The strategic frame from the Phase 1/2/2b analysis, with tactical execution detail layered in. One coherent playbook for recovering non-paid Amex Platinum sales.
Amex Platinum non-paid sales fell ~20% year-over-year in the March–June window (≈1,475 → 1,175, −300 sales). The hard part: it's not a conversion problem. CVR is up. Rankings are up. Impressions are up 40%. Click-through has roughly halved because Google is absorbing clicks at the SERP — AI Overviews, zero-click features, and a structural re-classification of UP as a "derivative layer" rather than a primary source.
The plan recovers ~50 non-paid Amex Plat sales/month through two parallel arcs:
The 90-day work is high-confidence; the longer rebuild is necessary because site-quality recoveries materialise on subsequent core updates (typically 6+ months per Glenn Gabe's evidence base).
Impressions held or grew, clicks collapsed. The clicks are being absorbed at the SERP itself — AI Overviews, featured snippets, knowledge-panel answers. The biggest loser page is cited in 24 AI Overviews + 23 ChatGPT citations + Gemini, Perplexity, Grok. UP is feeding the very thing eating its traffic.
Example: /news/175k-sign-up-bonus/ — impressions +119%, clicks collapsed.
Google has shifted UP's positioning from "primary source" to "derivative aggregator." Authority hasn't changed — /100k-offer/ has 2,000 referring domains, DR 74, and lost 85% of sessions. /benefits/ has 15 referring domains and grew 198%. Links are not the lever; source-type fit is.
Cross-card pattern: same head-term collapse on Amex Gold, CSP, Venture X. UP holds zero top-20 presence for any flagship card's brand head term.
Top 10 decliners account for ~32,000 lost clicks. Two failure modes correspond to two refresh strategies — covered in pillars below.
| URL | Clicks lost | Pattern |
|---|---|---|
/100k-offer/ | −11,563 | Ranking + CTR collapse (−85% sessions) |
/news/175k-sign-up-bonus/ | −5,079 | CTR collapse (impressions +119%) |
/news/150k-welcome-offer/ | −3,335 | Stale offer |
/news/business-platinum-changes/ | −2,006 | Page faded |
/vs-visa-infinite/ | −1,905 | Ranking loss (impressions −96%) |
Three of the four biggest gainers are new pages. The established winner is /benefits/, which grew +198% by refreshing headline figures (not body content): 24 → 27 benefits, $4,316 → $7,600 value, 80K → 175K bonus.
Ran 12 representative Amex Plat queries through Google. UP appears in top 10 for 9 of 12 (75%) — better than expected. But three critical gaps:
| Query | UP in top 10? | Pos | Read |
|---|---|---|---|
| best amex cards 2026 | Yes | #6 + #8 | Two UP pages cited |
| amex card tiers explained | Yes | #1 | UP owns it |
| amex platinum 175k bonus | Yes | #1 | UP owns it |
| best way to redeem 175k Amex | Yes | #2 + #3 + #4 | Cited 3× |
| amex platinum oura ring credit | Yes | #2 | TPG #1 |
| amex platinum military benefits | Yes | #3 | Behind niche site |
| amex platinum welcome bonus | Yes | #2 | TPG won #1 |
| amex platinum benefits 2026 | Yes | #3 | Mixed citation share |
| is amex platinum worth it | Yes | #4 | Lost summary to fool.com, cnbc.com |
| centurion lounge access how-to | NO | — | GAP high-intent how-to |
| amex platinum vs gold | NO | — | GAP high-volume comparison |
| amex platinum vs CSR | NO | — | GAP existing page leans CSR |
The strategic shift: from "rank and capture clicks" to "become the cited, recommended, un-summarisable source."
Phase 1's data shows the click-capture model is structurally degrading. Rankings get better; clicks get worse. The defence is to make UP's content the kind of source AI summaries cite (not just consume) — original data, first-hand testing, proprietary calculators, named experts — and to surface that source-type signal everywhere Google looks.
That strategy expresses through four pillars, each with near-term tactical work and longer-term rebuild work:
Current state: UP cited for military, redemption, and 175k-bonus queries. Shut out for the three highest-value: benefits (60.5K/mo), "is it worth it," and centurion lounge access.
/credit-cards/reviews/american-express-platinum-card/centurion-lounge-access/ — engineered for "how to use centurion lounge" intent/credit-cards/amex-gold-vs-amex-platinum/ — UP not in top 10 today, high-volume comparisonCurrent state: UP absent from top 100 for "amex platinum" (201K/mo); #35 for "american express platinum card"; #50 for annual-fee terms; #95 for "is amex platinum worth it." This is the source-type re-classification problem.
/american-express/ as the commercial flagship. Republished April 2026, already 50K impressions/mo and 6 sales in 3 months. Add 175K bonus callout at top. Internal-link from every Amex sub-page, every comparison, every roundup./american-express-card-levels/ as a feeder. 152K lifetime sessions but 0.8% click-to-affiliate rate. Add top-of-page commercial banner linking to the flagship; add inline apply buttons to the tier table./american-express/:
/credit-cards/best-amex-cards-for-travel//credit-cards/best-amex-cards-for-cash-back//credit-cards/best-amex-business-cards//credit-cards/best-amex-cards-for-points//credit-cards/best-amex-cards-no-annual-fee/Current state: Site-wide CTR halved despite improved rankings. Risk that on-page quality signals (ad density, sponsored content, thin variants) are accelerating the source-type re-classification.
/100k-offer/ as the prime example)/credit-cards/reviews/american-express-platinum-card/sign-up-bonus/ as bonus-agnostic durable URL. Stage 301s of /100k-offer/ → /150k-welcome-offer/ → /175k-sign-up-bonus/ over 8 weeks. Each new bonus is a content edit, not a URL change.This is where most of the near-term +20–50 sales/mo comes from. Existing traffic on existing pages with conversion infrastructure that doesn't currently fire.
The Amex Plat benefit pages have 300K+ combined sessions and weak conversion:
| URL | Sessions | Lifetime sales |
|---|---|---|
/credit-cards/reviews/.../lululemon-credit | 99,930 | 9 |
/credit-cards/reviews/.../resy-dining-credit | 87,038 | 8 |
/credit-cards/reviews/.../oura-ring-credit | 76,537 | 8 |
/credit-cards/reviews/.../airport-lounge-access | 45,675 | 8 |
/credit-cards/reviews/.../equinox-credit | 30,300 | 4 |
Action: standard above-the-fold callout — "Get the Amex Platinum with this benefit (current bonus: 175K points →)" — linking to the canonical sign-up-bonus page. Template-level, one editor day.
+5–10 sales/mo
UP already has 10+ city-specific Centurion Lounge pages with 491K combined lifetime sessions but only 28 lifetime sales — 0.006% sales rate. Cards are clearly under-placed.
| City page | Lifetime sessions | Lifetime sales |
|---|---|---|
/amex-centurion-lounge-lax-los-angeles | 74,148 | 4 |
/amex-centurion-lounge-new-york-jfk | 68,449 | 5 |
/amex-centurion-lounge-san-francisco-sfo | 43,456 | 6 |
/amex-centurion-lounge-dfw | 51,893 | 1 |
/amex-centurion-lounge-miami | 49,012 | 3 |
/las-vegas-amex-centurion-lounge | 53,243 | 1 |
| (plus LGA, CLT, PHX, SEA) | — | — |
Action: template-level top-of-page Amex Plat callout box framing "Visit this lounge with the Amex Platinum" + the 175K bonus callout in the first H2 + cross-link to the national Centurion Lounges hub.
+8–15 sales/mo if conversion goes from 0.006% → even 0.05%
| Page | Action |
|---|---|
/credit-cards/amex-platinum-card-vs-chase-sapphire-reserve/ | Build new as Amex-Plat-favored sister to the existing CSR-favored page |
/credit-cards/amex-gold-vs-amex-platinum/ | Build / refresh canonical — UP not in top 10 for this high-volume query |
/amex-platinum-card-vs-marriott-bonvoy-brilliant/ | Refresh (CVR jumped 15% — double down) |
/amex-platinum-card-vs-hilton-honors-aspire/ | Refresh (lost traffic) |
/mastercard-black-card-vs-amex-platinum-card/ | Refresh (28% historical CVR niche page) |
/amex-platinum-vs-delta-platinum-detailed-comparison/ | Reposition to cover both Delta Platinum AND Delta Reserve |
+8–12 sales/mo
Six pages that had real Amex Plat sales in 2025 (4–8 each), zero in 2026:
/los-angeles-lax-airport-lounges (lost 4 sales)/american-express-centurion-lounges (national hub, lost 8 sales)/atlanta-atl-airport-lounges (1)/newark-liberty-ewr-airport-lounges (1)/boston-logan-bos-airport-lounges (1)/seattle-tacoma-sea-airport-lounges (1)Action: update with 2026 Centurion Lounge entry rules; add above-the-fold Amex Plat CTA; cross-link to canonical sign-up-bonus page.
+5–10 sales/mo
Lost 13 sales YoY. Only page where refresh-recommend assigned uplift_tier: medium. Concrete actions: add missing H2 "American Express Platinum Card Military Benefits"; add 2 PAA FAQs ("How to apply for Amex Platinum military", "Can military spouse get Amex Platinum"); expand additional-cards section.
+3–5 sales/mo
As affiliate clicks decline with AI absorption, hedge by adding fixed-fee / pay-per-citation placements. Not a near-term lever but defends revenue as the structural shift continues.
The Amex Plat news cluster has clear format-winners. Past pieces with double-digit lifetime sales:
| Past winner | Lifetime sales | Format |
|---|---|---|
/news/amex-platinum-up-to-175k-bonus | 133 | Live-offer announcement |
/news/apply-amex-platinum-before-2025-refresh | 26 | Urgency / deadline |
/news/redeem-amex-platinum-175k-bonus-first-class | 11 | Specific redemption |
/news/have-both-amex-platinum-delta-reserve-cards | 9 | Pairing logic |
/news/apply-amex-platinum-end-of-year | 9 | Urgency / seasonal |
/news/apply-amex-platinum-before-summer | 8 | Urgency / seasonal |
/news/amex-platinum-2025-major-refresh-details | 8 | Benefit-change explainer |
/news/amex-platinum-best-luxury-travel-card | 8 | Curated list / opinion |
/sign-up-bonus/ page (no redirects yet)/american-express-card-levels//american-express//american-express/ flagship from all Amex sub-pages/100k-offer/ → canonical sign-up-bonus page/amex-gold-vs-amex-platinum/ canonical/amex-platinum-vs-chase-sapphire-reserve/ Amex-favored sister/150k-welcome-offer/ → canonical/best-amex-cards-for-travel/, /best-amex-business-cards/)/175k-sign-up-bonus/ → canonical (final cutover, only after new URL is conclusively stronger)Aleyda Solís's 3-layer model. The rule: if a metric can't change a decision, remove it from the dashboard.
KPIs: prompt coverage, recommendation rate, linked-citation rate, comparative-win-rate
Method: 50–250 prompts tested 5–7× each, tracked at topic level. Each AI platform (ChatGPT, Perplexity, Gemini, Claude) tracked separately — they don't move together.
KPIs: diagnostic score against 10 winning-brand traits — accessible, extractable, useful, fresh, differentiated, recognisable, consistent, corroborated, credible, transactable.
Prioritisation formula: (visibility-gap impact × commercial importance) ÷ effort
Four confidence tiers:
Reporting reframe: lead with citation/recommendation share and branded-search lift. Clicks and sales beneath — they're trailing indicators now.
Honest bottoms-up math. Near-term work compounds with the rebuild over 6+ months.
| Lever | Pillar | Sales / mo |
|---|---|---|
| Offer-page consolidation (301 to canonical sign-up-bonus) | 3 | +10–20 |
| CTR fixes on retained-impressions pages (175k page, benefits page, etc.) | 4 | +5–15 |
| On-page conversion fixes (benefit sub-page CTAs, Centurion city CTAs) | 4 | +5–15 |
| Military benefits refresh | 4 | +3–5 |
| Airport-lounge guide refreshes | 4 | +5–10 |
| Comparison page refreshes + new CSR / Gold pages | 4 | +8–12 |
| New news features at 3/mo cadence | 4 | +15–25 |
| Near-term subtotal (achievable within 90 days) | +20–50 | |
| Lever | Pillar | Sales / mo |
|---|---|---|
| AI citation share recovery (3 un-summarisable pages, points index) | 1 | +10–20 |
| Head-term reclamation (category hubs, flagship internal-link rebuild) | 2 | +10–20 |
| Centurion how-to + Gold-vs-Plat + worth-it canonical pages | 1+4 | +5–15 |
| Brand-demand program (PR, original-data hits) | 2 | +5–15 |
| Video layer | 1+2 | +3–10 |
| Rebuild subtotal (6+ months out) | +30–80 | |
| Target | +50 / mo |
| Near-term range | +20–50 / mo |
| Rebuild range (cumulative) | +50–130 / mo |
| Risk | Mitigation |
|---|---|
301 cutover from /175k/ → canonical loses momentum during transition | Stage it. Only flip after the canonical is conclusively stronger. Monitor sales daily for first 14 days post-cutover. |
| Amex changes the bonus mid-build | The canonical URL is bonus-agnostic by design — handle as content edit, not URL change. |
| Sub-hub category pages cannibalise the flagship | Clear internal-link hierarchy (subs → flagship, never reverse). Watch for impressions migrating between them. |
| News feature CVRs don't replicate past winners | Run formats with smallest editorial commitment first. If a format misses, drop it. |
| Per-card attribution isn't available at URL level | Spec the one-off Hex query that breaks card_sales by card_id × url × date. Biggest measurement gap. |
| AI Overview citation share decays as Google's algorithm shifts | Repeat the AEO benchmark quarterly. Reprioritise page work based on which queries are most at risk. |
| Site-quality issues take 6+ months to recover (Gabe) | Don't promise the long arc until the near-term sprint is shipping. Defensive work in Sprint 1 starts the clock; rebuild work in Q2-Q3 lands the recovery on a subsequent core update. |
| Editorial bandwidth doesn't sustain 3 news features / month | Start with the urgency format (lowest creative cost). Use the past-winner formats as scaffolding so writers aren't designing each piece from scratch. |