Reputation as a discipline is in the middle of three structural shifts. AI-generated review content is rising as a share of submitted reviews. Platform moderation is leaning harder on signal-based detection over user reports. Consumers are getting more cynical about perfect five-star averages. The numbers below show those shifts in motion and also show what stayed flat under all the noise.
How the 2025 vs 2026 Comparison Works
Both periods measure the same client cohort using the same dispute taxonomy. 2025 covers full-year disputes (n=78,902 in 2025; n=120,318 in the trailing 12 months ending March 2026, reflecting cohort growth). Where rates are reported, they are normalised against the dispute count. Where consumer numbers appear, both years used the same panel methodology with the same screening, weighted to the same demographic profile.
Same definitions, same methodology, two time points. The numbers that moved are real movement, not measurement drift.
The 12 Numbers That Moved
Each pair below is the same metric measured at two time points using the same definition. Movement larger than one point is real and traceable to a specific change in either platform behaviour, evidence-collection process, or consumer panel response.
- Network removal rate: 12.6% in 2025, 14.0% in 2026 (+1.4 points, driven by a stronger BRF-with-evidence mix)
- Business Redressal Form yield with two pieces of evidence: 44% in 2025, 51% in 2026
- Median time to outcome on the in-product flag: 6.1 days in 2025, 4.6 days in 2026
- Share of disputed reviews flagged as AI-generated: 7% in 2025, 14% in 2026
- Coordinated-attack share of network disputes: 4% in 2025, 9% in 2026
- Consumer trust in 5.0-star averages (panel score): 4.1 in 2025, 3.6 in 2026 (out of 10)
- Consumer trust in 4.7-star averages (panel score): 7.4 in 2025, 7.6 in 2026
- Average reviews read before service-business decision: 9.1 in 2025, 11.2 in 2026
- AI Overview appearance rate on informational local queries: 41% in 2025, 64% in 2026
- Map pack click share for position 1 on mobile: 47.8% in 2025, 44.2% in 2026 (compression)
- Share of disputes routed to legal channel due to defamatory factual claims: 3.1% in 2025, 4.0% in 2026
- Cost per removed review (BRF path): $52 in 2025, $48 in 2026 (efficiency gain on evidence templates)
Four Things That Conspicuously Did Not Change
Stable numbers are sometimes more interesting than the moving ones. These four held within a one-point margin across both windows.
- Share of buyers who would not consider a business under 3.7 stars: 57% (2025) and 58% (2026)
- Median analyst time per Business Redressal Form filing: 39 minutes (2025) and 38 minutes (2026)
- Removal rate for subjective complaints that are honest but harsh: 3% (both years)
- Removal rate by jurisdiction (US vs UK vs Australia): variance under 2 points across all three years
Consumer trust in 5.0-star averages dropped from 4.1 to 3.6 on a ten-point scale in 18 months. Trust in 4.7-star averages with visible negative reviews handled professionally rose.
What Actually Changed Under the Surface
The headline numbers tell one story. Underneath, three structural changes drove most of the movement.
First, evidence quality improved. Our analysts started attaching two pieces of evidence by default in mid-2025 because the internal data showed the BRF-with-two-evidence rate doubled the BRF-no-evidence rate. The shift accounts for most of the network-wide removal lift from 12.6% to 14.0%.
Second, AI-generated reviews became the fastest-growing category of inauthentic content. The 7% to 14% jump is not subtle. The detection signals (templated phrasing, generic praise, profile photo cluster matches with stock libraries) have improved at roughly the same rate, but the volume of inauthentic submissions is rising faster than enforcement.
Third, consumer behaviour around perfect ratings shifted. The decline in trust for 5.0-star averages from 4.1 to 3.6 on a 10-point scale is a meaningful drop. Buyers learned that perfect ratings are usually filtered, paid for, or unmaintained. The 4.7 with visible negative reviews handled professionally outperformed the 5.0 in every cell of the panel.
What This Means for 2026 Planning
If you ran your reputation plan against 2025 numbers, three things are now under-budgeted. AI-generated review detection needs more analyst time per case because the signals are subtler than templated text used to be. Coordinated attacks (ten or more reviews from new accounts inside 72 hours) are showing up more often and need a fast-flag protocol that does not exist in most agency playbooks. AI Overview presence is suppressing map pack click share on informational queries by enough percentage points to change media-mix decisions for any business that depends on local discovery.
Three things stayed the same and should keep their budget allocation: BRF-with-evidence as the workhorse channel, professional public response as the right answer for harsh-but-honest reviews, and a sustained review acquisition cadence as the floor below which rank decay is measurable.
