Scoring matrices play a crucial role in efficiently prioritizing SEO recommendations. Whether we're creating new content or updating existing pages, these matrices have become an integral part of our workflow. To streamline this process, we've adopted ChatGPT, which has proven to be an invaluable tool in our efforts.
[TIP] Our team even put together an SEO Priority Scorecard Template to help align your SEO recommendations to business goals.
Scoring Process
The prioritization analysis identifies necessary pages to update based on various data points. Traditionally, we manually reviewed each page and assigned a score ranging from 1 to 5, with 5 reflecting the highest importance. However, to expedite the process, we now export this data into a CSV file that consists of additional columns specifically for scoring.
In our test exercise we exported:
-
- Total and top 20 keywords from SEMRush. Additionally, we meticulously organize
- Organic clicks and impressions from Google Search Console
- Conversions from Google Analytics
Other potential factors to support your prioritization efforts include:
-
- Monthly search volume
- Paid clicks and conversions
- Page type
- Search intent
Updating the Process with GPT-4
To assist us in scoring the columns and rank the metrics in order of importance, we uploaded the CSV file to GPT-4. You'll need a ChatGPT Plus plan since GPT-3.5 doesn't have this capability. We used GPT-4's Advanced Data Analysis to easily attach the file into our chat.
After confirming GPT's processing capabilities, we use this prompt for ChatGPT:
I need assistance with the "Score" columns. I need help to score these metrics on a scale of 1-5, where 5 is the highest value and 1 is the lowest. The scoring should be based on dividing the range of values for each metric into five equal parts. Assist me in processing this data and providing the scored output in an updated Excel file.
Provide the ranges you used for each score along with the number of URLs in each range. I would like for these ranges to be as distributed as necessary in order for multiple URLs to be listed within ranges 1, 2, 3, 4, and 5.
Initially, we faced challenges with zeros and high values influencing the scoring process. Subsequently, we modified our criteria and assigned a score of 1 to zeros, a score of 5 to outliers, and distributed the remaining scores between 2, 3, and 4. After several iterations, GPT provides us with an updated range.
After updating the criteria, ChatGPT was able to provide prioritization scores for 1,400+ blog pages so our team understands which pages to focus on first.
What’s Next?
With the assistance of ChatGPT we can efficiently analyze priority pages and identify areas for improvement. Some ideas to iterate on this priority scoring workflow could be:
-
- Refining our prompt and inputs to improve consistency
- Creating a CustomGPT built for this specific purpose
Our team will continue testing how we can leverage AI to enhance traditional processes.
