Regression analysis to improve Google Ads performance
Advanced digital marketing requires us to go beyond what everyone else is doing and approach from new angles. One of the ways to stand out in your SEM analysis and performance is through advanced techniques like regression analysis. Regression is actually a form of basic machine learning (ML) and a relatively simple mathematical application. This type of analysis can help you make better predictions from your data, beyond educated guessing.
Regression might sound scary, but it’s not that advanced in the world of mathematics. For anyone who’s passed year 10 maths, you have probably already worked with regression formula previously. We’re going to look at using regression in your Google Ads to predict the conversion volume you can achieve by adjusting campaign spends. Building the model and applying it is far easier than you would think!
What is regression?
A regression model is an algorithm that tries to fit itself to the presented data best. In essence, it is a line of best fit. It can be linear, as a straight line through the data, or non-linear, like an exponential curve, which curves upwards. By fitting a curve to the data, you can then make predictions to explain the relationship between one dependent variable and one or more independent variables.
The plot below shows a simple linear regression between an independent variable “cost” (daily spend on Google Ads) on the x-axis and a dependent variable “conversions” (daily conversion volume on google ads) on the y-axis. We have fit a linear regression line (blue). We can now say that at $3k on the axis, that point on the regression line would match up to 35 conversions. So, based on the regression model fitted to the data, if we spend $3k, we are predicted to receive 35 conversions.
Headstart on feature selection
I’ve been running many of these regression models and I’ll share what I’ve found to be true, which will give you a headstart in where to start looking
Multiple regression is where some independent variables are used (rather than just one, as in the example above), to predict one dependent variable. With Google Ads, I’ve found that there is always one independent variable that is the strongest predictor of conversions. You could probably have guessed which one it is already.
When running ML model’s on daily labeled training data to predict whether certain features would lead to a conversion, we continually found that all other things being equal, campaign spend is the strongest predictor of conversion volume.
The following table shows the “Root Mean Squared Error” (RMSE) for different ML models.
RMSE is a measure of error, it shows how far off the fitted model is from the training data. The lower the error the better – it means the model is more accurately fitted to the data. (2) All features include: Day of week, keyword, CTR, CPC, Device, final URL (landing page), ad position & Cost.
We ran five different machine learning algorithms: Decision Tree, K Nearest Neighbours, Linear Regression, Random Forest and Support Vector …read more
Read more here:: marketingland.com-sem