Predicting conversion intent part 4 - Scoring, deploying and monitoring the model

• Blog 1: capture the behavioural ‘rules’ that lead to your target
• Blog 2: feature selection analysis
• Blog 3: training/evaluating the model
• Blog 4: deploying the model

In my third blog about predicting conversion intent, I explained how to train a Conversion Intent model in Google Cloud Platform. This model can be used to score visitors on what their future Conversion Intent will be, this result can be used as an Audience in Google Marketing Platform. How to do this is, is explained in this blog. Also, since the world is ever-changing and performance must be secured, I will explain training and monitoring in Google Cloud as well. 

Scoring and using the model in Google Marketing Platform

A trained model is nice, but it has no actual business impact if we don’t put it into action. We want to use the model to predict Conversion Intent of visitors of the website. And we want to use these predictions as an Audience in for example Google Ads for retargeting purposes.

We created this Conversion Intent model for a large international fashion retailer. Together, we decided we want to score visitors near-real time so that we could quickly retarget them with the right advertisement. This means that we score visitors every 15 minutes using the last trained model. Scoring means that we put all the visitor features into the model, and get a prediction as a result. This can easily be done with the BigQuery function ML.PREDICT, where you input a table of features (e.g. visitors of the last 15 minutes) and a model (e.g. the last trained model) and you get a prediction as a result. In this case, it is whether someone is likely to convert or not.

These predictions combined with a clientID, are pushed to Google Analytics. To be able to do this we have created a Cloud Function that consists of a Python script that grabs the visitor scores and uses the Measurement Protocol to send these as events to Google Analytics. Then in Google Analytics marketers can create audiences and put these audiences to action in the Google Marketing Platform.

Retraining & monitoring the performance of the model 

The world is ever changing and so are our models. In the blogs before I mentioned how to create a prediction model that you can use to predict the conversion intent of visitors based on their previous behaviour. To make sure our models are still relevant and useful we want to retrain our models from time to time. For the large fashion retailer mentioned before we have created a model training pipeline in Google Cloud Platform. By using Cloud Functions we retrain our predictive models once a week. An automated evaluation takes place, and when good enough, the model overwrites the last model, and is being used to score new visitors. In this way, we make sure that the model always looks at the most recent behavioural patterns in predicting conversion intent.

Part of the model training pipeline we monitor the actual performance of our models. As mentioned before, the trained model is used to score visitors on conversion intent. In the two weeks after, we can actually see whether they converted or not. By comparing the scores of the visitors and their actual behaviour, we use evaluation metrics like accuracy, precision, and recall to keep track of the model. In that way, we can see whether retraining the model weekly is enough, or that we should go through the whole process again of feature creation, feature selection analysis, and training based on different models or parameters.

Conclusion

In the last 4 blogs, I have shared how to predict which customers are likely to convert, and how to set this up in a successful way in Google Cloud Platform and use it in the Google Marketing Platform. This is only one example of the impressive amount of opportunities within the Google Cloud Platform. If you are interested in learning more, don’t hesitate to give us a call.

Read the other blogs

ABOUT CRYSTALLOIDS

Crystalloids helps companies improve their customer experiences and build marketing technology. Founded in 2006 in the Netherlands, Crystalloids builds crystal-clear solutions that turn customer data into information and knowledge into wisdom. As a leading Google Cloud Partner, Crystalloids combines experience in software development, data science, and marketing, making them one of a kind IT company. Using the Agile approach Crystalloids ensures that use cases show immediate value to their clients and frees their time to focus on decision making and less on programming.