We’ve heard before that businesses will log into their Google Analytics account once a year or so, stare blankly at their dashboard for 30 seconds, then close the browser as if nothing happened. 😅
If you’re already using Claude, then you can (and should) put it to work for you to process the data and help break down what it means for your website - starting with your top performing blog posts.
Understanding how and why your blog is performing the way it is can give you better insight into the type of content you create, where you promote your blog, and even if you’re missing the opportunity to pick up new customers.
Step 1: Prep Your Claude Space to Get Started
The first step is to prep your Claude space.
If you’re working in Claude Chat: Start a new chat conversation.
If you’re working in Claude Cowork: Create a task that will live in a dedicated website project going forward.
Assuming you haven’t already trained Claude on your business, upload (or explain) your company’s mission, vision and values. Be sure to break down your product/service model, as well as your target audience demographic. You want Claude to know as much as you do about your business so it has full context going forward.
Here’s an example training prompt:
My business [business name] is located in [city, state] but serves [general service location]. The business offers [products/services] to [primary audience/demographic]. Our mission is to [mission statement]. We value [list of values].
Step 2: Define Your Goals and Data Points
Goal: Determine which blog posts users are engaging with the most.
Primary data points:
- Average Engagement time: How long users are actively interacting with the page, not just having it open.
- Engaged Sessions: Google Analytics defines this as a user having 10+ seconds of active engagement time, having a conversion event, or viewing 2+ pages.
- Engagement Rate: This is a calculated ratio of engaged sessions ÷ total sessions expressed as a percentage.
Another secondary data point to consider is New vs. Returning Users. You can add this as a secondary dimension to your reporting as returning users on a blog post signals it's reference-worthy or memorable. However, Google Analytics tends to undercount returning users since it’s cookie-based, meaning that the same user could view the blog post on two different devices, or in two different browsers, and they would each be counted uniquely.
Step 3: Download the Data from Google Analytics
You’ll want to pull two separate reports using the general website report as a benchmark so Claude can see how your blog posts perform relative to your site as a whole.
Average Website Engagement
- Go to Engagement > Pages and Screens.
- Set the date range to “past 90 days” or a custom date range.
- Change “page path and screen class” to “page title and screen class” to see blog posts by name.
- Click the “Share this Report” icon > Download File > Download CSV.
Blog Post Engagement
- Go to Engagement > Pages and Screens.
- Set the same date range as above.
- Add a filter to find only blog posts. (This will be based on the page path for your blog. For example: All of Pixel Jam’s blog posts live under /blog/.) Here’s how you can set up the filter based on page path:
- Set dimension to “Page path and screen class”
- Set match type to “contains”
- Enter “/your-path/” into value
- Change “page path and screen class” to “page title and screen class” to see blog posts by name.
- Click the “Share this Report” icon > Download File > Download CSV.
Step 4: Prompt Claude To Assist with Your Blog Post Analysis
Upload the Google Analytics report to Claude, along with the following prompt:
I want to determine the top five most engaging blog posts on my website. I’m attaching several reports from Google Analytics over the past [your specified number] days including: general website engagement across all pages, and engagement specifically from blog posts. I want you to factor in the following metrics to determine engagement: average engagement time, engaged sessions, and engagement rate. Please provide an itemized list of the top five blog posts, with a summary of why they’re considered to be engaging.
Claude should then output a list of the blog posts with a description as to why they're considered engaging.
Step 5: Get Additional Insights into Blog Post Performance
Now that you know which blog posts are the most engaging, you can use this information to determine the following:
Which blog posts could stand to be promoted more. Claude can help find content that’s resonating with readers but just hasn't been seen by enough people yet. This will determine which posts could stand to be promoted more as they aren’t getting the page views they deserve.
The most searchable, evergreen blog posts. These will be the blog posts with search-friendly topics, strong dwell time, and the staying power to keep pulling organic traffic without being republished constantly.
The best blog posts that are already positioned to capture intent. Once you understand which blog posts present the best opportunity to capture leads, you can add calls-to-action on these pages to point to either a lead generation form, or a free digital download to keep them in your marketing funnel.
Which blog posts would be best to expand upon as part of a series breakdown. These posts would be the best candidates for a series that already prove audience interest, cover a topic broad enough to sustain multiple installments, and connect to your services.
Here’s an example prompt to get started with these additional insights:
Looking at these top performing blog posts, I want to understand [new insight request].
For example:
Looking at these top performing blog posts, I want to understand which posts are the most evergreen so I can create similar blog posts that will continue to attract organic traffic.
A Quick Caution When Working with Real Data
Remember, AI is confident even when it’s wrong.
While we don’t expect you to turn into a data nerd overnight, you can reference the Google Analytics data while reading through the results to see if the numbers and explanations match up. For example: if Claude says a post has a 4.5 minute engagement time, you should be able to verify that number in your GA4 report.
Plus, the more you look at the data, the more confident you’ll become in analyzing your site’s performance! (And when that happens, we’d love for you to share it with us!)
We Care About Your Website
If analyzing your website data starts to feel overwhelming, that's exactly what we're here for. Our Monthly Web Care Plan includes real-time data review, performance monitoring, and a monthly check-in where we walk through what the numbers actually mean for your business.
We can’t promise that your eyes won’t glaze over, but we’ll do our absolute best to break everything down into actionable steps.
Schedule a consultation so we can chat about how we can help!