What changed in Division50's Google visibility?
A transparent comparison of two consecutive 28-day periods. Clicks increased while low-value impressions fell, but Dubai's non-brand commercial visibility remained largely unproven.
Data through 26 August 202656 days comparedDivision50's own domain
Executive Summary
Useful visits increased. Google Search clicks rose from 378 to 463, while click-through rate moved from 2.6% to 4.7%.
The impression decline was concentrated. Total impressions fell by 33.1%, with legacy, weakly matched city pages accounting for most of the largest page-level losses.
Evidence content gained. The B2B lead generation statistics resource added 1,030 impressions, the largest useful gain in the reviewed page cohort.
Dubai is still the measurement gap. The available Dubai query set was overwhelmingly branded. This baseline does not prove ranking for non-brand agency terms.
Download the study data
The aggregate observations behind this page are available in machine-readable JSON and CSV. They contain no person-level or client data. Reuse is permitted under CC BY 4.0 with attribution to Division50 and a link to this study; the limitations below travel with the data.
A smaller result set produced more clicks
The latest period ran from 30 July to 26 August 2026. The comparison period ran from 2 July to 29 July 2026. Search Console's average position is an impression-weighted aggregate, so it is not a fixed rank and should not be read in isolation.
Interpretation: the 9.3-position improvement does not mean every page moved up nine places. Thousands of low-position impressions disappeared, changing the mix used to calculate the aggregate.
Research and core service pages carried the useful gains
These are selected page-level observations, not a complete export. The evidence asset is the clearest authority opportunity because it grew against non-brand research queries and already earns external citations.
| Page | Latest impressions | Previous | Change | Latest position |
|---|---|---|---|---|
| /resources/b2b-lead-generation-statistics | 2,097 | 1,067 | +1030 | 15.4 |
| / | 1,843 | 1,602 | +241 | 6.5 |
| /services/lead-generation-usa | 323 | 127 | +196 | 15.1 |
| /contact | 773 | 578 | +195 | 4.8 |
| /services/outbound | 568 | 415 | +153 | 8.9 |
| /services/ai-automation | 171 | 30 | +141 | 12.0 |
Legacy city pages explain much of the lost reach
The largest visible losses came from pages outside the focused market strategy, including Edmonton, Pretoria, Nagoya and Hamilton. Losing weakly matched reach is not automatically a business loss. The exception below is the sales-strategy page, whose decline was mainly a redistribution of branded demand to the homepage, not a reason to delete it.
| Page | Latest impressions | Previous | Change | Latest position |
|---|---|---|---|---|
| /services/seo/edmonton | 7 | 2,943 | -2936 | 80.3 |
| /sales-strategy-consulting/ | 32 | 622 | -590 | 3.8 |
| /services/seo/pretoria | 4 | 330 | -326 | 17.3 |
| /services/content/nagoya | 4 | 272 | -268 | 6.3 |
| /services/seo/hamilton | 0 | 268 | -268 | Not available |
Dubai demand exists, but non-brand agency visibility is not established
What the filtered data showed
A query filter containing “Dubai” produced 43 clicks and 128 impressions. Most were branded searches for Division50. The visible non-brand observations were limited:
- ai automation agency dubai: 11 impressions, average position 18.9
- ai automation agency in dubai: 2 impressions, average position 9.0
- paid ads dubai: 1 impression, average position 6.0
The reviewed local commercial branch was not deployed during this measurement window. Our next UAE comparison starts only after the reviewed UAE and Dubai SEO owner is released with explicit approval. Until then, this page is a baseline, not a ranking case study.
Google reported generative-AI visibility, not universal LLM citations
Google's generative-AI-features report recorded 867 impressions in the latest 28 days. The 90-day view recorded 2,269 impressions across 360 pages. This does not count ChatGPT citations, Perplexity mentions, Copilot answers or leads attributed to those systems.
| Page | Google generative-AI impressions |
|---|---|
| / | 430 |
| /blog/outbound-agency-vs-sdr-2026-cost-comparison | 59 |
| /resources/b2b-lead-generation-statistics | 59 |
| /services/outbound | 58 |
| /contact | 28 |
| /how-pricing-works | 19 |
GEO/AEO implication:original evidence, clear comparisons and stable service definitions are already receiving exposure in Google's AI surfaces. Google states that standard SEO practices remain the foundation for AI features; there is no special AI schema or guaranteed citation switch.
What we will measure next
- Release only the reviewed Dubai and UAE commercial owners after approval, then record the deployment date as the cohort boundary.
- Submit the truthful sitemap and inspect those exact URLs, without bulk-indexing noindex or unreviewed matrix pages.
- Track non-brand impressions and clicks by query family, then join qualified visits to CRM outcomes without treating rankings as revenue.
- Measure Google AI visibility separately from ChatGPT, Bing Copilot, Perplexity and referral traffic. A zero without provider coverage remains “not measured”, not “no citations”.
Methodology, sources and limitations
Method
Division50 reviewed the authenticated Search Console property for two consecutive 28-day periods through 26 August 2026. We compared property totals, selected page rows, filtered query cohorts and Google's separate generative-AI report. The figures are observational and rounded as displayed by Search Console.
Important limitations
- Filtered query tables can omit anonymized or low-volume queries.
- Property and page tables use different aggregation rules.
- Average position is not a fixed rank.
- Changes are not proof of causation or a client outcome.
- The undeployed local page matrix is excluded from the measured period.
Questions about this study
Did Division50's Google visibility improve?
The latest 28-day period was mixed. Clicks and click-through rate increased, and reported average position improved, while total impressions declined. Most of the largest impression losses came from legacy city pages. Because the set of impressions changed, average position cannot be read as every page moving up by the same amount.
Does this prove Division50 ranks for lead generation in Dubai?
No. The visible Dubai query cohort was dominated by branded searches, and the reviewed data contained almost no non-brand commercial visibility for Dubai lead-generation or marketing-agency terms. The new local pages were not live during this baseline.
Are Google generative-AI impressions the same as ChatGPT citations?
No. They describe visibility within Google's generative AI search features. They do not measure ChatGPT, Perplexity, Copilot, Gemini citations outside Google Search, or leads attributed to those systems.
Is this a client case study?
No. It is an observational study of Division50's own domain. It does not report a client outcome, claim causation, or guarantee that another website would produce the same result.