Analysis & Takeaways of Campaign Performance

Through a variety of new channels, including Google Ads, social media networks, and directed email campaigns, we drove a sizable increase to WELL AP campaign pages, creating more awareness and converting more than 300 registrations
  • 127%

    User Increase (+4,438 New)

  • 39.5%

    Bounce Rate (Site avg 53%)

  • 1,152

    Registrations (304 Completions - 5% Conv Rate)

  • The decreased bounce rate reflects the fact that we honed in on a targeted audience. More visitors landed on the site and stayed.
  • However, the average number of pages per visit decreased slightly because the WELL AP campaign landing page served users more information than the normal page, meaning users did not have to visit multiple pages to get the desired information.
  • The number of conversions reflect last-click attribution, meaning if a user clicked an ad, left the site, and returned the next day via direct visit, the direct/none channel would get the conversion. In other words, these registration completions do not reflect assisted conversions.

Assisted Conversion: An instance where a particular channel plays a role in the conversion process, but was not the last click that drove it.
For example, below, a user visited the site by clicking on a social ad, but did not immediately convert. At some point in the future that same user entered the site directly, and then converted.

Social media played an assisted conversion role because it was not the last click that lead to the purchase, yet was the first entry point through which the user discovered the website. But because of the way registrations are tracked, the “Direct/None” channel is attributed that conversion.


  • Google Ads targeted mostly brand keywords, displaying the discount pricing ads to users searching for any term associated with "WELL AP." We are going to begin focusing on adding a non-brand approach next month.
  • The large number of impressions during the first days of the campaign are related to broad targeting that didn't convert to clicks. Over time, I modified the audience and click-through-rate increased dramatically as impressions leveled out. This means the ads were resonating with the correct audiences, resulting in the highest CTR (38%), and utilization of budget in the most economical way ($3.32 per click).
  • We had a few issues with the ads that likely hindered performance. For example, in the first week, Google limited the presence of our ads in the US because it algorithmically tagged some ads as copyright violations by using the word "AP".

  • Organic was a high performing channel, resulting in over 100 registrations. Many of the conversions has assistance from paid channels or non-digital advertising (like signage or paid ads). For example, a potential customer from Facebook learned about the discount via FB ad. She clicks on the post, and reads about the discount. The next day, she searches Google to find the registration page, and purchases the test. Organic / Google would get credit for that conversion, not Facebook.
  • I am currently building out different Attribution Models to give first-click social the credit for their conversions.
  • At the beginning of the campaign, I optimized all the WELL AP title tags and meta descriptions to increase visibility. As a result, the CTR for organic increased 37%. Optimizing titles and descriptions are one of the low hanging fruit tasks we are tackling in 2020.

  • LinkedIn had the most engaged users, with avg time on site over 6 minutes, and bounce rate of just 56%. Since there is a smaller pool of users, however, we need to switch up ads and run them weekly. We don't want high ad saturation with users.

  • Twitter appears to be the least successful social channel for selling WELL AP, or sales in general. Based on performance, Twitter would be better in building awareness with articles or resources to build out and fill the top of the WELL AP funnel.

  • Facebook performed best out of the channels, but the most budget was used here. We suggest using different ad types to address different sections of the audience, as well as more targeted campaigns with modified landing pages. Facebook also had an issue with too much text on the ads, resulting in diminished coverage.