Quwa Case Studies
Quwa is a defence and aerospace publication that Sun Talon owns and operates. It runs as one system: specialist research, daily production, distribution through search and Google Discover, and measurement that sets the next assignment.
The figures on this page cover August 2025 to July 2026. They come from Google Analytics 4 and Google Search Console, and every one of them can be reproduced from those accounts.
Editorial flywheel
One system, six stages, run without a break. Timely coverage earns distribution while a story is live. Durable coverage holds demand after the cycle ends.
Performance sets the next assignment
Signals
Search and Discover performance data enter the same queue, so reader demand sits next to market events rather than in a separate report. Every signal is logged with a date, a source and a note on why a specialist reader would care.
Quwa covers defence and aerospace procurement for specialist readers. Primary coverage is Pakistan, the Gulf and Türkiye, extending to supplier markets in Europe, China and North America. Readers are service personnel, industry staff, analysts and officials who follow named programs.
The audience is where the coverage is. Over the verified period Pakistan supplied 52% of Discover clicks, India 14%, the United Kingdom 7%, and the Gulf states 7% between them. Canada, the United States, Bangladesh and Malaysia each supplied 2% to 4%.
Country shares show where Discover clicks landed. Brand queries show demand for Quwa by name, not unique repeat readers.
≈51,466 of 166,406 Search clicks were brand queries.
2,362 more clicks came from queries for the subscriber login page.Source: Google Search Console, retrieved 9 September 2026. Country values are rounded; “Other markets” is the calculated remainder.
Quwa began in 2015 as a personal defence blog. The growth reported here came from running it as a publication — a defined territory, dependable sourcing, daily production and active distribution. The system had to absorb fast-moving events without dropping the sourcing standard or the publishing rate.
Two constraints shaped every decision. The publication works from public sources only, so speed can never come from an unverified claim. It also carries a two-tier subscription, so the free layer has to earn reach while the paid layers hold the depth.
Cover the decisions, programs, companies and capabilities that specialist readers already follow. Depth in one market outperforms coverage spread across several. The test is brand demand: over the period, queries containing the publication name drew about 31% of all search clicks.
Timely pieces earn Discover distribution while a story is live. Durable pieces answer the questions that keep returning — capability, supplier, cost, program status. The same research usually produces both.
Every claim ties to a public source or a stated editorial judgment. Specification tables, contract values and delivery dates carry attribution. This is slower per piece, and readers in this market verify claims themselves.
Search and Discover do different jobs, and Quwa plans them separately.
Search handles durable demand. Over the period the site drew 166,000 search clicks from 13.0 million impressions at an average position of 7.3. The work is technical and slow to move: schema, news sitemap, internal links mapped from query data, index coverage.
Discover handles timely demand. Over the period it delivered 1.33 million clicks from 31.0 million impressions, a 4.3% click rate. Discover is earned distribution rather than an owned channel, and Google can change how it ranks and displays any publisher without notice.
The two are not substitutes. Discover moves fast in both directions, and search moves slowly and holds. Running both means a change in one does not take the audience with it.
The two Search Console surfaces are planned and measured separately across the same verified period.
1,331,093clicks
166,406clicks
Source: Google Search Console, 1 August 2025 to 31 July 2026. The 8.0× comparison is calculated from 1,331,093 and 166,406 clicks.
Google shipped a Discover core update in February 2026. The effect on Quwa is measurable and is reported here in full.
In February, Quwa drew 111,918 Discover clicks from 1.67 million impressions, a 6.7% click rate. In March, impressions fell 8% to 1.53 million while clicks fell to 42,666 and the click rate fell to 2.8%. Google kept serving the pages, and readers stopped clicking them at the previous rate.
The response was headline and image work rather than more volume. Headlines were rebuilt to a fixed pattern — named entity, active verb, one specific detail, stated stake — and the pattern was written down so every piece follows it. Image production moved to a single specification.
By June and July, Discover delivered 219,026 clicks from 5.37 million impressions: an average of 109,500 clicks a month at a 4.1% click rate. Monthly clicks came back to within 2% of the pre-update level. The click rate did not fully recover, and most of the regained volume came from a 61% rise in monthly impressions rather than from restored click-through.
The click rate did not fully recover. Most regained volume came from broader distribution after headline and image changes.
Source: Google Search Console. June–July values are monthly averages derived from the verified two-month totals of 219,026 clicks and 5,373,372 impressions.
The publishing volume holds because the routine work is automated and the judgment is not.
Quwa runs a fleet of more than 20 internal servers connecting the editorial workflow to the systems it depends on: WordPress publishing, Google Analytics 4, Search Console, keyword data, subscriber records and email. An editor can pull performance data, draft against it, publish with the correct taxonomy and schema, and check the result without leaving one workspace.
Measurement runs through a BigQuery pipeline holding session, search and subscriber data in one place. A daily briefing surfaces movement across programs, threads and pages, and that briefing sets the assignment queue.
Sourcing, editorial judgment and every published claim stay with the editor. Automation covers retrieval, formatting, publishing steps and measurement.
Each pair compares the verified period with the same twelve months a year earlier.
Source: GA4 property 397087941. Each bar pair is normalized to its verified-period value; exact totals remain visible.
| Measure | Verified period | Prior period | Change |
|---|---|---|---|
| Sessions | 1,612,299 | 889,267 | +81% |
| Readers (total users) | 1,002,663 | 487,467 | +106% |
| Pageviews | 2,010,235 | 1,338,972 | +50% |
| Channel | Volume | Detail |
|---|---|---|
| Google Discover | 1,331,093 clicks | 31.0M impressions, 4.3% CTR |
| Google Search | 166,406 clicks | 13.0M impressions, 1.3% CTR, avg. position 7.3 |
| Direct-attributed sessions | 899,657 | 55% of sessions — see below |
Sources: GA4 property, 1 August 2025 to 31 July 2026, against the same range a year earlier. Search Console, same range.
GA4 attributed 899,657 sessions, or 55%, to Direct. Direct is a residual category. It collects typed visits and bookmarks, and also app referrals, untagged links, and any visit arriving without referrer data.
Discover accounts for much of it. Search Console recorded 1.33 million Discover clicks over the period while GA4 recorded 684,000 Organic Search sessions. A large share of Discover traffic lands in Direct rather than Organic Search, so the Direct figure should not be read as a loyalty measure.
Brand search is the better repeat-audience signal. Over the period, queries containing the publication name drew roughly 51,000 search clicks, about 31% of all search clicks. Queries for the subscriber login page alone drew 2,362 clicks.
Sessions are visits, not people. GA4 recorded 1.0 million users against 1.6 million sessions over the same period.
Discover is earned distribution and it moves. One month, May 2026, supplied 294,484 sessions, and 280,092 Discover clicks landed in that same month. Excluding May, the other eleven months averaged about 120,000 sessions.
Direct is an attribution category rather than a repeat-audience measure. GA4 and Search Console are separate systems that count different events, and their totals do not reconcile exactly.
This records what one publication did, in one market, over one period. It is evidence of operating capability. It is not a forecast for a client.
Operating Quwa gives Sun Talon working knowledge of defence and aerospace subject matter, public-source verification, editorial judgment under a daily production load, and distribution through both search and Discover. It also produced the systems themselves — the production process, the technical stack and the measurement path.
Client work differs in one respect. A supplier's claims need the supplier's own evidence and approvals, and the review cycle sits with them. The method transfers; the public-source shortcut does not.
Content assessment Direct recommendation
Share the market, the current content estate and the growth goal. Sun Talon will identify whether strategy, production or distribution is the strongest place to start.
