How to use Google Analytics audience data to make media and marketing decisions

Andrew Ruegger on
  • Categories: Analytics, Analytics & Marketing Column, Channel: Analytics & Conversion, Google, Google: Analytics
  • Market-leading site analytics providers such as Google and Adobe continue to bring new insights about consumer behaviors through advancements in their analytics and tag management offerings.

    Both Adobe (Marketing Cloud) and Google (360 Analytics) have made big strides in improving their offerings to include tag management, measurement, attribution, optimization and more. These are designed to equip brands with the tools needed to maximize the value of their media investments and web traffic data.

    One free and easy place to look for new insights into your consumers is in Google Audience reports, which can be freely enabled by implementing Google’s Universal Analytics (or RLSA tagging). Here Google is simply passing on additional data from the DoubleClick cookie, the Android Advertising ID and the iOS Identifier for Advertising (IDFA) to paint a clear picture of your website’s audience.

    For me, the most interesting areas to explore are the Interests categories, both In-Market Segments and Affinity categories, and the differences between each medium. Because programmatic display and paid social are buying audiences, while direct and organic traffic are earning them, comparing the two and using what you learn as a feedback loop to adjust strategies is often effective and insightful.

    These are the metrics and dimensions used for the following examples:

    We then switch out In-Market for Interest Affinity for charts further down. The reason we sometimes don’t use both at once is that Google does not have as much of the data paired, meaning there will be a lot less data returned if you want to see users for both dimensions.

    When you’re dealing with a lot of data (this example has 2.5 million rows), your only choices are:

    1. Visualize it and explore possibilities.
    2. Run scripts/algorithms to determine a specific output.

    Normally, it’s best to visually explore first to find helpful patterns or insights, and then create the scripted work to process the data afterward. Visualizations are also typically better for blog posts:

    Because our objective is to drive conversions on this website, everything is color-coded by conversion (green being high average conversions, red being low average conversions).

    Additionally, we set up a Geo Chart for each city, a Day of Month Sessions vs. Conversions and a Day of Week Sessions vs. Conversions chart and a Conversions by Sessions colored by Conversion rate for Audience segments to quickly drill down and understand the differences.

    For this example, what I’m interested in is the difference between In-Market and Affinity Segments by medium and location.

    Houston is a market of strategic importance for my brand, so I will start there:

    After seeing the overall state of affairs, I then want to see the difference between the channels my brand wants more strategic insight on —  in this case, organic search and paid social:

    Organic search

    What I notice is that Cooking and Auto Enthusiasts have the highest conversion percentage, that the middle toward the end of the month has a higher conversion percentage and that no one really is looking to convert on Saturday and Sunday.

    Paid social

    When looking at paid social, I see something drastically different:

    What I notice here is that the majority of the audiences who are clicking on our paid social advertisements have a lower conversion rate, where the highest-converting audiences — Do-It Yourselfers and Comics & Animations fans — are not even audiences we are seeing in organic search.

    The best days of the month tend to favor weeks 3 and 4 but aren’t overwhelmingly similar to organic search, and the best days of the week are virtually the opposite!

    Based on these simple discoveries, I would then work with the larger digital team to ensure flighting, specific client requests or any other variables that may be causing these discrepancies are addressed and tweak our strategy. These are the points I would make and the questions I would ask:

    1. The audiences that are organically discovering us and showing high-performance metrics are different from those we are buying in social. Are we targeting audiences with similar affinities, and they’re just not converting? If we are not, let’s add these characteristics/interests and see if it improves our campaigns.
    2. Each of our channels is most effective on different days of the week. Are we flighting based on the performance shown in analytics? If not, why not?
    3. Are we using paid social for awareness or to drive conversions? If the intent is to drive conversions, we should re-evaluate the assets we’re putting media against in the Houston market.

    It should go without saying that in any media strategy, there is a lot going on. Ensure that you or your data analyst has a full contextual understanding of the information; otherwise, the data is useless.

    All in all, we’ve seen a lot of success leveraging this information to craft better holistic digital strategies. I encourage you all to explore it and see what nuggets you can find.


    About The Author

    Andrew Ruegger
    Andrew Ruegger is Senior Partner, Head of Data Science at GroupM Search & Social - Catalyst. He is a digital veteran with over 7 years of experience in advertising. He currently is responsible for running the data and insights practice for search and social marketing.