Predictive modeling for domain experts is a huge step forward towards predictive marketing.
How much money can a company earn from a single ad? That’s the first question every company considers before launching a web, TV or print marketing campaign.
The 2017 Salesforce report on Marketing showed that 60% of French marketing experts are now using AI tools for product recommendation or market maturity prediction.
Tools and techniques to anticipate customer behavior thanks to the collected data, also known as predictive marketing, are now key to help marketing experts make operational decisions.
Predictive models allow to quickly identify what makes a good advertisement.
Problems to solve
- What are the triggers to a successful ad?
- How to improve your return on investment?
- When is the best time to launch an ad to make it successful?
- How can AI help to understand if the customer wants to receive an ad?
Benefits of TADA
Marketing and communication specialists could benefit from predictive models to optimize their campaigns and improve their Return on Investment. However, they are not Data Scientists and are not skilled enough in Machine Learning and coding to build predictive models.
Marketing specialists have access to data from their previous campaigns. This data is considered Small Data, as it is usually built around a limited quantity of campaigns (a few thousands). Moreover, Small Data doesn’t work well with traditional Data Science tools and techniques.
In this context, TADA allows marketing specialists to build predictive models on user behavior in upcoming campaigns so that they can publish the best message possible at the best time.
No data science training is required to use TADA. Marketing and communication specialists can use their own data without preprocessing or normalization.
Every year, companies spend great amounts of money on TV, web and print advertising. Predicting an advertisement’s return on investment becomes harder as the number of variables increases, especially as some are unknown.
By using TADA, marketing and communication specialists can identify and optimize these variables to improve their campaigns for a product or a full range of products. By doing so, they can maximize their return on investment on the whole production range and find the best scenarios.
Such an approach is a precious help to decision-making, especially when marketing specialists must establish priorities between various products and campaigns.
Automated Machine Learning tools help users to predict the future thanks to historical data. To predict a future result, you must compile your descriptive data and the past results obtained.
TADA allows you to easily create a relevant predictive model from your data and apply it to future data.
In this use case, descriptive data are descriptive information from previous marketing campaigns as well as their sales results. TADA aims to predict the sales of a product after an advertising campaign.
You can generate a model in just 4 steps:
- Step 1: create your project and upload your data as a CSV file (with data in rows and variables in column).
- Step 2: Select the variable you want to predict, called “Goal”. In this use case, the goal is the “sales” variable.
- Step 3: Select your data for the model generation. This step is called "Creating the Variable set" and allows you to manually select the descriptive variables you want to use. By default, they are all selected.
TADA identifies the relevant descriptive variables by itself which affects the calculation time required to create the model.
The fewer variables selected the faster the model creation.
- Step 4: Create your model. When creating your model, some default values are proposed for the name of the model, the size of the population and the number of iterations.
You can start model generation by validating the default values or editing them according to your preferences. You’ll find best practices at your disposal to guide you in the choice of these parameters in the TADA UI.
According to the size of the file, this step can take between a few seconds and ten minutes. Once the model created, you have access to metrics and graphs to evaluate its relevance.
How can we go further?
You have various options to put your model into practice:
- Use the « Predict » feature of TADA: upload a CSV file with the data to predict. In return, TADA will generate a CSV file with the calculated predictions.
- Retrieve the associated mathematical formula and apply it (for instance on Excel).
- Retrieve the source code of the mathematical formula and use it on your own apps. The source code is available in R, Java, C++ and Python soon. (This option is only available in TADA Premium and Pro).
The below screenshot shows an extract of the dataset. Each row equals to a whole day of advertising campaigns and each column equals to a variable.
The variables of the dataset are the following:
- Date Local: Date of the advertising campaign
- TV Sponsorships: TV sponsorship expanses
- TV Cricket: Ad expanses during cricket games
- TV Ron: Paying channels expanses
- Radio: Radio expanses
- NPP: Newspaper advertising expanses
- Magazines: Magazine advertising expanses
- OOH: Out of Home advertising expanses
- Social: Social network advertising expanses
- Programmatic: Programming websites advertising expanses
- Display Rest: Internet advertising expanses
- Search: Search Engine advertising expanses
- Year: Year of expanses
- Day: Day of expanses
Model type: Regression
Column number: 15
Row number: 180
The results show how the predictive model performs.
The predictive model type and its metrics are linked to the Goal and its values. The model type is shown on the model results display.
Three types of prediction can be done according to the Goal data (here, the Goal is “sales”):
Binary classification: a discrete value taking only two values, such as Yes/No.
Multiclass classification: a discrete value with more than two values, such as status of state with values like “On”, “At Risk”, “Down”, etc.
- Regression: a continuous value that can take an infinite number of values, such as a temperature, a pressure, a turnover or the price of a house.
When generating the model and according to the state of the art of Machine Learning, your data will be divided in three parts by TADA:
Part 1: A Training part which represents 40% of the data and is used to train a certain number of models,
Part 2: A Validation part which represents 30% of the data and is used to validate and select the best models found in the previous step,
- Part 3: A Test part which represents 30% of the data and is used to test the model approved during the validation step.
The performance measurement and the model evaluation must be done on the Test part (according to Machine Learning standards) as the data used during this phase was not used to build the model and is just used to measure its performance.
Here, the prediction is a regression. For every row to predict, TADA will compute a numerical prediction. The difference between the predicted value and the actual value is called “error”, which can be positive or negative.
Errors are used to compute the below metrics which help us to assess the model quality.
The “sales” variable has an average value of 1.49 and a standard deviation of 0.2. TADA predicts this goal with a 3% variation (MAPE is 0.03).
Furthermore, the TADA model has an average error of 0.06 on “Sales” prediction.
Finally, the model is very precise to predict the “sales variation”, as shown by the R2 coefficient of 0.89.
These three points show than TADA is precise on the goal prediction and performing well.
MAPE (Mean Absolute Percentage Error) is the percentage of average error for each prediction. In this example, MAPE is 3.2%.
MAE (Mean Absolute Error) is the average of the absolute value of errors. In this example, MAE is ±0.05.
RMSE (Root Mean Square Error) is the average of the sum of squared errors. This value is more sensitive to outliers. In this example, RMSE is ±0.06.
R2 determines if the model explains the goal variations in comparison with a prediction of the average value (R²random=0). Here, R2 is 0.89 (R²max=1).
Ready to use TADA?
You don't have immediate data?
No problem, data are available to make your trial as relevant as possible!Try it now!