What Problem Were They Trying to Solve?
For years, Innovative Automations relied on outside SEO consultants to help improve their website performance. While those engagements produced technical analysis, it rarely translated into a clear understanding of what was actually wrong or meaningful forward progress. The goal in this episode was to find out whether Claude could replace that consulting relationship and produce better, more actionable results.
Why Did Google PageSpeed Insights Matter So Much?
Google PageSpeed Insights measures how quickly a page loads and how good the user experience is for someone requesting that page. The scores feed directly into how Google decides whether to recommend a site in search results. Before making any changes, the Innovative Automations site scored a 52 on desktop and a 35 on mobile. Given that Google places significant weight on the mobile experience, that 35 was especially damaging to their search visibility. The site had been professionally designed and looked polished, but those design choices were quietly hurting its performance in Google's eyes.
How Did They Use Claude to Analyze the Problem?
The team pulled their existing PageSpeed Insights reports and fed that data directly into Claude through the chat interface. From there, they had a back-and-forth conversation asking Claude to explain what each metric meant, why it mattered, and what could be done about it. One of the key advantages here was the ability to ask follow-up questions without any hesitation. Unlike a consulting engagement where time is billed and there is pressure to appear informed, Claude made it easy to say "I don't understand this" and keep asking until the answer was genuinely clear.
What Kinds of Questions Did They Ask Claude?
The team pushed past surface-level explanations by asking things like: Are you sure about this? How does Google interpret this metric? How would fixing this actually improve the score? That iterative questioning process helped them move from a list of technical issues to a prioritized set of recommendations they could act on.
How Did They Use Claude Code to Implement the Fixes?
Once they had a clear set of recommendations from the Claude chat session, they brought those findings into Claude Code. They provided the current site metrics and the recommendations, then asked Claude Code to help implement the changes directly. The process was iterative. They made changes, reviewed the results, and continued tweaking until the scores reflected the improvements they were looking for.
What About Tags and Tracking Scripts They Could Not Remove?
Some elements on the site, such as analytics tags and tracking scripts that help the team understand visitor behavior, could not simply be deleted. Claude helped them find a middle path: adjusting when and how those scripts loaded so they had a smaller impact on PageSpeed scores without sacrificing the tracking functionality the team depends on.
What Were the Results?
After applying Claude's recommendations through Claude Code, the results were substantial. The mobile score jumped from 35 to 97. The desktop score moved from 52 to 100. Metrics that had been dragging the scores down, including blocking time and first contentful paint, were addressed through the implementation changes.
How Do They Use This Process on an Ongoing Basis?
The team now runs PageSpeed Insights on a weekly basis and checks every new page or post they publish before it goes live. Whenever the score shows a problem, they run it through Claude, get feedback, and optimize before moving on. This turns what used to be an occasional consulting project into a routine part of their publishing workflow.
How Can You Reach the Team With Questions?
If you want to learn more about the specific tools and process used in this episode, the team is open to walking you through it. You can reach them at ideas@inovativeautomations.ai.