Web Analysis

By /

To understand how your website is performing, you can use Google Analytics (GA4), to examine key metrics such as user acquisition, page views and unique page views, and demographic details. User acquisition shows how many new users your website is getting as well as the sources the users are coming from. It allows you to see which platforms or sources are the most effective for gaining users for your website. Page views and unique page views show which blog posts are the most popular to website visitors. This allows you to see which types of content are most appealing to your target market. Demographic details allows you to see if you are targeting the specific geographic regions you were trying to reach. We will also use ChatGPT to analyze the GA4 data.

User acquisition – First user source / medium

We use first user source / medium to show where the new users are coming from. The metrics we focus on in this section are new users, engagement rate, and average engagement time. New users show you how many first time users are visiting your sits. Engagement rate shows the percentage of sessions that were engaged. Average engagement time shows the average amount of time a user typically spends engaged on your website. Below is an image of the GA4 analytics that show the user acquisition with these three stated metrics for my blog website you are currently visiting.

After gathering this information, I entered the relevant information into an excel sheet to make the information more organized and easier to put into ChatGPT to analyze. Below is the excel sheet information.

After organizing the information into excel, I entered the information into ChatGPT along with specific prompts to further analyze and understanding the meaning of these metric values. I entered the prompt followed by copying and pasting the first six cells under values in the excel sheet related to average engagement time. The prompt I used was: When I sorted the average engagement time in descending order in the First user source/medium classification, the following result came out. Please explain. Here is the result ChatGPT gave me.

The table you’ve shown represents user traffic sources and their corresponding average engagement times. Here’s an explanation of each element:

  1. (direct) / (none) – 0s: This refers to users who accessed your website directly, either by typing the URL into their browser or using a bookmark. The “0s” means these users spent no measurable time engaging with your website (likely bounced immediately).
  2. d2l.sdbor.edu / referral – 0s: This indicates that users were referred from the domain d2l.sdbor.edu (likely a learning platform or educational resource). The engagement time being 0 seconds suggests minimal or no interaction with your website.
  3. google / organic – 0s: These users arrived via organic (unpaid) search results on Google. The “0s” engagement time indicates no effective interaction with your content.
  4. linkedin.com / referral – 0s: Traffic referred from LinkedIn’s main domain. Again, the engagement time being 0 seconds shows no meaningful interaction.
  5. lnkd.in / referral – 0s: lnkd.in is LinkedIn’s URL shortener. Users coming through shortened links also show 0 seconds of engagement time.
  6. mail.google.com / referral – 0s: This reflects referrals from Gmail (e.g., clicking a link in an email). A 0-second engagement time suggests that users clicked the link but didn’t stay or interact with the site.

Possible Reasons for 0-Second Engagement Times:

  • Users bounced (left immediately after visiting the page).
  • Technical issues with tracking (e.g., delayed JavaScript execution for analytics).
  • Low relevance of landing pages to the referring sources.
  • Links leading to pages with insufficient or unclear calls to action (CTAs).

Improving these metrics might involve analyzing the landing page content, ensuring it aligns with user expectations, and testing analytics setups for accuracy.

For the next section, I used ChatGPT to analyze the data collected on new users to the website. I entered the prompt into ChatGPT followed by copying and pasting the next six cells under the values heading that corresponded to new users. The prompt I used was: When I sorted the new users in descending order in the First user source/medium classification, the following result came out. Please explain. This was the result that ChatGPT produced. 

The table shows the breakdown of new users by their source/medium in descending order, explaining how people first found or accessed your website. Here’s what it means:

  1. (direct) / (none) – 192:
    • These 192 new users accessed the website directly, without a referring source. This often indicates:
      • Users typed your URL directly into the browser.
      • Used a bookmark to visit your site.
      • Came from sources that aren’t tracked (e.g., certain apps or secure environments).
  2. d2l.sdbor.edu / referral – 6:
    • Six new users were referred from the d2l.sdbor.edu domain, likely from a link embedded within this platform (possibly an educational platform or portal).
  3. google / organic – 5:
    • Five new users arrived from unpaid Google search results. This suggests that people found your site by searching for relevant terms or keywords on Google.
  4. linkedin.com / referral – 2:
    • Two new users came via referral links from LinkedIn. This could be from posts, profiles, or LinkedIn groups where links to your website were shared.
  5. lnkd.in / referral – 1:
    • One new user came via LinkedIn’s URL shortener (lnkd.in), which is often used when sharing links on LinkedIn posts, messages, or ads.
  6. mail.google.com / referral – 1:
    • One new user found your website by clicking a link in an email accessed through Gmail. This might indicate an email marketing campaign or a shared link.
Analysis:
  • The majority of new users (192 out of 207 total) come from direct traffic. This indicates strong brand awareness or traffic from untracked sources.
  • Referrals (e.g., from d2l.sdbor.edu, LinkedIn, Gmail) and organic search are contributing smaller but relevant portions of new users.
  • If engagement times are low for these new users (as in your earlier data), it could mean:
    • Landing pages may not be meeting user expectations.
    • Referrals or campaigns may need better targeting or messaging.

To grow these numbers, you might consider:

  • Boosting organic search rankings by improving SEO.
  • Expanding outreach efforts on platforms like LinkedIn.
  • Refining referral strategies, especially for educational platforms like d2l.sdbor.edu.

The last section, I used ChatGPT to analyze the data collected on the engaged users per active session. I entered another prompt into ChatGPT followed by copying and pasting the last six cells under the values heading that are related to engaged users per active session. The prompt I used was: When I sorted the engaged sessions per active user in descending order in the First user source/medium classification, the following result came out. Please explain. This was the results ChatGPT produced.

The table shows the engaged sessions per active user percentage (sorted in descending order) across different traffic sources. This metric measures the percentage of sessions where users were actively engaged (e.g., interacted with the page, scrolled, or clicked) compared to the total number of sessions. Here’s what it means for each row:

  1. (direct) / (none) – 0.21%:
    • Only 0.21% of sessions from direct traffic were classified as “engaged.” While direct traffic brings the most users, engagement from this group is low, indicating that many users leave the site quickly or without meaningful interactions.
  2. d2l.sdbor.edu / referral – 0.67%:
    • This source has the highest engaged session rate at 0.67%, suggesting that users referred from the educational domain d2l.sdbor.edu are more likely to interact with the site compared to other sources. This could be due to targeted content or relevance to their interests.
  3. google / organic – 0.00%:
    • No engaged sessions were recorded from organic Google search traffic. This could indicate:
      • Landing pages are not aligned with user search intent.
      • Users are bouncing without interacting after clicking through search results.
  4. linkedin.com / referral – 0.00%:
    • No engaged sessions occurred from users referred via LinkedIn. This suggests that links shared on LinkedIn are not leading to relevant or engaging content for these users.
  5. lnkd.in / referral – 0.00%:
    • Similar to LinkedIn’s main domain, no engagement is recorded from links using LinkedIn’s shortener (lnkd.in).
  6. mail.google.com / referral – 0.00%:
    • Users coming from Gmail (via links in emails) also show no engagement. These users might be quickly leaving the page or not interacting with the content.
Analysis and Recommendations:
  • Low Engagement Across Sources: Except for d2l.sdbor.edu, most sources have extremely low engagement rates. This indicates an issue with the site’s ability to capture user interest or guide users toward meaningful actions.
  • Focus on Landing Pages:
    • Ensure that landing pages are tailored to match user expectations based on the source/medium.
    • Use clear, visible calls-to-action (CTAs) to encourage interactions.
  • Analyze Content Relevance:
    • For organic search, review the keywords driving traffic and align content with user intent.
    • For LinkedIn and email referrals, ensure that shared links lead to pages that deliver on the promise of the source (e.g., specific offers or content).
  • Improve Tracking Accuracy:
    • Ensure that the analytics setup is capturing all interactions accurately, especially for engaged sessions. Misconfigured analytics can also lead to low engagement rates.

These adjustments can help boost the overall engagement rate for your website traffic.

Engagement – Page title and screens

We use page title and screens to determine which pages or blogs posts are seeing the most engagement. The metrics we look at in this section are average engagement time, views, and new users. Average engagement time is used to show how long users are spending on the page content. The higher the average engagement time, the more interested people are in your content. Views shows how much the content is being discovered and viewed. The higher the views are, the more effective you know your promotional strategies have been. New users is how many new people have visited your content. Below is an image of the GA4 analytics that show the engagement with these three stated metrics for my blog website you are currently visiting.

After collecting this information, I entered the relevant information into an excel sheet to help make the information more easily readable and easier to copy into ChatGPT. Below is the excel sheet information.

Now that the information is laid out in excel, I entered the information along with specific prompts into ChatGPT to further analyze this information related to these three metrics. I entered the prompt followed by copying and pasting the first six cells of information on the far right side related to average engagement time. The prompt I entered into ChatGPT was: When I sorted the average engagement time in descending order in the Page title and screen class classification, the following result came out. Please explain. This was the result ChatGPT gave me.

The table represents average engagement time for different page titles or screen classes sorted in descending order. The data suggests that all these pages have 0 seconds of average engagement time, which means users are not spending any measurable time actively engaging with the content. Here’s an explanation for each row:

  1. About – Digital Marketing Blog – 0s:
    • Users visiting the “About” page of your blog are leaving immediately without interaction. This could indicate a lack of engaging or relevant content on the page, or users are not finding the information they are looking for.
  2. Blog – Digital Marketing Blog – 0s:
    • The main blog page, which likely lists your articles or provides an overview, is also showing no engagement. This could mean users are not clicking through to articles or finding the page appealing.
  3. Contact – Digital Marketing Blog – 0s:
    • A “Contact” page typically serves as a way for users to reach out. The lack of engagement may mean:
      • Users don’t see a compelling reason to stay (e.g., no clear or easily accessible contact information).
      • There might be a technical issue preventing proper tracking of user activity.
  4. Content Creation with AI for Scheels – Digital Marketing Blog – 0s:
    • This blog post is receiving no engagement, suggesting users aren’t scrolling, clicking, or reading. It’s possible that:
      • The content isn’t relevant or enticing to users.
      • The page layout or design might be discouraging interaction.
  5. Content Creation with AI for The Homeslice Media Group – Digital Marketing Blog – 0s:
    • Similar to the post above, users visiting this page are not interacting. The same potential issues apply: relevance, design, or targeting.
  6. Customer Persona for Scheels – Digital Marketing Blog – 0s:
    • This page, which presumably focuses on creating a customer persona for Scheels, also shows no engagement. This could indicate the content is not resonating with the audience or there are technical issues affecting engagement tracking.
Analysis:
  • Common Issues:
    • Bounce Rate: Users might be bouncing (leaving immediately) from these pages, leading to 0s average engagement time.
    • Relevance: The content might not align with the audience’s expectations or needs.
    • Tracking Issues: There may be problems with how analytics is configured, resulting in an inability to accurately capture user interactions.
Recommendations:
  1. Review Analytics Setup:
    • Verify that tracking is working correctly, especially for engagement-related events (e.g., scrolling, clicks).
  2. Improve Page Content:
    • Evaluate whether these pages deliver value to users. For example:
      • Is the “About” page providing meaningful information about your business or blog?
      • Are blog posts addressing topics that resonate with your audience?
      • Is the “Contact” page easy to navigate, with prominent and actionable contact options?
  3. Enhance Page Design:
    • Ensure that pages are visually appealing and encourage interaction. Include CTAs (e.g., “Learn More,” “Contact Us”) that guide users to take action.
  4. Optimize for Audience:
    • Use audience insights (e.g., customer personas) to tailor your content and ensure that it meets user expectations.

Addressing these areas will help improve engagement on your site. Additionally, check for any technical issues that might skew data.

For the next section, I used ChatGPT to analyze data regarding views on the site. I entered a prompt into ChatGPT followed by copying and pasting the next six cells on the far right side related to views. This was the prompt I entered: When I sorted the views in descending order in the Page title and screen class classification, the following result came out. Please explain. This was the result ChatGPT produced.

The table reflects page titles or screen classes sorted by the number of views, with all entries showing 0 views. This indicates that none of the listed pages were viewed during the period analyzed. Here’s what this could mean and how to interpret the results:

Possible Reasons for 0 Views:
  1. No Traffic to These Pages:
    • These pages were not visited by any users during the analyzed period.
    • This could result from a lack of promotion, visibility, or relevance in your audience’s interests.
  2. Issues with Website Analytics Setup:
    • The tracking code may not be properly implemented on these pages, leading to no data being recorded for views.
    • Misconfigured page titles or screen classes in your analytics setup might result in no data being attributed to these specific pages.
  3. Pages are Inaccessible:
    • The pages might be offline, unpublished, or have broken links preventing users from accessing them.
    • There could also be navigation issues, such as links to these pages not being prominently displayed or functional.
  4. Incorrect Filters or Data Sampling:
    • If filters were applied to exclude certain types of traffic or data, these pages might have been inadvertently excluded.
    • Sampling issues in your analytics platform could also lead to incomplete data for less popular pages.
Recommendations:
  1. Verify Analytics Implementation:
    • Ensure that tracking codes (e.g., Google Analytics or other tools) are properly installed on all pages.
    • Check if the page titles or screen classes are correctly reported in the analytics setup.
  2. Test Accessibility:
    • Confirm that these pages are live, accessible, and linked correctly from other areas of your website.
    • Use tools like site crawlers to detect any broken links or issues preventing users from reaching these pages.
  3. Promote These Pages:
    • If these pages are important, ensure they are effectively promoted. For example:
      • Include links to these pages in your blog posts, navigation menu, or social media campaigns.
      • Optimize them for search engines (SEO) to attract organic traffic.
  4. Analyze User Behavior:
    • Review your website’s traffic flow and entry/exit points to identify why users are not reaching these pages.
    • Adjust content or navigation to encourage users to visit these pages.
  5. Remove or Update Low-Relevance Pages:
    • If these pages no longer serve a purpose, consider updating their content or removing them entirely to maintain a clean and focused website structure.

By addressing these potential issues, you can identify whether the lack of views is due to technical errors, user behavior, or a need for improved visibility.

In the last section, I used ChatGPT once again to analyze the new users that visited the website. I entered the following prompt into ChatGPT followed by copying and pasting the last six cells of information that are related to users. The prompt I used was: When I sorted the users in descending order in the Page title and screen class classification, the following result came out. Please explain. This was ChatGPT’s response.

This table represents the number of users (sorted in descending order) who visited specific pages or screen classes. Each entry shows how many unique users accessed a page during the analyzed period. Here’s the interpretation:

Explanation of Each Page:
  1. Contact – Digital Marketing Blog – 31 Users:
    • The “Contact” page received the highest number of users, likely because users wanted to find ways to reach out or get more information about your business. High traffic here suggests interest in communication, but earlier data (0 average engagement time) indicates that users might not be finding what they need or are leaving quickly.
  2. Content Creation with AI for Scheels – Digital Marketing Blog – 19 Users:
    • This blog post received a significant amount of traffic. Users may have found the topic engaging or relevant, especially if it was promoted well or aligned with search queries.
  3. Customer Persona for Scheels – Digital Marketing Blog – 18 Users:
    • Another blog post with substantial user traffic. This indicates user interest in topics related to customer personas, particularly for a well-known brand like Scheels.
  4. Blog – Digital Marketing Blog – 11 Users:
    • The main blog page, typically an overview or a list of posts, had moderate traffic. This is expected, as users may be browsing through blog content before clicking on specific posts.
  5. About – Digital Marketing Blog – 5 Users:
    • The “About” page had relatively low traffic, which could mean users are less interested in learning about the blog or company itself, compared to other pages.
  6. Content Creation with AI for The Homeslice Media Group – Digital Marketing Blog – 4 Users:
    • This blog post had the least traffic among the listed pages. While it may be relevant, it likely needs more promotion or optimization to attract users.

Analysis and Recommendations:

1. High Traffic but Low Engagement:
  • Although some pages (e.g., “Contact” and blog posts) attract significant traffic, earlier data showing 0 seconds of engagement time suggests users are not interacting with the content.
  • Recommendations:
    • Improve the design and usability of these pages (e.g., ensure the “Contact” page has clear forms or CTA buttons).
    • Review the blog posts’ content to ensure they are engaging, informative, and aligned with user expectations.
2. Promote High-Value Pages:
  • Pages like “Content Creation with AI for Scheels” and “Customer Persona for Scheels” are attracting decent traffic, but they could be further promoted via:
    • Social media campaigns.
    • Email newsletters.
    • Internal linking from other blog posts.
3. Improve Low-Traffic Pages:
  • The “About” page and “Content Creation with AI for The Homeslice Media Group” may need content updates or better promotion to increase their visibility. For example:
    • Use SEO techniques to target keywords relevant to these pages.
    • Highlight these pages in navigation menus or homepage banners.
4. Content Alignment:
  • Align page content with the intent of the audience. For example:
    • The “Contact” page should have intuitive forms, phone numbers, or email links.
    • Blog posts should have clear CTAs (e.g., “Read More,” “Contact Us for Help”).

By addressing these issues, you can convert high user traffic into meaningful engagement and action on your website.

User attribute – Demographic details

For the demographic details, we focus on country and city. We focus on country and city to see if the website content is reaching the desired target audience. The metrics we focus on in this section are active users, engagement rate, average engagement time per active user, engaged session per active user. Active users shows how many people visit from each location. Engagement rate shows how many users from each location are engaged with your content. Average engagement time per active user focuses on how deep the engagement is from each engaged active users. And engaged session per active user shows how often engaged users from each location return to your site. Below is an image of the GA4 analytics that show the user attribute of demographic details with these four stated metrics from my blog website.

Now that I have identified some relevant information, I entered a few select locations and their information into an excel sheet to easier identify and use this information for further analysis with ChatGPT. Below is the excel sheet with the information.

Now that the information is summarized in excel, we’ll use ChatGPT to further analyze this information. For this section, we group all four metrics into one prompt and had ChatGPT only give one response instead of multiple. The prompt that was used was: Let’s begin the analysis assuming my cities of interest are Rapid City, Spearfish, Washington, Cheyenne, and Des Moines. When I sorted Active users in descending order in the Country and city classification, the following result came out. Please explain <add data on Active users here> When I sorted Engagement rate in descending order in the Country and city classification, the following result came out. Please explain <add data on Engagement rate here> When I sorted Average engagement time per active user in descending order in the Country and city classification, the following result came out. Please explain <add data on Average engagement time per active user here> When I sorted Engaged sessions per active user in descending order in the Country and city classification, the following result came out. Please explain <add data on Engaged sessions per active user here> Finally, please provide a conclusion based on the answers to these four questions. I would copy and paste the corresponding excel sheet information into each of the areas that said to add data here to get the completed prompt. This was the result ChatGPT gave me.

Here’s a breakdown of the results and analysis for the cities of interest based on Active users, Engagement rate, Average engagement time per active user, and Engaged sessions per active user:

1. Active Users (Descending Order):
  • Des Moines (26 users): Highest number of active users.
  • Cheyenne (14 users): Moderate traffic.
  • Washington (11 users): Slightly lower traffic than Cheyenne.
  • Rapid City (10 users): Comparable to Washington but lower traffic.
  • Spearfish (7 users): Smallest user base among the cities of interest.

Implication: Des Moines drives the most traffic, indicating strong interest or outreach effectiveness in this area. Spearfish has the fewest active users, which may indicate less awareness or reach.

2. Engagement Rate (Descending Order):
  • Spearfish (72%): Exceptional engagement rate, indicating that users from this city are highly engaged and find the content relevant or appealing.
  • Rapid City (42.31%): Moderate engagement, showing some level of interest from users.
  • Washington, Cheyenne, Des Moines (0%): No measurable engagement, which suggests users from these cities may leave immediately without meaningful interaction.

Implication: Spearfish and Rapid City show promise with active user engagement. The lack of engagement in Washington, Cheyenne, and Des Moines indicates potential issues with content relevance, user targeting, or page experience for these cities.

3. Average Engagement Time per Active User (Descending Order):
  • All cities (0 seconds): None of the cities show any measurable engagement time.

Implication: This metric highlights a significant concern—users across all cities are not spending enough time on the site to register meaningful activity. This could result from: – Users bouncing (leaving immediately). – Misconfigured tracking preventing proper recording of engagement time.

4. Engaged Sessions per Active User (Descending Order):
  • Spearfish (2.57 sessions): Users in Spearfish have a high number of engaged sessions per user, indicating strong interaction levels.
  • Rapid City (1.10 sessions): Moderate engagement, with users interacting somewhat but less than those in Spearfish.
  • Washington, Cheyenne, Des Moines (0 sessions): No engaged sessions recorded, indicating a complete lack of interaction from users in these cities.

Implication: Spearfish stands out as a city with engaged users. Rapid City also shows some engagement, but other cities are entirely disengaged.

Conclusion and Recommendations:
  1. Spearfish and Rapid City Are Key Focus Areas:
    • Spearfish exhibits a high engagement rate (72%) and strong engaged sessions per user (2.57) despite having fewer active users. This city shows the highest interaction potential and should be prioritized for further marketing and content strategies.
    • Rapid City also performs well with moderate engagement levels, indicating that efforts in this city are reasonably effective.
  2. Des Moines, Cheyenne, and Washington Need Attention:
    • While Des Moines has the highest number of active users (26), the 0% engagement rate and lack of engaged sessions suggest these users are not finding the content relevant or appealing.
    • Similar patterns in Cheyenne and Washington suggest the need to review targeting strategies and optimize content to better engage these audiences.
  3. Critical Issue with Engagement Time:
    • The 0 seconds of engagement time across all cities is concerning. This indicates either:
      • Users are leaving immediately (high bounce rate).
      • Tracking errors are failing to record engagement properly.
    • Action: Conduct a review of website analytics setup to ensure accurate tracking and resolve any technical issues.
  4. Recommendations to Improve Overall Engagement:
    • Content Optimization:
      • Tailor content to resonate with audiences in Des Moines, Cheyenne, and Washington.
      • Highlight Spearfish and Rapid City success factors to replicate across other cities.
    • Better Targeting:
      • Ensure campaigns in underperforming cities are aligned with audience preferences and needs.
    • Improve Page Experience:
      • Evaluate landing pages to ensure they are engaging and encourage meaningful user interactions.
    • Technical Review:
      • Verify analytics configuration to rule out data collection issues.

By focusing on these strategies, you can boost engagement and interaction in underperforming cities while capitalizing on the success in Spearfish and Rapid City.

Conclusion

This process shows how you can utilize Google Analytics (GA4) to review how your website is performing and who it is effectively reaching. It allows you to understand where you are succeeding and where you need to make improvements. With the use of an AI Chatbot, ChatGPT, we were able to take the information gathered from Google Analytics and input that into ChatGPT along with specific prompts to further analyze and understand what each of these values mean for the website.

Scroll to Top