Pension-Course Campaign & Landing Performance
Meta paid acquisition • 2026-05-13 → 2026-06-10 • generated 2026-06-11 10:00
Snapshot 2026-06-11 10:00 Asia/Jerusalemauto-refreshes hourly · next in …
Executive Summary
Real Facebook-ad leads (all-time)
1matches Meta (1) ✓
spend ₪1,083 · ad CPL ₪1,083 · paid 0 · ad ROAS 0.00× · net ₪-1,083
Definition (reproduce manually): a real Facebook-ad lead = a
pension_course_leads row that survives the test/dedupe filter and carries a Facebook source. Exact query — copy/paste to verify:SELECT * FROM pension_course_leads
WHERE superseded_by IS NULL
AND lower(utm_source) IN ('fb','facebook','meta','instagram','ig','facebook_paid')
AND lower(coalesce(email,'')) NOT LIKE '%@campaigner.co.il'
AND lower(coalesce(email,'')) NOT LIKE '%@goola-group.com'
AND coalesce(utm_campaign,'') NOT ILIKE '%smoketest%'
ORDER BY created_at;Course context (all real registrations, all-time — NOT the Facebook ad)
Real registrations
26
all sources, tests/dupes excluded
Real paid
13
50.00% of real
Course revenue
₪2,340
all real paid × price
of which from FB ad
1
0 paid
Tests excluded
2
your own QA submissions
Dupes excluded
10
superseded dedupe rows
Total DB rows
38
pension_course_leads, all-time
Meta Spend
₪1,083
window 2026-05-13→2026-06-10
Unprofitable: ROAS 0.00× — revenue ₪0.00 vs spend ₪1,083.
Ad destination misrouting — 299 paid-social sessions landed on the generic
/l/pension-course vs only 0 on the purpose-built /l/pension-course-fb. The Meta ad set's destination URL is pointing at the wrong page; the variant you built for this campaign is starved. Repoint the ad creative's link to /l/pension-course-fb.5 real lead(s), 0 paid in this window (tests excluded) — leads but no payments yet. The leak is lead→checkout→paid: review reminder timing/copy and the checkout page, not the ad.
Source availability: Meta GA4 Supabase
Meta Campaign Layer (paid acquisition)
Campaign
קמפיין קורס פנסיוני
Status
PAUSED
OUTCOME_LEADS
Daily budget
—
campaign/adset
Spend
₪1,083
window
Impressions
9,707
Reach
5,096
unique
Frequency
1.9
avg views
CTR (all)
5.61%
Link CTR
3.70%
link clicks
CPC
₪1.99
CPM
₪111.57
per 1K
Leads
1
pixel/CAPI
CPL
₪1,083
Ad sets
| Ad set | Status | Type | Spend | Impr | CTR | CPC | Leads |
|---|---|---|---|---|---|---|---|
| IL 40-65 Lead Optim — Multi-aspect (regular) | CAMPAIGN_PAUSED | — | ₪1,083 | 9,707 | 5.61% | ₪1.99 | 1 |
Per-creative comparison & fatigue (first 7d → last 7d CTR)
| Ad | Status | Spend | Impr | Freq | CTR | CTR 1st→last | Δ | Leads | Verdict |
|---|---|---|---|---|---|---|---|---|---|
| Main-A multi-aspect | CAMPAIGN_PAUSED | ₪374.60 | 3,310 | 1.3 | 6.10% | 6.06%→0.00% | -100.0% | 1 | MONITOR |
| Medium-A multi-aspect | CAMPAIGN_PAUSED | ₪271.15 | 1,700 | 1.2 | 4.71% | 4.68%→0.00% | -100.0% | 0 | MONITOR |
| Short-A multi-aspect | CAMPAIGN_PAUSED | ₪259.28 | 2,925 | 1.4 | 7.11% | 7.11%→0.00% | -100.0% | 0 | MONITOR |
| תמונה 1 | CAMPAIGN_PAUSED | ₪89.42 | 934 | 1.8 | 3.32% | 3.30%→0.00% | -100.0% | 0 | MONITOR |
| תמונה 3 | CAMPAIGN_PAUSED | ₪66.58 | 633 | 1.6 | 3.00% | 3.01%→0.00% | -100.0% | 0 | MONITOR |
| תמונה 2 | CAMPAIGN_PAUSED | ₪21.94 | 205 | 1.5 | 2.44% | 2.44%→0.00% | -100.0% | 0 | MONITOR |
Leads use the deduplicated
lead action_type (fallback: sum of offsite/onsite variants). RETIRE = CTR collapsed >40% with frequency ≥2.5.Daily trend
| Metric | Trend | First day | Last day | Total |
|---|---|---|---|---|
| Spend | ₪136.59 | ₪57.82 | ₪1,083 | |
| Impressions | 1,702 | 224 | 9,707 | |
| Link clicks | 149 | 8 | 359 | |
| Leads | 0 | 0 | 1 |
Placement breakdown
| Platform | Position | Spend | CTR | Link clicks | Leads |
|---|---|---|---|---|---|
| Facebook Reels | ₪423.39 | 4.13% | 122 | 1 | |
| Feed | ₪393.33 | 6.14% | 79 | 0 | |
| Instagram Reels | ₪101.80 | 2.69% | 17 | 0 | |
| Facebook Stories | ₪48.25 | 3.23% | 5 | 0 | |
| Instagram Stories | ₪32.06 | 4.07% | 6 | 0 | |
| Feed | ₪28.45 | 3.86% | 6 | 0 | |
| Instream Video | ₪24.99 | 4.76% | 9 | 0 | |
| Audience_Network | An Classic | ₪14.28 | 15.95% | 42 | 0 |
| Audience_Network | Rewarded Video | ₪10.40 | 41.45% | 70 | 0 |
| Right Hand Column | ₪3.01 | 0.00% | 0 | 0 | |
| Facebook Reels Overlay | ₪2.27 | 4.23% | 3 | 0 | |
| Marketplace | ₪0.43 | 0.00% | 0 | 0 | |
| Facebook Notification | ₪0.24 | 0.00% | 0 | 0 | |
| Threads | Threads Feed | ₪0.06 | 25.00% | 0 | 0 |
Age × gender (top 14 by spend)
| Age | Gender | Spend | CTR | Leads |
|---|---|---|---|---|
| 65+ | Male | ₪257.15 | 5.75% | 0 |
| 55-64 | Male | ₪244.55 | 5.11% | 1 |
| 55-64 | Female | ₪159.33 | 6.28% | 0 |
| 45-54 | Male | ₪135.25 | 5.47% | 0 |
| 65+ | Female | ₪107.22 | 8.31% | 0 |
| 45-54 | Female | ₪74.26 | 6.89% | 0 |
| 35-44 | Male | ₪47.59 | 3.77% | 0 |
| 25-34 | Male | ₪23.33 | 4.53% | 0 |
| 35-44 | Female | ₪16.02 | 5.94% | 0 |
| 25-34 | Female | ₪7.02 | 0.00% | 0 |
| 55-64 | Unknown | ₪5.28 | 0.00% | 0 |
| 45-54 | Unknown | ₪2.84 | 0.00% | 0 |
| 65+ | Unknown | ₪1.93 | 7.69% | 0 |
| 35-44 | Unknown | ₪1.14 | 0.00% | 0 |
Meta — Deep KPIs (efficiency · video · delivery)
Click efficiency & engagement
Outbound clicks
359
Outbound CTR
3.70%
Unique clicks
417
Unique CTR
8.18%
of people
Unique link CTR
5.79%
Cost / unique click
₪2.60
Cost / link click
₪3.02
Cost / outbound
₪3.02
CPP
₪212.51
per 1k reached
Landing-page views
247
LP-view rate
68.80%
LPV ÷ link clicks
Click→LPV drop
31.20%
JS/bounce loss
Video performance
Video plays
7,532
ThruPlays
1,051
≥15s / 97%
Cost / ThruPlay
₪1.03
25% / 50%
1,293 / 698
75% / 95%
322 / 193
100% watched
184
Hold rate
2.44%
100% ÷ plays
Avg watch (s)
7.0
Delivery diagnostics (Meta relevance rankings)
| Ad | Impr | Spend | Quality | Engagement | Conv-rate |
|---|---|---|---|---|---|
| Main-A multi-aspect | 3,310 | ₪374.60 | UNKNOWN | UNKNOWN | UNKNOWN |
| Medium-A multi-aspect | 1,700 | ₪271.15 | UNKNOWN | UNKNOWN | UNKNOWN |
| Short-A multi-aspect | 2,925 | ₪259.28 | UNKNOWN | UNKNOWN | UNKNOWN |
| תמונה 1 | 934 | ₪89.42 | UNKNOWN | UNKNOWN | UNKNOWN |
| תמונה 3 | 633 | ₪66.58 | UNKNOWN | UNKNOWN | UNKNOWN |
| תמונה 2 | 205 | ₪21.94 | UNKNOWN | UNKNOWN | UNKNOWN |
Rankings are vs other advertisers competing for the same audience. 'below average' on quality/engagement is the lever to raise CTR & lower CPL.
Full action breakdown (everything Meta recorded)
| Action type | Count | |
|---|---|---|
| page_engagement | 3,586 | |
| post_engagement | 3,583 | |
| video_view | 3,179 | |
| link_click | 359 | |
| landing_page_view | 247 | |
| omni_landing_page_view | 247 | |
| post_interaction_net | 48 | |
| post_interaction_gross | 45 | |
| post_reaction | 41 | |
| onsite_conversion.post_net_like | 38 | |
| offsite_content_view_add_20_s_calls | 8 | |
| onsite_web_view_content | 8 | |
| onsite_web_app_view_content | 8 | |
| view_content | 8 | |
| offsite_content_view_add_meta_leads | 8 | |
| omni_view_content | 8 | |
| offsite_conversion.fb_pixel_view_content | 8 | |
| onsite_conversion.post_unlike | 3 | |
| like | 3 | |
| onsite_conversion.post_save | 2 | |
| onsite_conversion.post_net_save | 2 | |
| onsite_web_lead | 1 | |
| comment | 1 | |
| post | 1 | |
| onsite_conversion.post_net_comment | 1 |
Hour-of-day (ad-account TZ)
| Hour | Spend | Impr | CTR | Leads |
|---|---|---|---|---|
| 00:00 | ₪21.86 | 263 | 4.94% | 0 |
| 01:00 | ₪17.63 | 105 | 6.67% | 0 |
| 02:00 | ₪6.49 | 67 | 5.97% | 0 |
| 03:00 | ₪5.85 | 34 | 2.94% | 0 |
| 04:00 | ₪8.22 | 65 | 1.54% | 0 |
| 05:00 | ₪18.96 | 142 | 4.23% | 0 |
| 06:00 | ₪43.73 | 314 | 4.46% | 0 |
| 07:00 | ₪83.61 | 409 | 5.38% | 1 |
| 08:00 | ₪62.94 | 425 | 3.76% | 0 |
| 09:00 | ₪50.22 | 364 | 6.32% | 0 |
| 10:00 | ₪45.31 | 371 | 5.93% | 0 |
| 11:00 | ₪69.35 | 616 | 17.37% | 0 |
| 12:00 | ₪118.16 | 846 | 5.79% | 0 |
| 13:00 | ₪87.26 | 782 | 4.35% | 0 |
| 14:00 | ₪63.93 | 570 | 6.49% | 0 |
| 15:00 | ₪49.74 | 508 | 2.17% | 0 |
| 16:00 | ₪36.84 | 441 | 5.44% | 0 |
| 17:00 | ₪54.39 | 502 | 5.98% | 0 |
| 18:00 | ₪31.22 | 381 | 4.46% | 0 |
| 19:00 | ₪32.48 | 365 | 5.21% | 0 |
| 20:00 | ₪33.25 | 352 | 5.11% | 0 |
| 21:00 | ₪44.88 | 611 | 4.09% | 0 |
| 22:00 | ₪54.24 | 658 | 4.26% | 0 |
| 23:00 | ₪42.41 | 516 | 3.29% | 0 |
Day-of-week
| Day | Spend | Impr | Leads |
|---|---|---|---|
| Sun | ₪154.29 | 1,617 | 0 |
| Mon | ₪225.69 | 1,887 | 0 |
| Tue | ₪195.93 | 1,176 | 0 |
| Wed | ₪57.82 | 224 | 0 |
| Thu | ₪136.59 | 1,702 | 0 |
| Fri | ₪166.96 | 1,740 | 1 |
| Sat | ₪145.69 | 1,361 | 0 |
Impression device
| Device | Spend | Impr | CTR | Leads |
|---|---|---|---|---|
| android_smartphone | ₪791.76 | 7,275 | 5.95% | 1 |
| iphone | ₪192.32 | 1,800 | 5.39% | 0 |
| desktop | ₪94.39 | 574 | 2.09% | 0 |
| android_tablet | ₪3.44 | 41 | 7.32% | 0 |
| ipad | ₪1.07 | 17 | 0.00% | 0 |
Landing Pages — GA4
/l/add-milion
SessionsNone
UsersNone
PageviewsNone
Bounce rate0.00%
Engagement rate0.00%
Avg session0.0s
Conversions (GA4)None
Devices—
/l/pension-course
SessionsNone
UsersNone
PageviewsNone
Bounce rate0.00%
Engagement rate0.00%
Avg session0.0s
Conversions (GA4)None
Devicesmobile: 346, desktop: 121, tablet: 2
GA4 reports 0 conversions for both pages — the 'Lead' conversion event/key event is not configured in GA4 property 498071676. Supabase lead counts are the source of truth for conversions throughout this report.
Daily sessions
| Page | Trend | First | Last | Total |
|---|---|---|---|---|
| /l/pension-course | 45 | 4 | 469 |
Traffic sources — /l/pension-course
| Source | Medium | Campaign | Sessions | Users |
|---|---|---|---|---|
| fb | paid | 120246648939920749 | 137 | 124 |
| an | paid | 120246648939920749 | 99 | 99 |
| (direct) | (none) | (direct) | 94 | 78 |
| organic | (organic) | 34 | 13 | |
| ig | paid | 120246648939920749 | 21 | 21 |
| paid_social | pension_course | 18 | 18 | |
| m.facebook.com | referral | (referral) | 17 | 17 |
| InforuMail | pension-course-arutz7-2026-04-28 | 9 | 3 | |
| (not set) | smoketest | 6 | 1 | |
| bing | organic | (organic) | 5 | 1 |
| eventsmanager.facebook.com | referral | (referral) | 5 | 1 |
| facebook.com | referral | (referral) | 5 | 5 |
Behavior analytics — cross-validated (7 sources)
Multi-source view of on-page user behavior. PostHog, Clarity and Cloudflare Web Analytics fire site-wide; Sentry/Smartlook/OpenReplay/LogRocket are opt-in per page via
src/config/behavior-tools-activation.ts. Sources marked 'missing key' are wired in code but their vendor account/env-var has not been configured yet.PostHogOK
rage=2 · clicks=25 · summaries=0
Microsoft ClarityOK
dashboard (API dead)
Cloudflare WAOK
pv=0 · visits=0
Matomo CloudOK
visits=0 · pv=0 · heatmaps=0
ContentsquareOK
MCP-only (REST API token missing)
SmartlookOK
heatmaps=0
OpenReplayOK
dashboard
LogRocketOK
sessions=0
capture: 8/8 sources OK · for the window 2026-05-13 → 2026-06-10
✓ PostHog
Top rage-clicks by element
| element | rage | users |
|---|---|---|
input.bg-white/10.border.border-white/20.disabled:cursor-not-allowed.disabled:op | 1 | 1 |
input.bg-white/10.border.border-white/20.disabled:cursor-not-allowed.disabled:op | 1 | 1 |
✓ Microsoft Clarity
Pixel-coord visual heatmaps in Clarity dashboard (API dead for our project; visual inspection only).
✓ Cloudflare WA
0 pageviews / 0 visits (server-side, ad-blocker-immune).
✓ Matomo Cloud
✓ Contentsquare
✓ Smartlook
✓ OpenReplay
✓ LogRocket
Audience — Meta × GA4
Meta — region (Israel)
| Region | Spend | Impr | CTR | Leads |
|---|---|---|---|---|
| Central District | ₪431.72 | 3,726 | 5.56% | 0 |
| Tel Aviv | ₪245.00 | 2,014 | 6.65% | 0 |
| Heifa | ₪127.75 | 1,239 | 3.95% | 0 |
| Northern District | ₪101.94 | 988 | 5.97% | 0 |
| Jerusalem | ₪92.19 | 774 | 6.07% | 0 |
| Southern District | ₪84.37 | 966 | 5.07% | 0 |
/l/pension-course
GA4 geo — /l/pension-course
| City | Region | Sessions | Bounce |
|---|---|---|---|
| Tel Aviv-Yafo | Tel Aviv District | 138 | 60.14% |
| Petah Tikva | Center District | 43 | 62.79% |
| Jerusalem | Jerusalem District | 23 | 65.22% |
| Dublin | County Dublin | 22 | 36.36% |
| Rishon LeZion | Center District | 21 | 66.67% |
| Holon | Center District | 13 | 76.92% |
| Haifa | Haifa District | 12 | 75.00% |
| Netanya | Center District | 11 | 63.64% |
| Ramat Gan | Tel Aviv District | 11 | 90.91% |
| Rehovot | Center District | 11 | 81.82% |
| Prineville | Oregon | 8 | 87.50% |
| Ra'anana | Center District | 8 | 37.50% |
GA4 tech — /l/pension-course
| Device | OS | Browser | Sess | Bounce | Engaged |
|---|---|---|---|---|---|
| mobile | Android | Chrome | 162 | 71.60% | 28.40% |
| mobile | Android | Android Webview | 113 | 73.45% | 26.55% |
| desktop | Windows | Chrome | 106 | 39.62% | 60.38% |
| mobile | iOS | Safari | 52 | 86.54% | 13.46% |
| desktop | Windows | Edge | 11 | 63.64% | 36.36% |
| mobile | Android | Samsung Internet | 11 | 72.73% | 27.27% |
| mobile | iOS | Chrome | 4 | 25.00% | 75.00% |
| desktop | Linux | Chrome | 2 | 50.00% | 50.00% |
| mobile | iOS | Safari (in-app) | 2 | 0.00% | 100.00% |
| tablet | Android | Android Webview | 2 | 50.00% | 50.00% |
| desktop | (not set) | (not set) | 1 | 100.00% | 0.00% |
| desktop | Macintosh | Chrome | 1 | 0.00% | 100.00% |
GA4 new vs returning — /l/pension-course
| Type | Sessions | Users | Bounce | Avg sess |
|---|---|---|---|---|
| new | 355 | 356 | 70.70% | 46.1s |
| returning | 72 | 35 | 25.00% | 291.4s |
| (not set) | 40 | 26 | 90.00% | 3.0s |
Conversion + Funnel (Supabase)
Real leads (window)
5
2026-05-13→2026-06-10 · tests excluded
FB-ad (window)
1
fb_campaign bucket
Other real (window)
4
email/Arutz7/direct
Tests excluded (window)
2
your QA submissions
Paid (window, real)
0
0.00% of real
Revenue (window)
₪0.00
real paid × price
Paid (all-time, real)
13
unique real payers
Revenue (all-time)
₪2,340
real paid × price
Avg days to pay
0.0
lead→paid
Daily leads (FB vs other)
| Date | FB | Other | Total |
|---|---|---|---|
| 2026-05-15 | 1 | 0 | 1 |
| 2026-05-25 | 0 | 1 | 1 |
| 2026-06-05 | 0 | 1 | 1 |
| 2026-06-07 | 0 | 1 | 1 |
| 2026-06-08 | 0 | 1 | 1 |
Funnel (canonical 4-stage model)
① Invitation
0
owned blasts — context only
② Registration
5
leads (window)
③ Reminders
145
20 leads engaged
④ Completion
0
0.00% of leads paid
These four stages are not a single monotonic funnel: Stage 1 (invitation) counts owned email/WhatsApp blasts and is outside the Meta-paid acquisition scope; Stage 3 (reminders) is a volume metric (multiple sends per lead), not a subset of leads. The true monotonic acquisition funnel (impressions→clicks→sessions→leads→paid) is in the End-to-End section.
Meta CAPI delivery health (server-side, recency-aware)
CAPI Lead — last
OK
2026-06-08 05:05
Fails last 48h
0
none
Lead lifetime
5/5 ok
100.00%
CAPI Purchase
2/2 ok
last OK
Post-lead reminder/notification mechanics (page health, not acquisition)
| Campaign | Ch | When | Recip | Deliv% | Open% | Click% | Bounce% |
|---|---|---|---|---|---|---|---|
| pension-reminder_24h-4098c48c-wa | 2026-06-09 05:10 | 1 | 0.00% | 0.00% | 0.00% | 0.00% | |
| pension-reminder_24h-4098c48c | 2026-06-09 05:10 | 1 | 100.00% | 0.00% | 0.00% | 0.00% | |
| pension-reminder_immediate-f690741a-wa | 2026-06-08 15:28 | 1 | 0.00% | 0.00% | 0.00% | 0.00% | |
| pension-reminder_immediate-f690741a | 2026-06-08 15:28 | 1 | 100.00% | 0.00% | 0.00% | 0.00% | |
| pension-reminder_24h-f690741a-wa | 2026-06-08 14:45 | 1 | 0.00% | 0.00% | 0.00% | 0.00% | |
| pension-reminder_24h-f690741a | 2026-06-08 14:45 | 1 | 100.00% | 0.00% | 0.00% | 0.00% | |
| pension-lead_received-4098c48c-wa | 2026-06-08 05:05 | 1 | 0.00% | 0.00% | 0.00% | 0.00% | |
| pension-lead_received-4098c48c | 2026-06-08 05:05 | 1 | 200.00% | 150.00% | 0.00% | 0.00% | |
| pension-next_session-d66b497f-wa | 2026-06-07 18:20 | 1 | 0.00% | 0.00% | 0.00% | 0.00% | |
| pension-next_session-d66b497f | 2026-06-07 18:20 | None | 0.00% | 0.00% | 0.00% | 0.00% | |
| pension-next_session-15aae757-wa | 2026-06-07 18:20 | 1 | 0.00% | 0.00% | 0.00% | 0.00% | |
| pension-next_session-15aae757 | 2026-06-07 18:20 | 1 | 100.00% | 0.00% | 0.00% | 0.00% | |
| pension-next_session-971a6cfd-wa | 2026-06-07 18:20 | 1 | 0.00% | 0.00% | 0.00% | 0.00% | |
| pension-next_session-971a6cfd | 2026-06-07 18:20 | 1 | 100.00% | 0.00% | 0.00% | 0.00% | |
| pension-next_session-584231ba-wa | 2026-06-07 18:20 | 1 | 0.00% | 0.00% | 0.00% | 0.00% | |
| pension-next_session-584231ba | 2026-06-07 18:20 | 1 | 100.00% | 200.00% | 0.00% | 0.00% | |
| pension-next_session-655be24c-wa | 2026-06-07 18:20 | 1 | 0.00% | 0.00% | 0.00% | 0.00% | |
| pension-next_session-655be24c | 2026-06-07 18:20 | 1 | 100.00% | 100.00% | 0.00% | 0.00% | |
| pension-next_session-df6a1d93-wa | 2026-06-07 18:20 | 1 | 0.00% | 0.00% | 0.00% | 0.00% | |
| pension-next_session-df6a1d93 | 2026-06-07 18:20 | 1 | 100.00% | 0.00% | 0.00% | 0.00% |
Lead Quality — every registration, all-time
Every
pension_course_leads row, all-time, classified by a deterministic first-match rule set (zero false-positives on this set). The headline counts only fb_campaign. Reproduce manually:SELECT * FROM pension_course_leads
WHERE superseded_by IS NULL
AND lower(utm_source) IN ('fb','facebook','meta','instagram','ig','facebook_paid')
AND lower(coalesce(email,'')) NOT LIKE '%@campaigner.co.il'
AND lower(coalesce(email,'')) NOT LIKE '%@goola-group.com'
AND coalesce(utm_campaign,'') NOT ILIKE '%smoketest%'
ORDER BY created_at;fb_campaign
1
0 paid — THE headline
real_other
25
13 paid — email/Arutz7/direct
test
2
your QA submissions (excluded)
superseded
10
dedupe duplicates (excluded)
TOTAL rows
38
pension_course_leads all-time
Real paid (unique)
13
fb_campaign + real_other
Real registrations by month
| Month | Real leads | Paid | From FB ad |
|---|---|---|---|
| 2026-04 | 10 | 5 | 0 |
| 2026-05 | 13 | 8 | 1 |
| 2026-06 | 3 | 0 | 0 |
Test submissions excluded (2) — itemised with the rule that caught each
| id | date | name | phone | rule | |
|---|---|---|---|---|---|
| 7bdbb561 | 2026-06-07 | בדיקה | yaak…@gmail.com | …5274 | name contains בדיקה |
| 9c92126b | 2026-05-27 | יוסי | yosi…@goola-group.com | …3044 | internal admin email domain |
Full classification — all rows (contact masked)
| id | date | name | utm_source | utm_campaign | pay | bucket | |
|---|---|---|---|---|---|---|---|
| 514c186b | 2026-04-24 | שלומית | shlo…@gmail.com | — | — | paid | real_other |
| 9a1109b3 | 2026-04-24 | מנחם | 0508…@gmail.com | — | — | — | real_other |
| 56f39b12 | 2026-04-26 | יאיר | baba…@gmail.com | — | — | pending | superseded |
| 2e16d721 | 2026-04-26 | יאיר | baba…@gmail.com | — | — | pending | superseded |
| 24bdedee | 2026-04-26 | יאיר | baba…@gmail.com | — | — | pending | superseded |
| 909b0bfb | 2026-04-26 | יאיר | baba…@gmail.com | — | — | pending | superseded |
| 7a2409d2 | 2026-04-26 | יאיר | baba…@gmail.com | — | — | pending | superseded |
| 19ca2863 | 2026-04-26 | יאיר | baba…@gmail.com | — | — | pending | superseded |
| c8d3eede | 2026-04-26 | יאיר | baba…@gmail.com | — | — | pending | superseded |
| 971a6cfd | 2026-04-26 | יאיר | baba…@gmail.com | — | — | paid | real_other |
| dffcbc3e | 2026-04-26 | עודד | oded…@gmail.com | — | — | pending | superseded |
| d16d7491 | 2026-04-27 | קובי | kobi…@gmail.com | — | — | pending | real_other |
| 9bb2a988 | 2026-04-28 | נתן | nata…@gmail.com | InforuMail | pension-course-arutz7- | — | real_other |
| bd36af14 | 2026-04-29 | עודד | oded…@gmail.com | — | — | pending | superseded |
| 655be24c | 2026-04-29 | אבי | avim…@gmail.com | InforuMail | pension-course-invitat | paid | real_other |
| b6b30043 | 2026-04-29 | מנחם | lauf…@bezeqint.net | InforuMail | pension-course-arutz7- | paid | real_other |
| a8cf3fc0 | 2026-04-29 | אהרון | ahr2…@gmail.com | InforuMail | pension-course-arutz7- | pending | real_other |
| 5cc5d3bc | 2026-04-30 | יובל | oded…@gmail.com | — | — | — | real_other |
| b5309a78 | 2026-04-30 | מריה | mari…@gmail.com | InforuMail | pension-course-invitat | paid | real_other |
| 115ad5b2 | 2026-05-01 | אילנה | ilan…@gmail.com | InforuMail | pension-course-MAIN-ba | paid | real_other |
| f503388e | 2026-05-01 | שלום | bs11…@gmail.com | InforuMail | pension-course-MAIN-ba | paid | superseded |
| eed28299 | 2026-05-01 | שלום | bs11…@gmail.com | InforuMail | pension-course-MAIN-ba | paid | real_other |
| 11485b9e | 2026-05-02 | יגיל | yagi…@gmail.com | — | — | — | real_other |
| 9152555b | 2026-05-02 | הדר | hada…@gmail.com | — | — | — | real_other |
| 584231ba | 2026-05-03 | משה | mema…@gmail.com | — | — | paid | real_other |
| e1cc47b6 | 2026-05-05 | אילון | 123e…@gmail.com | — | — | paid | real_other |
| 13995cb6 | 2026-05-06 | מיכל | erli…@gmail.com | — | — | paid | real_other |
| d66b497f | 2026-05-07 | יעקב | jay…@shames.co.il | — | — | paid | real_other |
| a7ae7099 | 2026-05-10 | אורלי | orly…@gmail.com | InforuMail | pension-course-MAIN-ba | — | real_other |
| 15aae757 | 2026-05-11 | שירה | shir…@gmail.com | — | — | paid | real_other |
| df6a1d93 | 2026-05-11 | חיים | haim…@gmail.com | InforuMail | pension-course-MAIN-ba | paid | real_other |
| 09dc2174 | 2026-05-15 | רונן | r_da…@yahoo.com | fb | 120246648939920749 | — | fb_campaign |
| 2e5afcce | 2026-05-25 | יאיר | yair…@gmail.com | InforuMail | pension-course-arutz7- | — | real_other |
| 9c92126b | 2026-05-27 | יוסי | yosi…@goola-group.com | — | — | — | test |
| bdd90b50 | 2026-06-05 | אהרן | aaro…@gmail.com | — | — | — | real_other |
| f690741a | 2026-06-07 | שביט | shav…@gmail.com | — | — | — | real_other |
| 7bdbb561 | 2026-06-07 | בדיקה | yaak…@gmail.com | — | — | — | test |
| 4098c48c | 2026-06-08 | איתן | ets3…@gmail.com | — | — | — | real_other |
Funnel — Deep KPIs (round · timing · attribution)
Round selection
| Round | Leads | Paid | Lead→paid |
|---|---|---|---|
| round2 | 4 | 0 | 0.00% |
| round3 | 1 | 0 | 0.00% |
Leads by hour-of-day (Asia/Jerusalem)
| Hour | Leads | |
|---|---|---|
| 07:00 | 1 | |
| 08:00 | 1 | |
| 13:00 | 1 | |
| 17:00 | 1 | |
| 22:00 | 1 |
Leads by day-of-week
| Day | Leads | |
|---|---|---|
| Sun | 1 | |
| Mon | 2 | |
| Fri | 2 |
Source → paid attribution
| Source | Leads | Paid | Conv% | Revenue |
|---|---|---|---|---|
| other | 4 | 0 | 0.00% | ₪0.00 |
| fb | 1 | 0 | 0.00% | ₪0.00 |
'fb' = utm∈{facebook,fb,meta,instagram,ig} or landing_url contains pension-course-fb; 'other' = organic / email / direct.
Reminder/notification effectiveness
| Notification type | Leads reached |
|---|---|
| session_reminder_5m | 13 |
| session_reminder_1h | 13 |
| next_session | 13 |
| session_reminder_24h | 13 |
| session_reminder_5m_s0 | 7 |
| session_reminder_24h_s0 | 7 |
| next_session_s0 | 7 |
| session_reminder_1h_s0 | 7 |
| session_reminder_1h_s2 | 6 |
| session_reminder_1h_s1 | 6 |
| next_session_s1 | 6 |
| session_reminder_24h_s2 | 6 |
| session_reminder_5m_s2 | 6 |
| session_reminder_5m_s1 | 6 |
| lead_received | 5 |
| reminder_1w | 4 |
| reminder_24h | 3 |
| reminder_immediate | 3 |
| confirmation | 2 |
0 of 0 paid leads received ≥1 notification before paying. 20 of 5 leads got any notification.
GA4 top events — /l/pension-course
| Event | Count | Users | |
|---|---|---|---|
| page_view | 558 | 389 | |
| session_start | 442 | 373 | |
| first_visit | 347 | 341 | |
| scroll | 137 | 94 | |
| user_engagement | 136 | 63 | |
| form_start | 32 | 17 | |
| generate_lead | 7 | 4 |
End-to-End Funnel & Unit Economics
Acquisition funnel (paid → revenue)
Meta impressionsCPM ₪111.57
9,707
Meta link clicks3.7% of priorCPC ₪1.99
359
GA4 sessions (/l/pension-course)0.0% of prior ▼ 100% lost
0
DB leads (window)prior n/a₪216.59/lead
5
Paid0.0% of prior
0
Meta link-clicks vs GA4 sessions will not match 1:1: GA4 needs JS to fire (mobile/iOS blockers), and Meta uses the ad account's local date while Supabase stores UTC — expect up to a 1-day boundary shift in daily comparisons. Treat hop %s as directional.
פילוח: סדנה (lead-magnet) מול קורס (paid)
סדנת מיליון לפנסיה (free workshop, Meta-targeted)
GA4 sessionsNone
GA4 usersNone
Bounce rate0.00%
Avg session0.0s
Workshop registrations—
Zoom attendees—
Reg→attended—
קורס פנסיוני (paid course, organic-only)
GA4 sessionsNone
GA4 usersNone
Bounce rate0.00%
Avg session0.0s
DB leads (attributed)5
DB paid0
Lead→paid0.00%
Post-pivot 2026-05-24: workshop is the new Meta-campaign target; pension-course is organic + distributor only. Workshop registrations + attendance counts are placeholders pending a follow-up form_submissions fetcher (the existing pension-course funnel queries don't cover /api/conference-webhook submissions).
Optimization & Data Quality
Campaign is unprofitable: ROAS 0.00× (₪0.00 rev vs ₪1,083 spend). Cut spend on losing creatives or raise lead→paid before scaling.
Ad destination misrouting — 299 paid-social sessions landed on the generic
/l/pension-course vs only 0 on the purpose-built /l/pension-course-fb. The Meta ad set's destination URL is pointing at the wrong page; the variant you built for this campaign is starved. Repoint the ad creative's link to /l/pension-course-fb.5 real lead(s), 0 paid in this window (tests excluded) — leads but no payments yet. The leak is lead→checkout→paid: review reminder timing/copy and the checkout page, not the ad.
GA4 'Lead' key event not configured (property 498071676) — you cannot optimise Meta toward GA4 conversions or see funnel drop-off in GA4. Configure the key event; DB remains the conversion source of truth meanwhile.
Reconciled — Meta 1 = DB real FB-ad 1. 2 test + 10 dupe rows excluded; 26 real registrations total (see Lead Quality). The numbers match — no attribution gap.
Data-quality flags
| Check | Status |
|---|---|
| GA4 Lead goal | NOT CONFIGURED |
| CAPI Lead (recent) | last OK · 5/5 lifetime |
| Dedupe-suppressed leads | 0 |
| Meta token expiry | 29d left |
| Group-mode send attribution | customer_message_id may be absent pre-2026-05-14 |
CRO: prior Playwright renders of both pages exist at
/tmp/pension-course.html and /tmp/pension-course-fb.html (2026-05-17) for qualitative review against the page-cro framework.External AI peer review — independent, facts-only
Generated 2026-05-20T07:24:57Z from a facts-only data bundle (snapshot 2026-05-20T07:00:03Z). Each model saw the raw numbers with our analysis, recommendations and verdicts withheld — and did NOT see the other models' output or any prior review. Blocks are verbatim per model, deliberately not blended.
Gemini gemini-3.1-pro-preview · 23.3s
High ad engagement but severe landing page drop-off and zero ROAS yield an unsustainable 1,024 ILS CPL.
What is working
- Strong top-of-funnel ad engagement with an overall CTR of 5.62% and a low CPC of 1.92 ILS.
- The 'Short-A multi-aspect' ad is highly efficient at driving traffic, achieving a 7.11% CTR and 1.24 ILS CPC.
- Email automation sequences are successfully delivering and achieving high open rates (e.g., up to 100%+ open rates on reminders).
Critical problems
- Catastrophic landing page conversion rate. — 351 link clicks and 241 landing page views resulted in only 1 Meta lead, driving the CPL to an unprofitable 1,024.72 ILS with 0 paid conversions.
- Severe mobile bounce rate on the FB-specific landing page. — The '/l/pension-course-fb' path has a 91.2% bounce rate and a 2.36-second average session duration, with Android Chrome users bouncing at 93.3%.
- WhatsApp automation is completely broken. — All WhatsApp broadcast metrics across the funnel show 0 delivered, 0 opens, and a 0% delivery rate.
Top actions
- Audit and fix the mobile landing page experience. [high] — A 93% bounce rate and 2-second session duration on Android indicates a technical error, slow load time, or immediate UX friction preventing conversions.
- Troubleshoot WhatsApp messaging integration. [high] — A 0% delivery rate on WhatsApp reminders means a critical follow-up channel for unpaid leads is failing, directly impacting the bottom-line ROAS.
- Re-evaluate demographic targeting. [medium] — Spend is heavily concentrated on Males 55-64 and 65+ (over 480 ILS combined); verify if this demographic is actually likely to purchase the course or if they are just clicking without intent.
Caveats
- The 7-day window is short and may not account for longer consideration cycles before payment.
- GA4 shows 0 conversions despite 1 lead recorded in the database and Meta, indicating a potential client-side tracking issue.
- WhatsApp delivery metrics might be a reporting API failure rather than actual message failure, but this requires immediate verification.
model confidence: high
OpenAI (GPT) gpt-5 · 146.8s
Traffic OK, conversion layer broken: 1 lead for ₪1,024.72 spend, ROAS 0; GA4 lead conversion not configured; mobile engagement very poor
What is working
- Top-funnel delivery is fine: CTR 5.62% (unique CTR 8.16%), CPC ₪1.92, 351 outbound clicks, 241 landing page views (≈69% of outbound clicks reached LP)
- Video engagement is solid: 1,027 thruplays; 181 full video views (p100), indicating creatives are being watched
- Older audiences respond best: Female 65+ CTR 8.19%, Female 45–54 CTR 6.99%; regions like Tel Aviv CTR 6.75%
- Time-of-day shows strong click intent: 11:00 hour CTR 17.37%; 14:00 CTR 6.50%
- Creative signal: Short-A had highest CTR 7.11% and lowest CPC ₪1.25 among ads (though no leads); Main-A produced the only lead
Critical problems
- Extremely poor efficiency: only 1 lead, ROAS 0 — Spend ₪1,024.72; Meta-attributed leads 1; ad_cpl ₪1,024.72; revenue 0; roas 0.0
- Severe LP drop-off from click to lead — 351 outbound clicks → 241 landing page views (≈69%) → 1 lead (≈0.41% LPV→lead), far below typical benchmarks
- GA4 conversions not configured, obscuring measurement — ga4_lead_event_configured=false; GA4 shows generate_lead events (e.g., 3 on /l/pension-course-fb) but ‘conversions’=0 everywhere
- Very high mobile bounce and low engagement on landing pages — /l/pension-course: Android Chrome bounce 74.78% (engagement 25.22%), iOS Safari bounce 86.54% (engagement 13.46%); /l/pension-course-fb: Android Chrome bounce 93.33%, iOS Safari (in-app) bounce 100%
- Cross-platform tracking mismatch suggests instrumentation gaps — GA4 paid sessions to LPs (fb+ig+an with campaign 120246648939920749) ≈338 vs Meta Landing Page Views 241 (≈+40% in GA4), indicating missing/faulty Meta events or inconsistent tagging, especially on /l/pension-course-fb
- Performance decay over time — Daily CTR fell from 11.05% (May 14) to 3.60–4.09% (May 18–19) while spend rose (e.g., ₪225.69 on May 18)
- Low-quality placements risk — Audience Network ‘rewarded_video’ CTR 41.45% and ‘an_classic’ CTR 15.95% with minimal spend—signals click inflation potential; GA4 shows weak mobile engagement broadly
Top actions
- Keep campaign paused until measurement is fixed end-to-end (GA4 + Meta Pixel/CAPI) [high] — Without trustworthy conversion tracking, optimization is blind; current CPL is untenable (₪1,024.72) with ROAS 0
- Configure GA4 lead as a conversion and QA events on both LPs (/l/pension-course and /l/pension-course-fb) [high] — ga4_lead_event_configured=false while generate_lead is firing; mark it as a conversion, align event names/parameters across both LPs, and validate in GA4 DebugView
- Audit Meta Pixel events (Lead, ViewContent/LandingPageView) on all LP variants and unify UTM tagging [high] — Meta shows 241 LPVs vs GA4 ≈338 paid sessions; ensure LPV/Lead fire on both pages, deduplicate, and use consistent utm_source/medium/campaign across creatives
- Fix mobile LP experience to reduce bounce and increase form starts [high] — Mobile bounce is 75–100% with very low engagement; prioritize performance (load speed), clarity (above-fold CTA), and friction reduction on the form to lift the 0.41% LPV→lead rate
- Constrain placements and rebalance budget to higher-quality inventory [medium] — Exclude Audience Network (especially rewarded video); emphasize Facebook Feed and high-quality placements where engagement is stronger; monitor Instagram Reels if on-site quality remains weak
- Refocus creatives based on early signals and performance decay [medium] — ‘Short-A’ drives cheapest clicks (CTR 7.11%, CPC ₪1.25) and older female segments click most; test iterations of high-performing angles but prioritize post-click fixes first
- Time/dayparting tests once tracking is reliable [low] — 11:00 had CTR 17.37% and mid-day bands were stronger; could improve click quality and costs after conversion tracking is stable
Caveats
- Short window (May 13–19) and only 1 Meta-attributed lead limit statistical confidence
- GA4 conversions are unconfigured, so true conversion rate is uncertain despite event counts
- Placement-level and device-level quality inference is limited without reliable post-click conversion data
- Email/WhatsApp broadcast stats appear inconsistent (e.g., delivered_count > recipients_count in some rows), so downstream nurture impact is unclear
model confidence: medium
Grok (xAI) grok-3 · 23.2s
High engagement but catastrophic CPL of 1024 ILS with zero revenue in 7-day window
What is working
- Strong overall CTR of 5.62% and video views (3117 total)
- Main-A ad drove the only Meta lead (CTR 6.06%, 1 lead)
- Decent reach (5014) with frequency under 2
- Older demographics (55-64 and 65+) showing higher CTR and spend share
Critical problems
- Extremely high CPL of 1024.72 ILS against 180 ILS price point — 1024.72 spend for 1 meta_leads and 1 fb_lead; economics shows cpl:1024.72 and ad_cpl:1024.72
- Zero conversions and revenue from the campaign — revenue:0, paid:0, roas:0.0, net:-1024.72; GA4 shows 0 conversions across 358+ sessions on /l/pension-course
- Very high bounce rates and low engagement on landing pages — GA4 new users bounceRate 0.73, averageSessionDuration 43s; mobile Android dominant with bounce ~0.74-0.93
- Campaign paused with minimal Meta-attributed results — status:PAUSED; only 1 fb_campaign lead in db_leads; most real leads from email/other sources
Top actions
- Immediately pause or drastically cut budget until landing page and tracking fixed [high] — ROAS 0 and CPL 5-6x course price with only 1 lead in 7 days
- Fix GA4 conversion tracking and test form_start to lead flow [high] — 23 form_start events but 0 conversions; ga4_lead_event_configured:false
- Shift targeting toward 55+ age groups and test new creatives [medium] — Highest CTR and spend in 55-64/65+; Main-A ad performed best
- Review placement mix and reduce Audience Network spend [low] — Audience Network has high CTR but low volume; main spend on FB reels/feed
Caveats
- Many db metrics appear all-time while ad data is 7-day window
- GA4 lead events not configured; conversions shown as 0
- Suppressed_count_window:0 but 10 superseded leads in all-time data
- Daily data missing May 13; campaign start mid-day
model confidence: medium
Perplexity sonar-reasoning-pro · 59.4s
Meta pension-course campaign: strong click engagement but effectively zero sales — 1 very expensive lead (~₪1,025 CPL), ROAS 0, with the funnel breaking after the landing page.
What is working
- Top-of-funnel engagement is solid: 9,482 impressions, reach 5,014, CTR 5.62% and link CTR 3.70% with CPC ≈₪1.92 indicate the ads are compelling enough to win clicks at a reasonable cost.
- Video engagement is decent: 3,117 video views, 1,027 thruplays, and 682 50% views on only 9,482 impressions show creatives are attention-grabbing and watchable.
- Frequency is healthy at 1.89 over 7 days, so there is no obvious ad fatigue or overexposure yet.
- Traffic volume to the landing funnels is substantial: Meta reports 241 landing-page views; GA4 sees ≈339 sessions on the two pension-course landing URLs from this campaign ID alone (fb/an/ig), so there is enough traffic to evaluate the funnel.
- Server-side tracking (CAPI) appears healthy: both lead and purchase events show 100% success (no recent fails), and Meta attributes 1 lead via pixel (offsite_conversion.fb_pixel_lead).
- The underlying product and checkout funnel are validated by other channels: all-time there are 22 real leads and 13 paid buyers (₪2,340 revenue) from email/other sources, so the offer and payment flow can convert when traffic is qualified.
- Among creatives, "Short-A multi-aspect" achieves the best CTR (7.11%) and lowest CPC (₪1.25), and overall video variants outperform static images on engagement metrics, suggesting video is the right format direction.
Critical problems
- Economics are currently unviable: 1 lead for ₪1,024.72 spend (CPL≈₪1,025) and 0 recorded sales from the campaign (ROAS=0). — Campaign spend is ₪1,024.72; Meta-attributed leads=1.0; db_leads=1 from bucket "fb_campaign" in the window; paid=0 and revenue=0 in the economics block; ad_cpl=₪1,024.72 and roas=0.0. With course price ₪180, even a 100% lead-to-sale rate would not recover ad spend at this CPL.
- Massive funnel drop-off between clicks/landing-page views and leads — conversion rate is near zero for Meta traffic. — Meta shows 351 link clicks and 241 landing_page_view actions (≈69% click→LPV), but only 8 view_content events and 1 lead. GA4 for /l/pension-course and /l/pension-course-fb shows ≈339 sessions from this campaign (fb+an+ig, campaignName=120246648939920749) with 0 configured conversions and only 1 fb_campaign lead in the DB. That implies a session→lead rate of roughly 1/339≈0.3%, far below the email/other channels where 13 of 22 real leads are paid.
- Landing pages for paid social traffic show very poor engagement and extremely high bounce, especially the dedicated FB variant. — On /l/pension-course, new users have a 73% bounce rate and only 102 scroll events and 23 form_start events across 339 total sessions (all sources). On /l/pension-course-fb, new users show 91.2% bounce (new_vs_returning), and device breakdown shows iOS Safari (in-app and regular) with bounce ≈100% and 0 engagement for 49 sessions. Despite 149 page_view events and 125 session_start events on this FB variant, there are only 3 generate_lead events (1 user) and no recorded purchases from Meta traffic in the window.
- Meta user quality from certain placements (especially Audience Network) is very low — high click volume with zero downstream conversions. — Placement stats: Audience Network rewarded_video CTR=41.45% and an_classic CTR=15.95% on a combined 409 impressions, but GA4 shows 99 sessions from source "an" (sessionMedium=paid, campaignName=120246648939920749) to /l/pension-course with 0 conversions and high bounce. By contrast, placements like Facebook feed and Instagram feed have more moderate CTRs (6.12% and 4.02%) and align better with standard user intent for a pension course.
- Measurement in GA4 is incomplete/misaligned: lead events are not configured as conversions, making it hard to analyze or use GA as a reliable source of truth. — ga4_lead_event_configured=false at the property level. On /l/pension-course-fb, GA4 records generate_lead eventCount=3 (1 user) but conversions=0 for all days/sources. On /l/pension-course there are 23 form_start events but no generate_lead events and conversions are 0. This misconfiguration prevents GA4 from reflecting actual leads and complicates cross-checking Meta attribution vs onsite behavior.
- Meta is optimizing for leads but has almost no signal (only 1 lead), so the algorithm cannot effectively learn or improve. — Campaign objective is OUTCOME_LEADS; summary shows meta_attributed_leads=1.0 and offsite_conversion.fb_pixel_lead value=1.0. With only a single conversion over 9,482 impressions and 5,014 reached users, there is insufficient conversion volume for Meta’s optimization to distinguish good vs bad segments (age, placement, device, etc.).
- Email/owned-channel performance is far better than paid social in this period, suggesting current Meta spend is cannibalizing budget from a stronger channel mix. — DB shows 22 real leads all-time with 13 paid (₪2,340 revenue) mostly tagged with utm_source=InforuMail/email or untagged direct, many before the Meta campaign window. In the same 7-day window the Meta campaign adds only 1 unpaid lead while consuming ₪1,024.72. Bucket_counts.window_real: fb_campaign=1 vs real_total=1 for this window, i.e., all new real leads in this window are from the FB campaign, yet revenue in the economics window is 0.
Top actions
- Keep the Meta campaign paused (do not scale) until the landing-page experience and tracking for Meta traffic are fixed and tested to a reasonable conversion rate. [high] — With CPL≈₪1,025 and ROAS=0, further spend is almost certainly unprofitable. The main bottleneck is not clicks but the LP→lead step (≈0.3% conversion from sessions). As long as the funnel and measurement issues persist, adding budget will only compound losses without giving the algorithm enough quality signals to improve.
- Redesign and/or clearly re-position the Meta-specific landing page(s) to align tightly with ad promises and reduce bounce; run an A/B test versus the existing /l/pension-course page focused on LP→form_start and LP→lead conversion. [high] — Meta is already delivering engaged visitors (good CTR/CPC and 241 LPVs), yet almost none start or complete the form. The biggest drop is from landing_page_view (241) to view_content (8) and generate_lead (1). This strongly points to a mismatch between ad expectations and page content, or UX friction. A dedicated variant for paid social, with clearer value prop, social proof, simplified form, and mobile-first layout, directly tackles the main bottleneck and can be evaluated with LP-level metrics (form_start, generate_lead).
- Exclude or severely limit Audience Network placements (especially rewarded video) and focus spend on Facebook/Instagram feeds and Reels where intent is higher and attribution is clearer. [high] — Audience Network shows extremely high CTR (up to 41%) but 99 GA sessions from source "an" produce 0 leads or conversions, indicating low-intent or accidental traffic. Removing AN will slightly raise CPC but should improve average traffic quality and reduce wasted budget, letting more of the remaining spend go to placements (FB/IG feeds/Reels) where pension-course intent is more realistic.
- Fix GA4 conversion tracking: mark generate_lead and purchase events as conversions, and ensure both /l/pension-course and /l/pension-course-fb fire these events consistently on successful submissions/payments. [high] — Currently ga4_lead_event_configured=false and conversions=0 for all landing pages despite known leads and purchases from other channels. This prevents reliable funnel analytics and makes cross-channel comparison and optimization impossible. Configuring lead/purchase as conversions will allow accurate LP conversion rates, device/placement analysis, and better validation of Meta-reported results against onsite behavior.
- Tighten placement/device targeting based on engagement: bias budgets toward desktop and high-engagement mobile (e.g., Android Chrome), and away from iOS in-app traffic where bounce is ~100%. [medium] — Meta device stats show most impressions on Android smartphones, while GA4 tech data shows much higher engagement on desktop Windows/Chrome (bounce≈32%, engagement≈68%) and lower engagement on iOS Safari (bounce≈86–100%, engagement≈0–13%). Rebalancing toward higher-engagement environments can improve effective LP conversion without changing the creative, especially once conversion tracking is fixed and can validate these patterns.
- Consolidate around the best-performing video creatives and retire underperforming or expensive ones (e.g., static image ads with higher CPC and lower CTR), then iterate new variants of the strongest concept. [medium] — Among ads, "Short-A multi-aspect" has the best CTR (7.11%) and CPC (₪1.25), while static image ads ("תמונה 1/2/3") have lower CTR (2.44–3.30%) and higher CPC (₪2.75–₪4.39). Although all creatives currently convert poorly on leads, it is more efficient to test funnel fixes and new LPs using the lowest-CPC, highest-engagement creatives and stop spending on clearly weaker ones.
- Leverage and scale the already-proven email/owned channels while Meta is being fixed, as they currently generate all revenue in this dataset. [medium] — Email campaigns (utm_source=InforuMail) and other non-Meta sources produced 13 paid buyers and ₪2,340 revenue, compared to 0 from Meta in this window. Redirecting a portion of the acquisition/test budget into additional email sequences, list growth, and reactivation flows can drive incremental revenue with far better economics while the paid social funnel is reworked.
- Once tracking and LP are stabilized, relaunch Meta with smaller test budgets and clear success thresholds (e.g., target CPL and session→lead rate), using primary KPIs (cost per lead / cost per purchase) as the only optimization criteria. [medium] — Per modern Meta best practices, primary KPIs like amount spent, number of leads/purchases, and cost per result should guide optimization, while secondary metrics (CTR, CPC, video views) explain WHY something works. After fixes, restarting with controlled budgets and explicit targets will allow a data-driven decision on whether Meta can be a sustainable channel for this funnel, instead of scaling based on engagement metrics alone.
Caveats
- The analysis is based on a 7-day window with only 1 Meta-attributed lead, so any segmentation by age, gender, region, placement, or device for conversion performance is statistically unstable; recommendations at that level focus on engagement and bounce patterns, not proven ROI.
- GA4 lead and purchase events are not fully configured as conversions, and there are discrepancies between Meta landing_page_view counts and GA session counts; while the direction of findings (very low conversion from sessions to leads) is robust, exact conversion percentages may shift slightly after tracking is corrected.
- Some generate_lead events on /l/pension-course-fb (3 events, 1 user) may represent tests or non-campaign traffic; the DB, economics block, and Meta pixel attribution are treated as the primary sources of truth for revenue/lead counts.
- Benchmarks for “good” CTR, CPC, and CPM vary by market and niche; comments on engagement quality are relative to typical performance ranges for paid social, not to a specific industry benchmark dataset.
model confidence: high
Appendix
Run metadata
Generated
2026-06-11 10:00
Asia/Jerusalem
Window
2026-05-13 → 2026-06-10
29d
Wall time
10.8s
Meta calls
19
Raw cache
raw/
/tmp/pension-report-20260610
Source errors / warnings
GA4totals /l/add-milion: HTTP 400: { "error": { "code": 400, "message": "Requests are limited to 10 metrics within a nested request.\n This request is for 12 metrics.", "status": "INVALID_ARGUMENT" } }
totals /l/pension-course: HTTP 400: { "error": { "code": 400, "message": "Requests are limited to 10 metrics within a nested request.\n This request is for 12 metrics.", "status": "INVALID_ARGUMENT" } }
Notes & honest caveats
• Meta pixel
•
• Returning paid contacts who resubmit get HTTP 409 and are counted nowhere.
• Microsoft Clarity Data Export API is deterministically dead for this project — use the dashboard: https://clarity.microsoft.com/projects/view/v1cdss7xz0 (heatmaps/recordings, filter by these URLs).
/stats mixes browser+server events; the DB pension_course_lead_events is ground truth for actual leads. If Meta leads > DB leads ×1.2, suspect broken pixel/CAPI dedup.•
/l/pension-course-fb is intentionally noindexed (ad variant) — organic search to it should be ≈0.• Returning paid contacts who resubmit get HTTP 409 and are counted nowhere.
• Microsoft Clarity Data Export API is deterministically dead for this project — use the dashboard: https://clarity.microsoft.com/projects/view/v1cdss7xz0 (heatmaps/recordings, filter by these URLs).