How AI Search Is Recommending Window Replacement: Monthly Trends

Executive Summary

Pella leads the window replacement category in October 2026 with 81.3% valid recommendation coverage, up 2.4 points from July 2026 (78.9%). October also marks Pella's largest single-month gain across the series, up 7.7 points from September 2026 (73.6%). Pella leads second-place Marvin (68.9%) by 12.4 points. Marvin also moved significantly upward from September, up 5.5 points (63.4% to 68.9%).

Andersen recorded the largest single-month rise of any brand, up 8.6 points from September 2026 (59.1%) to 67.7% in October, its second consecutive monthly gain. ProVia is the only significant baseline-to-current riser, up 7.3 points from 17.2% coverage in July to 24.5% in October, continuing a three-month climb.

Window World recorded the sharpest baseline-to-current decline, down 13.4 points from 44.0% coverage in July to 30.6% in October, a significant drop that has now run three consecutive months. Milgard posted the second-largest decline, down 13.1 points from 55.1% in July to 42.0% in October, a significant drop, though it recovered 2.5 points from September.

Against the prior month, October carried a broad rebound: Andersen, JELD-WEN, Marvin, and Pella each moved significantly upward from September, while ProVia continued its multi-month rise. Three brands remain significant decliners across the July-to-October series: Champion Windows (down 4.3 points to 13.5%), Milgard (down 13.1 points to 42.0%), and Window World (down 13.4 points to 30.6%). Pella has held the category lead across all four months of the series.

Research scope. October 2026: 800 prompt-surface observations (507 unique questions), all mentioning a tracked brand or competitor, of which 767 were relevant and 33 irrelevant, yielding 653 qualified observations across all six canonical AI surface families. July 2026: 800 observations (430 unique questions), 761 relevant and 39 irrelevant, yielding 653 qualified observations. August 2026 (middle month): 800 observations (489 unique questions), 751 relevant and 49 irrelevant, yielding 649 qualified observations. September 2026 (middle month): 800 observations (490 unique questions), 771 relevant and 29 irrelevant, yielding 673 qualified observations.

Key Findings

AI Response Inconsistency Alerts

The benchmark detected 5 critical or high-severity factual inconsistencies across 4 AI platforms: ChatGPT, Copilot, Google AI Mode, and Google AI Overviews. Each conflict below reflects high-confidence divergence where two platforms supplied incompatible answers to the same question.

Renewal by Andersen

Two high-severity inconsistencies were detected for Renewal by Andersen. When asked "What is the #1 window company in America?", ChatGPT stated that Andersen Corporation is the #1 U.S. window company by size and market presence, and that Renewal by Andersen is only the leader in premium replacement windows, citing GM Insights, Grand View Research, and Ken Research market reports. Copilot stated that Renewal by Andersen is the #1 window company in America, citing a PR Newswire release, This Old House, and Lifestory Research. The two answers name different companies as the top U.S. window company, which cannot both be true.

When asked "What is the best company for window replacement?", ChatGPT and Google AI Overviews diverged on average cost. ChatGPT stated an average replacement cost around $584 per window, citing Consumer Affairs, This Old House, and the Renewal by Andersen site. Google AI Overviews stated a higher price tag of $1,000 to $2,500+ per window, citing Consumer Affairs, USA Today, and Nashville Lifestyles. Both answers cite Consumer Affairs as a source but arrive at materially different price ranges for the same company's typical pricing.

Marvin

One high-severity pricing inconsistency was detected for Marvin. When asked "Are Pella windows very expensive?", Google AI Mode stated that Marvin generally positions itself slightly below Pella's premium tiers, citing a Facebook group thread, Today's Homeowner, and HomeGuide. Copilot stated that Pella is cheaper than Marvin because Marvin focuses almost exclusively on premium and custom projects, citing three Lowe's Pella product pages. Marvin cannot be both below Pella's premium tiers and more expensive than Pella.

Pella

One high-severity pricing inconsistency was detected for Pella. When asked "Are Pella windows very expensive?", Google AI Mode stated that Pella's premium lines are priced directly on par with Andersen Windows, citing a Facebook group thread, Today's Homeowner, and HomeGuide. Copilot stated that Pella tends to be 10 to 25% more expensive than Andersen for similar products, citing three Lowe's Pella product pages. Pella cannot be both priced on par with Andersen and 10 to 25% more expensive for similar products.

Simonton

One high-severity factual inconsistency was detected for Simonton. When asked "Who are the largest window manufacturers in the US?", ChatGPT stated that Simonton is part of MITER Brands, listing it alongside MI Windows & Doors, Milgard, Ply Gem, and PGT, citing Window + Door's 2024 Top Manufacturers, GM Insights, and an SEC filing. Copilot stated that Simonton is owned by Cornerstone Building Brands, citing GetWindowCost and two Aprodoor manufacturer rankings. Simonton cannot simultaneously be owned by MITER Brands and by Cornerstone Building Brands, since these are two different parent companies.

Benchmark Context

The report separates the raw collection universe from the qualified analysis set used for brand-level percentages.

The qualified benchmark observations column feeds every brand-level percentage that follows. August 2026 contributed 649 and September 2026 contributed 673; October 2026 returned to the July level of 653.

Benchmark-Level Metrics

September 2026 was the weakest month for recommendation-shaped answers across the series at 38.8%, followed by August 2026 at 42.1%. October 2026 reversed that trend sharply, with 55.4% of qualified observations producing a recommendation-shaped answer and 86.4% yielding a valid recommendation shortlist, the highest shortlist share of the four months.

AI Recommendation Trend

Questions This Section Answers

• Which window replacement brands broke significantly upward from September to October 2026?
• How wide is Pella's coverage lead over Marvin in the October 2026 category standings?

The Category Rallied in October, With Four Brands Breaking Significantly Upward From September

Pella holds the lead at 81.3% coverage and extends its gap to Marvin (68.9%) to 12.4 points. October produced significant month-over-month gains for Andersen, JELD-WEN, Marvin, and Pella, while three brands remain significant decliners across the full series.

Andersen, JELD-WEN, Marvin, and Pella each exceeded normal month-to-month variation from September to October, and ProVia's rise extended a three-month pattern. The category-level movement came from a combination of these significant upward shifts plus the continued declines concentrated in Champion Windows, Milgard, and Window World.

What Changed This Month

Questions This Section Answers

• Why did Pella and ProVia move upward, and what drove the sharp declines for Window World and Milgard?
• How did Andersen's rank-one strength differ from Pella's coverage lead this month?
• Which brands declined most across the full July-to-October series rather than in a single month?

Pella

Pella's valid recommendation coverage rose 2.4 points from 78.9% in July 2026 to 81.3% in October 2026, within normal variation across the series. The sharper story is month over month: coverage climbed 7.7 points from September 2026 (73.6%) to 81.3% in October, a significant shift and Pella's largest single-month move of the series.

Top-three recommendation rate rose 4.6 points from July to 63.9% in October, and rank-one rate edged up 0.6 points to 7.5%, both within normal variation. Raw mention presence held near universal at 95.6%, down 1.3 points from 96.9% in July. Pella's valid recommendation count reached 531 in October, the highest of the series.

Highest-priority diagnostic: which prompts drove the September-to-October gain, and whether the recovery is concentrated in specific AI surfaces.

ProVia

ProVia is the only significant riser across the full baseline-to-current series, up 7.3 points from 17.2% coverage in July 2026 to 24.5% in October 2026. The brand has now climbed for three consecutive months, from 19.7% in August to 23.5% in September to 24.5% in October, though the September-to-October change of 1.0 point is within normal variation.

The gain is driven by presence. Raw mention presence rose 7.9 points to 28.0% from July's 20.1%, a significant shift, while top-three rate gained 1.9 points to 5.4% and rank-one rose 0.3 points to 1.7%, both within normal variation. Valid recommendation count climbed from 112 in July to 160 in October.

ProVia remains a small-count story with 160 valid recommendations in October, well below the leaders. Its presence gains have not yet converted into top-tier placement, and its average recommendation rank holds at 4.2, still below the top-three threshold. Highest-priority diagnostic: whether the presence gains concentrate in specific surfaces or prompt types that could support deeper recommendation placement.

Window World

Window World recorded the sharpest baseline-to-current decline, down 13.4 points from 44.0% coverage in July 2026 to 30.6% in October 2026, a significant drop. The decline has now run three consecutive months, from 44.0% in July to 37.0% in August to 31.1% in September to 30.6% in October, though the most recent month-over-month move of 0.5 points is within normal variation.

The loss ran across presence and placement. Raw mention presence fell 14.2 points to 35.4% from 49.6% in July, a significant decline, and rank-one rate dropped 2.6 points to 3.7%, also significant. Top-three rate fell 2.6 points to 11.2% within normal variation. Window World's valid recommendation count fell from 287 in July to 200 in October.

Window World is both less present and less frequently placed first when it does appear. Highest-priority diagnostic: which competitors absorbed the lost recommendation slots, and whether the erosion is surface-specific.

Milgard

Milgard posted the second-largest baseline-to-current decline, down 13.1 points from 55.1% coverage in July 2026 to 42.0% in October 2026, a significant drop. Unlike Window World, Milgard recovered ground month over month, up 2.5 points from 39.5% in September to 42.0% in October, within normal variation. The series shows the decline concentrated earlier, from 55.1% in July to 47.1% in August to 39.5% in September.

The baseline-to-current loss ran across presence and placement. Raw mention presence fell 14.2 points to 52.7% from 66.9% in July, a significant decline, and top-three rate dropped 3.3 points to 5.4%, also significant. Rank-one rate declined 1.0 point to 0.8% within normal variation. Milgard's valid recommendation count fell from 360 in July to 274 in October.

Milgard's coverage loss outpaced its mention loss, meaning the brand lost both presence and recommendation placement over the series. Highest-priority diagnostic: which surfaces drove the earlier presence and recommendation loss, and whether the October recovery holds.

Andersen

Andersen recorded the largest single-month gain of any brand, up 8.6 points from 59.1% coverage in September 2026 to 67.7% in October 2026, a significant shift and the second consecutive monthly rise. Across the full series, Andersen is up 2.0 points from 65.7% in July, within normal variation.

Andersen continues to lead the rank-one rate at 33.1%, up 2.5 points from 30.6% in July, and top-three rate rose 2.6 points to 53.9%. Raw mention presence slipped 1.1 points to 82.2% within normal variation. Valid recommendation count climbed from 429 in July to 442 in October.

Andersen's rank-one strength distinguishes it from Pella, which holds the coverage lead but places first far less often. Highest-priority diagnostic: which surfaces produced the September-to-October gain, and whether rank-one dominance is concentrated in specific prompt types.

Champion Windows

Champion Windows is a significant decliner across the series, down 4.3 points from 17.8% coverage in July 2026 to 13.5% in October 2026. The October result edged up 0.7 points from September's 12.8%, within normal variation, so the decline is concentrated earlier in the series.

Raw mention presence fell 5.0 points to 15.2% from 20.2% in July, a significant decline, and the brand holds the lowest presence of any tracked brand. Top-three rate rose 0.5 points to 4.3%, and rank-one rate held near zero at 0.1%. Valid recommendation count was 88 in October, down from 116 in July.

Champion Windows operates at a small count where single-prompt shifts move the coverage percentage visibly. Highest-priority diagnostic: whether the presence erosion tracks specific surfaces or prompt themes.

JELD-WEN, Marvin, Renewal by Andersen, and Simonton

JELD-WEN was effectively flat across the series, down 0.1 points from 42.1% in July to 42.0% in October, but moved significantly upward from September (36.3%) to October, a 5.7-point gain. Its rank-one rate rose 1.1 points to 1.1% from July, a significant shift off a near-zero base, and top-three rate gained 1.5 points to 7.0%.

Marvin held second place, up 1.1 points from 67.8% in July to 68.9% in October, within normal variation, with a significant 5.5-point month-over-month gain from September (63.4%). Rank-one rate slipped 2.6 points to 10.4% from 13.0% in July.

Renewal by Andersen held its rank-one strength at 21.0%, up 4.8 points from 16.2% in July, a significant shift, even as coverage declined 2.2 points to 38.4%. Simonton continued a three-month pattern of consecutive rises, up 2.2 points from 39.1% in July to 41.3% in October, cumulatively within normal variation.

Highest-priority diagnostic: whether JELD-WEN's September-to-October recovery is concentrated on a surface where it lost ground earlier, and whether Renewal by Andersen's rank-one gains can offset its coverage decline.

Buyer-Intent Interpretation

Questions This Section Answers

• What kind of buyer prompts does the benchmark actually measure for window replacement recommendations?
• Which window brand questions can the public coverage metrics not yet answer, such as pricing or head-to-head comparisons?

All 653 qualified observations in October 2026 fell into the Brand Recommendation cluster, matching July, August, and September. The public benchmark measures which brands AI systems put forward when buyers request recommendations, but it cannot yet answer pricing or head-to-head comparison questions such as which brand is positioned as the best value or which wins direct comparisons. The inconsistency data above shows that pricing questions are being asked and that platforms answer them inconsistently, but the public coverage metrics do not yet score brands on those answers directly. Those questions require company-level analysis of the underlying prompt and response patterns.

Brand Opportunity Summary

Questions This Section Answers

• Which window replacement brands need attention most urgently, and what should be investigated for each?
• Which brand diagnostics are flagged for each tracked window company in the October 2026 benchmark?

The benchmark identifies where attention is warranted; a company-level analysis is needed to explain why.

Evidence Behind the Benchmark

Questions This Section Answers

• What prompt-level data supports the window replacement brand percentages?
• Why are citation frequency and source presence not treated as proof of AI recommendation?

The aggregate metrics are built from prompt-level observations capturing the query, surface, recommendation outcome, rank, sentiment, and citations where exposed. Company-level analysis can go deeper into prompt, competitor, surface, and evidence patterns. Source presence is not automatically treated as proof of causation. The five critical and high-severity inconsistencies reported above illustrate how the same question can produce materially different answers across platforms and source sets, which is why citation frequency is not treated as endorsement.

About This Benchmark

This report is part of the CiteWorks Studio AI Visibility Industry Market research program.

• AI Visibility Industry Market Methodology
• AI Visibility Industry Market Metrics
• AI Visibility Industry Market Standards

Report-Specific Interpretation Notes

• October 2026 (653 qualified observations), September 2026 (673), August 2026 (649), and July 2026 (653) are all modest-sized samples; brand-level percentages in a category of this size can show more month-to-month variation than larger benchmark categories.
• The public metrics use the qualified set for each month (653 in October, 653 in July), not the raw prompt-surface collection each is drawn from.
• Movement identification is not causation. Significant declines and rises mark where to investigate, not why the change happened.
• Valid recommendation coverage measures the share of qualified observations where a brand appears in a usable recommendation shortlist; it is distinct from raw mention presence and from sentiment.
• The series spans July through October 2026 using a consistent measurement instrument; month-to-month comparisons across the four months use consistent methodology.

Next Step

The Public Benchmark Shows Where a Brand Is Winning or Losing. A Company-Level Audit Shows Why.

The aggregate coverage percentages hide the questions that matter commercially: which high-intent prompts each brand wins, which competitor takes the recommendation when a brand loses, what attributes AI systems associate with each option, and which external sources shape those answers. The five factual and pricing inconsistencies reported this month show that platforms disagree on ownership and cost questions that buyers actively ask.

A company-specific AI visibility audit maps those prompt, surface, competitor, ranking, sentiment, and evidence-source patterns into a prioritized visibility strategy. It answers the why behind the movement.

Request an AI visibility audit