How to measure paid social’s impact on PPC

If your paid social campaigns aren’t converting, you may be undervaluing their impact. Your brand’s exposure on social media can influence other parts of your marketing that platform metrics don’t capture.

Here’s how to design and measure a test to understand how paid social influences your other marketing channels, including PPC.

Step 1: Determine your hypothesis

Start with what you want to learn, then define a hypothesis you can realistically evaluate with your data.

For example, this is a common hypothesis for measuring paid search lift from social traffic:

• Search lift hypothesis: Increasing spend on social media will increase brand search volume and overall PPC CTRs.
• Logic: 
  • Social ads build brand awareness. As more people become familiar with our brand, they will search for it more often when making research and purchase decisions. 
  • As more people are exposed to our brand, they will increasingly click on our PPC ads regardless of their search term (i.e., increasing non-brand and brand CTRs).
  • People exposed multiple times to our brand will have a higher trust factor in our products, and therefore, our conversion rates will increase. 
• Measurement: 
  • Impression and click volume for our branded terms.
  • CTR changes for brand and non-brand terms.
  • Conversion rate changes for brand and non-brand terms. 

Your hypothesis could have a different scope, such as measuring paid and organic lift from social spend or an increase in direct traffic. 

Step 2: The test

The next step is to set up the test parameters. Generally, measuring before and after a change is a mistake, as seasonality or other factors can affect your test results.

The most common test setup is a geographic split. In this test, we’ll increase social spend for only a set of geographies. Then we’ll examine the PPC data for the geographies where we ran the test and compare them with areas where we did not.

As you choose geographies, you’ll want to control for other variables that may affect your test. Here are some common issues that companies have run into and need to control for in their tests and measurements:

• You sponsor a sports team, and they’re playing during your test.
  • If the game is regionally televised, this can dramatically affect your test results.
• You’re running TV commercials in only certain regions.
• You choose experimental geographies with many out-of-region commuters, such as New York City, and include New Jersey and Connecticut in your control group.
  • In these instances, grouping a region and its surrounding commuter areas together, and placing other cities with similar characteristics, such as Chicago and…