We conducted an Ordinary Least Squares (OLS) regression analysis to investigate the relationship between CPM (cost per thousand impressions) and CTR (click-through rate) for a sample of 72 Meta ad campaigns. The results showed that the model has a high R-squared value of 0.694, which indicates that approximately 69% of the variation in CTR can be explained by the variation in CPM. Additionally, the F-statistic of 159.0 and p-value of 1.10e-19 suggest that the model is statistically significant and that the regression coefficient for CTR is significantly different from zero.
The estimated regression coefficient for CTR was 2.776e+04, which means that for every one unit increase in CTR, we would expect to see an increase of 2.776e+04 in CPM. Furthermore, the 95% confidence interval for the coefficient ranged from 2.34e+04 to 3.21e+04, indicating that we can be fairly confident that the true population coefficient falls within this range.
The regression analysis also provided some diagnostic statistics, such as the Durbin-Watson statistic of 0.681 and the Jarque-Bera test for normality, which suggested that the residuals of the model were not autocorrelated and followed a normal distribution.
Overall, these results indicate that CPM has a significant impact on CTR for Meta ads, and that evaluating ad performance based solely on CTR may not provide a complete picture of the effectiveness of an ad campaign. By considering the cost per impression, marketers can gain a better understanding of the value of their ad campaigns and make more informed decisions about their advertising strategies.
import pandas as pd
import statsmodels.api as sm
# Load the data into a pandas DataFrame
data = pd.read_csv("Test Campaign Original Data 1c22fe19d2ee400c8c4e76f8ea39d0f9.csv")
# Fit a linear regression model to the data
model = sm.OLS(data["CPM (KRW)"], sm.add_constant(data["CTR"])).fit()
# Print the summary statistics of the model
OLS Regression Results
====================================================================================
Dep. Variable: CPM(1,000회 노출당 비용) (KRW) R-squared: 0.694
Model: OLS Adj. R-squared: 0.690
Method: Least Squares F-statistic: 159.0
Date: Fri, 24 Feb 2023 Prob (F-statistic): 1.10e-19
Time: 22:15:46 Log-Likelihood: -681.87
No. Observations: 72 AIC: 1368.
Df Residuals: 70 BIC: 1372.
Df Model: 1
Covariance Type: nonrobust
===============================================================================
coef std err t P>|t| [0.025 0.975]
-------------------------------------------------------------------------------
const 443.2147 619.593 0.715 0.477 -792.524 1678.953
CTR(링크 클릭률) 2.776e+04 2201.553 12.609 0.000 2.34e+04 3.21e+04
==============================================================================
Omnibus: 7.934 Durbin-Watson: 0.681
Prob(Omnibus): 0.019 Jarque-Bera (JB): 8.714
Skew: 0.527 Prob(JB): 0.0128
Kurtosis: 4.340 Cond. No. 6.17
==============================================================================
Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
Python
복사
