marketing affect each distribution channel's ROI?
- How does distribution affect our brand?
- How much should I invest in Media next year?
- What is the optimal timing for our next campaign?
- With a 1 million EUR media budget, which media groups should invest in?
- How many sales would we lose if we cut media by X%?
- How do we respond most effectively to changes in external factors (competitor initiatives, business cycles, pandemics)?
Actions
How our models can provide actionable steps
When we start working with a client we do a scoping session to determine what the goal of the model is and what data is available to build the model with. So insights the model can provide will be determined based on that, the obvious thing is if we don't have a set of data ex. media we can not provide any insights into media. Sometimes the data is limited to just sales and price, then the model usually focuses on sales forecasting and the impact of external factors such as macroeconomics.
But that being said, assuming we have the relevant data available to use in the model. Here is an example of the insights we can provide
But that being said, assuming we have the relevant data available to use in the model. Here is an example of the insights we can provide
- What factors drive our sales?
- What was the ROI of our latest campaign?
- How much revenue do I get per 1000 EUR invested in media?
- How many sales are we getting out of our SEO program?
- What is the synergy between offline and online media?
- How do price and media interact?
- How does marketing affect each distribution channel's ROI?
- How does distribution affect our brand?
- How much should I invest in Media next year?
- What is the optimal timing for our next campaign?
- With a 1 million EUR media budget, which media groups should invest in?
- How many sales would we lose if we cut media by X%?
- How do we respond most effectively to changes in external factors (competitor initiatives, business cycles, pandemics)?