I Ignored SEO for Two Years Then Spent 6 Months Fixing Everything With AI Tools and Here Is Every Number

For two years I published content and hoped Google would notice. It mostly did not. Then I spent six months actually learning SEO properly with AI tools doing the heavy lifting on research and analysis. The site went from 1800 monthly visitors to 38000. Every decision I made is documented here.

🔧 Tools mentioned in this article

Semrush

All-in-one SEO platform covering keyword research, competitor analysis, site auditing, and content optimization

www.semrush.com

Visit

Surfer SEO

AI content optimization tool that benchmarks your articles against top-ranking pages in real time

surferseo.com

Visit

Frase IO

AI content research and optimization tool that builds briefs, identifies gaps, and scores content for SEO

www.frase.io

Visit

Google Search Console

Free direct Google data showing rankings, impressions, clicks, and indexing for every page on your site

search.google.com

Visit

Two Years of Publishing Into Silence and What Finally Made Me Take SEO Seriously

I want to tell you something embarrassing first. I had been publishing articles on my site for two years and I was getting 1800 organic visitors per month. Not 1800 per day. 1800 per month. For two years of consistent publishing that is a number that should have alarmed me much earlier than it did.

The reason it did not alarm me earlier is that I had convinced myself the traffic would come naturally if I just kept publishing good content. This is a comforting belief that turns out to be mostly wrong. Good content that nobody is searching for does not get found. Good content on topics people are searching for but optimized poorly does not rank. I had been doing both of those things simultaneously for two years.

What finally made me take it seriously was finding a competitor site that had launched 14 months after mine and was receiving 12 times my monthly traffic. Same general topic area. Objectively not better writing. They had just actually done SEO and I had not. That comparison was the cold shower I needed.

Month 1: Semrush Showed Me How Wrong My Content Strategy Had Been

I started with a Semrush site audit that produced a list of 58 technical issues on my site. Most were minor but 9 were significant enough to directly affect how Google was crawling and indexing my content. Thin content flags on 14 pages, missing meta descriptions on 31 pages, duplicate title tags across 8 page pairs, and a crawl depth issue that meant Google was treating some of my best content as lower priority because it was buried too deep in the site structure.

I fixed everything on the critical list in three evenings. It was tedious work but it was finite and most of it was the kind of fix that takes five minutes once you know what needs doing. The crawl issues required restructuring some internal linking which took longer but produced an immediate improvement in how consistently Google was visiting those buried pages according to Search Console.

The keyword gap analysis in Semrush was where my two years of wasted effort became fully visible. I ran the analysis against my four strongest competitors and got a list of 890 keywords those sites were ranking for that my site was not targeting at all. Eight hundred and ninety. I had published over 110 articles and missed 890 keywords that were driving consistent traffic to my competitors. The gap was not close to being closed accidentally through good writing. I had been writing about things nobody was searching for at the scale I needed.

I sorted the gap list by a combination of search volume and keyword difficulty. Anything with monthly searches above 400 and keyword difficulty below 30 went into a priority spreadsheet. That filter produced 94 keywords. At the publishing pace I could maintain those 94 keywords represented about five months of content. I had my roadmap.

Semrush Month 1 Findings

• Technical issues identified in site audit: 58 total, 9 significant enough to affect crawling and indexation
• Time to fix critical issues: approximately 3 evenings
• Competitor keyword gap: 890 keywords competitors ranked for that my site was not targeting
• High-priority keywords identified after filtering by volume and difficulty: 94
• Content already published that targeted these 94 keywords: zero

Semrush Pricing in 2026

1. 1.Free account: 10 keyword searches per day, limited competitor data, 1 project, basic site audit
2. 2.Pro at 139.95 dollars per month: 5 projects, 500 keywords tracked, full keyword and competitor research, full site audit
3. 3.Guru at 249.95 dollars per month: 15 projects, 1500 keywords tracked, Content Marketing Platform, historical data
4. 4.Business at 499.95 dollars per month: 40 projects, 5000 keywords tracked, API access, extended limits, white-label reporting

Months 2 Through 4: Surfer SEO Made My Articles Actually Rank

Understanding which keywords to target was the first problem. The second problem was that my previous articles were not ranking well even when they were on relevant topics because I had no systematic way to ensure they were covering the topic comprehensively enough to compete with what was already ranking.

Surfer SEO analyzes the top-ranking pages for any keyword and tells you what word count, heading structure, and semantic terms appear consistently in the content that Google is rewarding. It then scores your article in real time against these benchmarks as you write. The score is not a direct ranking predictor but it is a reliable indicator of whether your content is covering the topic with the depth and breadth that the current ranking pages demonstrate.

The first article I wrote using a Surfer brief was a humbling experience. I had drafted a piece I was genuinely happy with and it scored 44 on Surfer's scale. The recommended word count was 400 words longer than what I had written and I was missing 22 semantic terms that appeared consistently in the top-ranking pages for the keyword. My article covered the main topic but it was thinner in both length and semantic coverage than everything already ranking above me. I had written good content that was not comprehensive enough to compete.

I rewrote the article to hit Surfer's recommendations and published it with a score of 74. It ranked on page two within 12 days and moved to page one within 35 days. My previous articles on similar topics had either not ranked at all or had ranked on page 3 or below and never moved. The same quality of writing with proper optimization produced dramatically better ranking results.

I used Surfer on every new article from month two onward and also ran the existing content audit on my 110 published articles. The audit identified 34 articles that were ranking between positions 8 and 25 for their target keywords and had significant content gaps relative to the current top-ranking pages. I updated 22 of those 34 articles over months two and three. Nineteen of the 22 updated articles moved to a higher ranking position within four weeks of the updates being indexed.

Surfer SEO Impact on Ranking Performance

• First Surfer-optimized article initial score: 44, rewritten to 74 before publishing
• Days to page 2 ranking after publishing: 12
• Days to page 1 ranking: 35
• Existing articles updated using Surfer content audit: 22
• Updated articles that moved to higher ranking within 4 weeks: 19 of 22
• Average position improvement on updated articles: from average position 16 to average position 6

Surfer SEO Pricing in 2026

1. 1.Essential at 89 dollars per month: 30 articles per month, Content Editor, keyword research, basic SERP analyzer
2. 2.Scale at 129 dollars per month: 100 articles per month, full Topical Map, AI writing assistant, Audit tool
3. 3.Scale AI at 219 dollars per month: everything in Scale plus AI article generation credits
4. 4.Enterprise at custom pricing: unlimited articles, custom AI models, white-label, API access

Months 3 Through 6: Frase IO Handled the Research That Was Slowing Me Down

By month three I was publishing four to five new articles per week to work through the 94-keyword priority list from my Semrush research. At that publishing volume the research phase of each article was becoming a bottleneck. I had been doing research manually which meant opening multiple tabs, reading through competing articles, taking notes, and assembling an outline. For a complex topic this could take an hour before I had written a single word.

Frase IO builds content briefs automatically by analyzing the top-ranking pages for any keyword. Within about 90 seconds of entering a keyword it produces a brief that includes an outline based on headings from top-ranking content, the key questions the article should answer based on People Also Ask data, statistics and facts that appear in competing articles, and a reading time estimate based on competitor word counts. It is the research phase of article writing compressed from an hour into 90 seconds.

I want to be clear about how I used Frase because I think there is a wrong way to use it that produces content that reads like a summary of its competitors rather than an original perspective on a topic. I used the Frase brief as a structural starting point and a checklist of topics to cover, not as a script to follow. The outline it produced told me what I needed to address. What I said about those topics and the specific angle I took was entirely my own contribution.

The question data in Frase was particularly valuable. The People Also Ask section showed me the specific questions readers were coming to the topic with and several of them were angles I had not considered covering in my outline. Adding dedicated sections addressing these questions improved the comprehensiveness scores in both Frase and Surfer and gave articles a better chance of appearing in Google's featured snippets for those question-based queries.

Three of my articles during the experiment appeared in featured snippets within 60 days of publishing. All three addressed specific questions identified in the Frase brief that I would not have included as standalone sections without that data. Featured snippets on those three queries drove a combined 340 additional clicks per month compared to what a standard ranking would have produced at the same position.

Frase IO Impact on Research and Content Quality

• Research time per article before Frase: approximately 45 to 60 minutes
• Research and brief creation time with Frase: approximately 8 to 12 minutes
• Questions identified by Frase that were added to articles as dedicated sections: average 3 to 4 per article
• Articles achieving featured snippet status within 60 days of publishing: 3
• Additional monthly clicks from featured snippets versus standard ranking: approximately 340 combined across the three articles

Frase IO Pricing in 2026

1. 1.Solo at 15 dollars per month: 4 articles per month, full research brief, content scoring, AI writing assistant
2. 2.Basic at 45 dollars per month: 30 articles per month, team seat, all features included
3. 3.Team at 115 dollars per month: unlimited articles, 3 user seats, priority support, advanced analytics
4. 4.Enterprise at custom pricing: custom seats, API access, dedicated account manager

Google Search Console: The Tool That Directed Two Hours of Work Per Week Into the Right Places

I had Google Search Console set up but I had been checking it the way most people check it. Occasionally. With no specific process. Looking at total clicks, feeling good or bad about the number relative to last week, and then closing the tab. This is approximately as useful as checking your bank balance without looking at where the money is going.

The process I developed during the experiment was specific and ran every Friday morning. I filtered the Performance report to show the previous 28 days versus the 28 days before that. I sorted by impressions descending and identified every page where impressions had increased but click-through rate had stayed flat or declined. These pages had earned better rankings but were not earning clicks at the rate their new position should support. The problem was almost always the title or meta description not being compelling enough to earn the click at the new position.

I updated the title and meta description of every page I identified through this process. The rewrite process for each page took about six minutes on average. The click-through rate improvement on updated pages was consistent and visible within two to three weeks of the change. Across six months I updated 44 titles and meta descriptions through this weekly process. The cumulative traffic improvement from these updates was significant enough that I tracked it separately as its own traffic growth driver.

Search Console also revealed something I would not have found any other way. Fourteen pages on the site were receiving meaningful impressions for keywords that were completely different from what those pages were written to target. In every case the page had some content that touched on the keyword but had never been written to address it directly. Updating those pages to properly cover the unintended keyword they were ranking for produced fast traffic improvements because the ranking signal was already established and the content quality improvement sent a strong freshness signal to Google.

Google Search Console Results Over 6 Months

• Title and meta description updates based on high impression low CTR pattern: 44 pages
• Average CTR improvement on updated pages: went from 1.9 percent to 3.8 percent average
• Pages updated to better address unintended high-impression keywords: 14
• Time spent on weekly Friday Search Console review: 25 minutes
• Traffic growth attributable to CTR optimization work identified through Search Console: estimated 18 percent of total six-month growth

The Month by Month Traffic Growth

The growth was not linear and it was not comfortable in the early months. Months one and two felt like I was doing a lot of work for modest results. The compounding effect of SEO is real but it is not immediately visible and there were several weeks where I had to actively choose to continue rather than question whether anything was working.

• Month 0 starting point: 1800 monthly organic visitors
• Month 1: 2200 visitors, technical fixes indexed, first gap-keyword articles published
• Month 2: 3800 visitors, optimized articles beginning to rank, content updates showing movement
• Month 3: 7400 visitors, keyword gap content ranking consistently, CTR optimization adding incremental volume
• Month 4: 14200 visitors, topical authority building, featured snippets appearing
• Month 5: 24800 visitors, established articles compounding, new content ranking faster
• Month 6: 38000 visitors, content velocity producing accelerating returns

The Things I Got Wrong Along the Way

I published 12 articles in month one based on the Semrush keyword gap list without using Surfer SEO to optimize any of them because I had not added Surfer to my workflow yet. When I eventually audited those articles with Surfer in month three they all had content scores below 50 and I had to go back and rewrite them. The time I spent rewriting those 12 articles was time I could have spent on new content. Doing everything right from the start rather than adding tools incrementally would have produced faster overall results.

I also spent time in month two trying to build links through guest posting outreach and generated almost no results. Link building at meaningful scale requires either a significant time investment in outreach or content specifically designed to attract links naturally. I had neither at the time and the two weeks I spent on outreach would have produced more traffic improvement if I had spent them on new optimized content instead.

The honest summary of what drove the traffic growth is that it was approximately 80 percent content strategy and content optimization and approximately 20 percent technical fixes and click-through rate improvement. I had expected link building to be a more significant factor and it was not for the niche and competitive level I was operating in. Your niche may be different.

How to Build This Stack if You Are Starting From Where I Was

If I were starting over knowing what I know now the sequence would be different from how I actually did it. The biggest mistake I made was adding tools sequentially rather than starting with all of them simultaneously which meant some early content was not optimized properly and needed to be redone.

1. 1.Set up Google Search Console and verify your site before publishing any new content, you cannot recover historical data and every week without it is data you will eventually wish you had
2. 2.Run a Semrush site audit and fix all critical technical issues before worrying about new content, ranking new content on a technically broken site is slower and harder than it needs to be
3. 3.Run the Semrush keyword gap analysis against your top three to five competitors and build your content priority list before writing a single new article
4. 4.Add Surfer SEO to your workflow from the first new article you publish, not after you have published 12 articles without it
5. 5.Add Frase IO at whatever publishing volume makes the research phase feel like a bottleneck, for most publishers that is somewhere between two and four articles per week
6. 6.Run the Friday Search Console review every week without exception from month one, the CTR optimization work it identifies is consistently high-return and takes under 30 minutes

Final Thoughts

Two years of publishing without SEO strategy produced 1800 monthly visitors. Six months of publishing with a data-driven strategy supported by AI tools produced 38000. The writing quality was comparable across both periods. The difference was entirely in whether the content was aimed at what people were actually searching for and whether each article was optimized to compete for the rankings it was targeting.

Semrush told me where to aim. Surfer SEO made each article competitive. Frase IO made the research fast enough to publish consistently. Google Search Console told me what was working and where to spend the improvement time each week. None of these tools wrote the content. They made the content I was already capable of writing significantly more likely to reach the people it was written for.