Plumber SEO Case Study: 137 to 1330 Clicks Over 6 Months

The starting problem was a weak search architecture, not simply a lack of activity. Average position sat around 34, and the plumbing site's commercial pages had little supporting depth behind them.

Across the case-study window, monthly clicks moved from 137 to 1,330 while impressions moved from 22,782 to 57,836. Tracked conversions moved from 4 to 45, and modeled revenue based on average job value moved from 1,360 to 15,300 per month. The useful part is the sequence: resolve crawl and intent ambiguity first, then build supporting coverage, then consolidate and reinforce the pages that can generate leads. Those observations come from the masked synthetic dataset described in the evidence policy, so they are a planning example rather than verified third-party performance.

This case follows an anonymized plumbing lead-generation site competing in local home services. The commercial goal was to turn relevant searches into calls, forms, and booked work, which makes page intent more important than raw publishing volume.

At kickoff, average position was around 34. The site also recorded Domain Rating 14, 26 referring domains, and 98 total backlinks. That is enough existing authority to work with, but not enough to justify a strategy based on link volume alone.

Three constraints shaped the plan: established local competition, limited content production capacity, and a firm evidence boundary for public claims. The practical consequence was to concentrate relevance rather than add pages indiscriminately, and to describe results as modeled observations instead of promises.

An average position near 34 was only the symptom. The more important issue was that commercial and transactional intent were not clearly separated across the site's conversion paths.

The main service URL was trying to answer emergency, pricing, comparison, and general service intent at the same time. That made it difficult to tell which page should own which query and where supporting internal links should point.

• Intent overlap. Commercial and transactional pages were not differentiated sharply enough, so relevant signals were split.
• Thin informational support. Homeowners researching repair, maintenance, pricing, replacement, or emergency issues had few useful supporting pages to enter through.
• Diffuse authority. Internal links and crawl attention reached pages that were less important to lead generation instead of concentrating on the strongest commercial destinations.

The decision was to fix page ownership before scaling content. More pages would have added more places for the same signals to compete.

The work was organized into six connected streams, with each stage intended to support the next rather than operate independently.

1. Clean the technical base before adding scale (months 1 to 3)

Crawl and indexation triage came first, followed by canonical and redirect cleanup, template duplication review, and checks on rendering, internal status codes, and schema. Average position changed from 34.5 toward 22.9 by the end of month three. That timing is consistent with cleaner page ownership, but the case does not treat it as proof that technical work alone caused the movement.

2. Separate intent and build supporting coverage (months 2 to 4)

Target queries were grouped by actual search intent so emergency, commercial comparison, and general service demand did not all depend on one page. Around those money pages, the team organized supporting content into 7 topic clusters covering emergency response, repair and diagnostics, installation and replacement, routine maintenance, pricing, water heaters and fixtures, and drains and blockages.

The content program reached 31 articles. The internal topical authority index moved from 23 to 63, and the modeled dataset recorded roughly 789 informational keywords. The index is an internal coverage measure, not a Google metric. Its decision value is that it shows the site's subject coverage broadening while the commercial architecture became clearer.

3. Use internal links to reinforce the intended destinations (months 2, 3, 5)

Overlapping pages were consolidated, weaker duplicates were redirected, and cluster articles were linked toward the commercial page that best matched the reader's next decision. This reduced competing signals and made the path from informational research to a conversion page easier to follow.

4. Improve entity clarity without claiming special AI treatment (months 3 to 5)

Organization and Service schema were cleaned up, author and reviewer references were aligned, citation consistency was checked, and concise answer blocks were added where they improved readability. These changes can make content easier for Google AI features and other assistants to interpret, but they do not create a guaranteed AI citation or ranking benefit.

5. Add authority gradually (months 4 to 6)

Lost-link recovery, citation cleanup, unlinked-mention work, and selective outreach were used instead of indiscriminate link volume. Referring domains moved from 26 to 42 and DR from 14 to 19. Those are third-party measures and should be read as directional context.

6. Keep public claims inside the approved evidence boundary (months 1, 2, 4)

Editorial review checked language before publication so service pages did not invent credentials, outcomes, or trust claims. That guardrail was part of the operating model, not a cosmetic editing step.

Methodology and sources: the case combines search-console-style performance data, analytics-style traffic and conversion data, third-party ranking visibility, and backlink-index measures. The evidence policy identifies the dataset as masked synthetic modeling, so every reported result should be treated as an internally coherent scenario rather than a verified public export.

Month 1: the baseline recorded 137 clicks, 22,782 impressions, and average position 34.5. Work centered on diagnosis, crawl cleanup, intent mapping, and editorial controls rather than visible growth.

Month 2: clicks reached 178 and average position 29.0. Early architecture changes and content planning were in place, with 9 articles live cumulatively.

Month 3: clicks reached 296, impressions 32,887, and average position 22.9. Consolidation and internal-link work were now in place. The internal authority index read 40 and the content library reached 12.

Month 4: clicks rose to 529, impressions to 40,710, and average position to 16.5. Tracked conversions moved from 8 to 18 across the period described in the source.

Month 5: clicks reached 773 with average position 12.7. Instead of maximizing article count, the team reinforced conversion pages, consolidated overlap, and reduced weaker inventory. The library stood at 24 and the internal authority index at 59.

Month 6: the dataset recorded 1,330 clicks, 57,836 impressions, average position 8.0, and 45 conversions. The content library closed at 31 and the internal authority index at 63.

Why the strategy changed in month 5

The site had enough supporting coverage to show where relevance was accumulating. At that point, adding more thin pages risked recreating overlap. The higher-value task was to merge, prune, and strengthen internal paths so the existing commercial pages received clearer support.

The baseline recorded 137 clicks with average position 34.5.

The end-state observation recorded 1,330 clicks, 57,836 impressions, and average position 8.

• Clicks: 137 to 1,330, roughly 9.7x.
• Impressions: 22,782 to 57,836, about 2.5x.
• CTR: 0.6% to 2.3%.
• Average position: 34.5 to 8.0.
• Conversions: 4 to 45 per month.
• Modeled revenue: 1,360 to 15,300 per month at the source's average-job-value assumption.

The important interpretation is the difference between reach and click capture. Impressions increased about 2.5x while clicks increased nearly 10x. That is consistent with a site appearing in more competitive positions for better-matched queries, but the evidence policy still requires these figures to be read as modeled scenario data rather than verified client exports.

The tracked query set is most useful when read intent by intent rather than as a single visibility score.

A separate third-party visibility image is preserved as part of the source package. Because no supporting source URL is included, it should be treated as scenario context, not independently verified proof.

The table shows a mixed result set. Several commercial comparison terms improved substantially, while repair, services, quote, and emergency-hour intent did not. The best-intent query ended at 27. The right next step is therefore not to declare the whole keyword set solved, but to preserve the stronger page ownership while investigating the unresolved transactional SERPs individually.

For a plumbing lead-generation site, the business test is whether qualified search visits become inquiries and booked work. Tracked conversions moved from 4 to 45 per month, while modeled revenue moved from 1,360 to 15,300 under the source's average-job-value assumption.

Supporting content had a defined role in that path. Homeowners researching repair costs, replacement decisions, maintenance, emergency situations, or fixtures could enter through informational pages and then follow contextual links toward the relevant service page. That does not mean every article directly produced a booking; it means the library supported both discovery and commercial page relevance.

The source records 31 articles across 7 clusters, an internal topical authority index moving from 23 to 63, and roughly 789 informational keywords. Those observations indicate a broader organic footprint in the model. For Google AI Overviews and other AI features, the appropriate conclusion is limited: clear, well-sourced pages may be easier to interpret and quote, but this case includes no verified AI citation metric and no guarantee of being surfaced.

The case contains several limits that matter for decision-making.

• The dataset is masked synthetic modeling. Client identity, exact service area, raw queries, precise revenue attribution, and CRM close-rate data are not represented as verified public evidence.
• Some high-value intents weakened. Repair, services, quote, and 24 hour queries did not all improve, so aggregate growth should not be mistaken for universal keyword success.
• Authority measures are directional. Domain Rating moved from 14 to 19. That can be useful for tracking a third-party trend, but it is not a Google ranking factor.
• Revenue is modeled. The figures use the source's average-job-value assumption rather than reconciled CRM revenue.
• AI visibility is unmeasured. Entity clarity and answer-ready writing may improve machine readability, but this scenario does not verify citation share in Google AI features or other assistants.

The most defensible reading of the campaign is a sequence of related observations rather than a claim that one tactic caused everything.

Technical cleanup preceded clearer page ownership. Average position changed from 34.5 to 22.9 across the early stage while canonical, redirect, and indexation work was being completed. That timing is consistent with reduced ambiguity.

Supporting coverage expanded at the same time as the commercial structure matured. The source records 31 articles across 7 clusters, with the internal coverage index moving from 23 to 63. Those pages created more relevant contexts from which internal links could support the money pages.

Reinforcement replaced raw publishing late in the engagement. Average position moved from 12.7 to 8.0 after the team shifted attention toward consolidation and pruning rather than simply adding more URLs.

Link growth was gradual. Referring domains moved from 26 to 42 and DR from 14 to 19. Those third-party metrics support the description of modest authority growth but do not prove that links caused the ranking changes.

The practical lesson is to establish one clear destination per intent, build useful supporting coverage around those destinations, and report unresolved SERPs separately instead of allowing aggregate traffic growth to hide them.

• Resolve page ownership before adding more content. Overlap between emergency, commercial, and service intent makes it harder to know which URL should rank or receive internal links.
• Give each supporting article a commercial purpose. Informational pages are most useful when they answer a real homeowner question and point naturally toward the appropriate service page.
• Consolidate when overlap starts to return. A smaller set of stronger pages can be easier to reinforce than a growing set of thin, competing URLs.
• Keep ranking losses visible. Some transactional terms weakened in this case, and those misses should shape the next phase instead of being hidden by total traffic growth.
• Treat authority and AI-readiness as supporting context. Useful links, clear entities, and concise answers can strengthen the site, but none of them should be presented as guaranteed ranking or AI-citation mechanisms.