Roofing Contractor SEO Case Study: 134 to 1528 Organic Clicks in 6 Months

The starting point was a roofing lead-generation site with about 134 non-branded clicks, an average position near 35, and weak commercial visibility. By the end of the modeled period, the same dataset recorded 1,528 clicks and 51 conversions per month, with modeled revenue moving from 1,360 to 17,340 under the source's average-job-value assumption.

The useful lesson is the sequence. Technical and indexation issues were addressed before supporting content scaled. Commercial intent was consolidated so fewer pages competed for the same job. Informational coverage was then built to support the money pages through clearer topical context and internal links. Later, production was slowed so weaker pages could be merged or pruned instead of allowing content volume to recreate the same ambiguity.

This case follows an anonymized roofing contractor site used for local lead generation through quote requests, calls, and emergency inquiries. The business operates in a competitive home-services market, so the search strategy had to prioritize commercially useful visibility rather than traffic for its own sake.

At kickoff, average position sat around 36. The source also records 29 referring domains and a domain rating of 15. That provided some existing authority, but the site still lacked a clear content architecture connecting early-stage homeowner research with the pages intended to generate leads.

The first decision was therefore architectural. Instead of increasing publishing immediately, the work focused on determining which page should own each commercial intent, which weaker pages should be consolidated, and where supporting content could answer real roofing questions without competing with conversion pages.

Three constraints governed the plan: established local competition, limited production capacity, and a strict evidence boundary for safety, warranty, and trust claims.

The source diagnosis also showed several commercial pages sitting deep in search results, including observed positions 55, 56, and 62 for repair, services, and quote intent. Those observations pointed to a larger problem than copy quality alone: page ownership was unclear, supporting informational coverage was thin, and internal authority was not concentrated on the pages most likely to generate leads.

The practical response was to avoid adding more overlapping pages. Technical cleanup and intent mapping came first so later content would reinforce a clearer structure rather than add another layer of competition.

The engagement used six connected workstreams. They were sequenced so that each stage produced a cleaner environment for the next.

1. Technical cleanup before scale (months 1 to 3)

Crawl and indexation triage came first, followed by canonical and redirect cleanup, template duplication review, renderability checks, Core Web Vitals review, internal status-code validation, and schema checks. The objective was to reduce unnecessary crawl noise and make the priority commercial URLs unambiguous before adding more supporting pages.

2. Rebuild information architecture and internal linking (months 2, 3, 5)

Pages competing for the same commercial purpose were reviewed and consolidated where appropriate. Internal links were then redirected toward the surviving money pages, with descriptive anchors and shorter paths from research content to quote-oriented pages.

3. Build authority content around real roofing decisions (months 2 to 4)

Target queries were grouped by search intent, then support content was organized around roof repair, replacement, storm damage, materials, maintenance, cost, financing, and contractor selection. The program reached 31 articles across 7 topic clusters and the modeled dataset recorded roughly 517 informational keywords.

The internal topical-authority index moved from 25 to 67. That index is an internal coverage measure rather than a Google metric, so it should be used comparatively: it shows that subject coverage expanded while the commercial pages were being reinforced with contextual internal links.

4. Clarify entities and answer surfaces (months 3 to 5)

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

5. Reinforce authority gradually (months 4 to 6)

Lost-link recovery, unlinked-mention work, citation cleanup, and selective resource outreach were used conservatively. Referring domains moved from 29 to 50 and domain rating from 15 to 20. Those third-party measures are directional context, not official Google signals.

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

Editorial QA used approved source material and a reviewer checklist to keep warranty, safety, and service claims within the client's stated evidence. This reduced the risk of using persuasive language that the source material could not support.

Month 1: the baseline recorded 134 clicks, 26,871 impressions, average position 35.58, and 4 conversions. Only 3 articles were published while technical diagnosis and page ownership were being established.

Month 2: clicks were 136 and impressions 22,743. The library stood at 6 articles and average position was 34.94. This stage remained focused on cleanup rather than visible growth.

Month 3: clicks reached 213, impressions 30,442, and average position moved from 34.9 to 24.7. The content library reached 11, and the internal topical-authority index read 43. Consolidation and internal-link changes were now in place.

Month 4: clicks reached 492, impressions 49,208, average position 15.08, and conversions 16. The library stood at 17 articles, giving the commercial pages a larger body of relevant supporting content.

Month 5: clicks were 541 and conversions 17, while average position moved from 15.08 to 15.96. The content library reached 24. Because the broad trend had flattened, the team shifted effort toward pruning weak pages, consolidating overlap, and reinforcing the pages already producing qualified demand.

Month 6: the modeled dataset recorded 1,528 clicks, 69,467 impressions, CTR 2.2%, average position 6, and 51 conversions. The library closed at 31 articles, the internal authority index at 67, and referring domains at 50. Because several workstreams matured together, the final jump should be treated as a combined observation rather than proof that one late action caused it.

Across the modeled period, clicks moved from 134 to 1,528, impressions from 26,871 to 69,467, conversions from 4 to 51, and modeled revenue from 1,360 to 17,340. Average position changed from 35.58 to 6, while CTR moved from 0.5% to 2.2%.

The baseline view shows a site appearing for relevant searches without capturing many clicks, which is consistent with weak average placement. The end-state view shows stronger placement and a larger share of clicks from the visible demand.

These figures belong to the masked synthetic scenario identified in the evidence policy. They are useful for understanding how ranking, click-through, and lead metrics can move together, but they should not be presented as independently verified client results.

The keyword table is most useful when read as a map of which intents responded to the consolidated structure and which did not.

The cost intent finished at 8, while installation moved from 28 to 61. The source associates that installation decline with consolidation, but the data does not prove a single cause. A safer next step is to re-evaluate whether installation deserves a distinct page with useful, non-duplicative content.

The two largest local terms remained difficult. Their recorded volumes were 4,400 and 2,400. The observed positions changed from 33 to 40 and from 42 to 41. Those are unresolved local-search opportunities. Any next phase should examine genuine location relevance, Google Business Profile completeness, and honest customer feedback practices without treating any undocumented activity as a guaranteed ranking factor.

The third-party visibility image is preserved as scenario context. With no supporting source URL in the JSON, it should not be described as independently verified. The modeled informational footprint ended near 517 ranking keywords.

The business measure in this case is qualified lead activity, not traffic by itself. Tracked conversions moved from 4 to 51 per month, while modeled revenue moved from 1,360 to 17,340 under the source's average-job-value assumption.

The supporting content also had a defined role in the lead path. The program included 31 articles across 7 clusters, and the internal topical-authority index moved from 25 to 67. Homeowners researching repairs, storm damage, materials, maintenance, costs, or contractor selection could enter through informational pages and then follow contextual links to the appropriate service page.

That does not prove each article caused a booking or that the resulting rankings will persist indefinitely. The more defensible conclusion is that the site finished the scenario with a broader base of useful roofing content and clearer internal paths into commercial pages. For Google AI Overviews and other AI features, the same clarity may improve machine readability, but the case contains no verified AI citation metric and makes no guarantee of inclusion.

This case is based on masked synthetic scenario data, not verified public client exports, and several unresolved areas remain important.

The largest local opportunities did not improve meaningfully, installation intent weakened, and cost intent remained volatile. The source also notes that the early cleanup period can look flat, but that timing should not be converted into a guaranteed waiting period. A single observation window of 30 days is not enough to establish a durable trend, and later performance can still change with competition, seasonality, site changes, and search-system updates.

Modeled revenue is based on average job value rather than reconciled CRM revenue. Third-party authority metrics are directional. AI visibility is unmeasured. The most useful way to read the case is therefore as a guide to sequencing and trade-offs, not as an outcome forecast.

The recorded growth from 134 to 1,528 clicks is best understood as a sequence of overlapping changes, not as proof that one tactic caused everything.

Technical and architectural work came first. Average position moved from 34.9 to 24.7 across the stage when consolidation and internal linking were being completed. That timing is consistent with clearer page ownership, but it is not a controlled experiment.

Supporting coverage broadened next. The source records 7 topic clusters and 31 articles, ending near 517 informational keywords, while the internal topical-authority index moved from 25 to 67. Those pages created more relevant contexts from which internal links could point toward commercial destinations.

Authority grew gradually. Referring domains moved from 29 to 50 and domain rating from 15 to 20. Those third-party metrics support the description of modest link-profile growth, but they do not prove that links produced the ranking gains.

Better average placement coincided with stronger click-through and lead volume. CTR moved from 0.5% to 2.2%, while tracked conversions moved from 4 to 51. The sensible interpretation is that stronger placement, better page-to-query fit, and a larger supporting content base all contributed to the modeled outcome.

• Resolve technical and intent ambiguity before adding more pages. A larger content library is only useful when the site already knows which commercial URL owns each important search intent.
• Use informational content to support real roofing decisions. Repair, replacement, storm damage, materials, maintenance, cost, and contractor-selection topics can create useful entry points and relevant internal links.
• Consolidate carefully. Removing overlap can clarify page ownership, but the installation decline in this case shows why every lost intent should be tracked after a merge.
• Do not hide local-search misses behind total traffic growth. The largest local terms remained unresolved and should shape the next phase.
• Treat third-party authority and AI-readiness as supporting context. Useful links, clear entities, and concise answer blocks can strengthen the site, but none should be framed as guaranteed ranking or citation mechanisms.
• Judge the program by qualified demand. Traffic is useful only when it brings the right homeowners closer to a quote, call, or service decision.