# Analit keyword decisions, 3 October 2026

The current data supports measurement and query-intent investigations before changing spend. It does not establish profitable campaigns or a reason to pause keywords. Google Ads was read-only throughout this run.

Customer `1938291268` returned `Analit Kft.`, HUF and Europe/Budapest through manager `9494412337`. The project display is Analit Utazásszervező Kft. Snapshot `20261003T102808-85cecb0a` contains 185 keyword identities, 880 daily rows, 818 disclosed search-term rows and 11 campaigns. Account, campaign-type and keyword totals reconcile in exact micros, clicks and impressions. All 170 prior criteria are represented, with 15 new criteria and no unexplained missing identity.

| Provider period | Spend | Clicks | Impressions | Google Ads conversion events | Average CPC |
|---|---:|---:|---:|---:|---:|
| 3 September to 2 October 2026 | 82,872.12 HUF | 965 | 6,117 | 6 | 85.88 HUF |
| 4 August to 2 September 2026 | 43,998.38 HUF | 820 | 5,187 | 2 | 53.66 HUF |

Spend increased 88.4%, clicks 17.7% and average CPC 60.1%. Campaign composition changed. The new `ANALIT_HU_Search_CsoportosBali_2026Q4-2027H1` campaign accounts for 37,128.05 HUF and 320 latest-period clicks, so this is not a clean comparison of unchanged campaigns. All six latest-period events use `ANALIT_FORM_SUCCESS_20260914`. The prior two use `balionline.hu (web) dynamic_trip_form_submit`. These changed event definitions cannot prove a lead or ROI improvement.

| Priority and lane | Investigation | Current evidence | Decision-changing evidence |
|---|---|---|---|
| 1 · CHANGE investigation | Reconcile paid form events with CRM receipt, qualification and booking on both active campaign routes. | Old campaign `23492984549`: 45,744.07 HUF, 645 clicks, 0 Ads events in the latest period. New campaign `24253795852`: 37,128.05 HUF, 320 clicks, 6 Ads events. Tracking coverage is unknown. | Per-campaign successful form receipt and deduplicated CRM record, followed by qualified lead, booking, attributable revenue and contribution margin. Zero events alone are not a stop signal. |
| 2 · CHANGE investigation | Compare the duplicated `bali utazás` phrase terms and explain the intended split before considering restructuring. | Old criterion `193948822753~303678317454`: 16,936.67 HUF, 241 clicks, 1,641 impressions. New criterion `199834050709~303678317454`: 17,225.51 HUF, 149 clicks, 1,355 impressions. Together they consumed 34,162.18 HUF, 41.2% of latest spend. | A read-only query overlap, device/location and landing-to-CRM comparison. Different CPC does not demonstrate waste or self-competition. |
| 3 · STOP investigation | Review informational search intent under the old `bali utazás` phrase row. Do not apply negatives automatically. | In the full 60-day disclosure, `bali látnivalók` consumed 2,885.16 HUF for 43 clicks and 309 impressions. The parent keyword meets the family evidence floor with 4,021 impressions and 541 clicks. | Confirm the query's fit with the offered tours and actual CRM outcomes, plus any approved exclusion policy. The disclosure is sampled and uses a 60-day period. Lack of platform conversions is insufficient. |
| 4 · KEEP investigation | Preserve the useful price and organized-tour intent while validating commercial outcomes. | `bali nyaralás árak` under old phrase `bali nyaralás`: 13,511.17 HUF, 220 clicks, 665 impressions over 60 days. `bali szervezett utazás`: 2,279.78 HUF, 36 clicks, 143 impressions over 60 days. These are plausible commercial searches, not proven profitable leads. | CRM qualification and bookings per search intent. Do not promote or scale using CTR alone. |
| 5 · WAIT | Keep thin and legacy paused rows out of a premature exclusion or reactivation queue. | 177 of 185 keyword rows are below the combined 200-impression and 10-click floor over 60 days. 146 have zero impressions. Current serving diagnostics show paused legacy campaigns. | Intentional campaign scope, eligible serving evidence and a fuller measured window. Zero spend is not itself a needs-attention reason. |

The two main active ad groups have configured destinations at the group-trip collection page. The old URL contains a duplicate slash and an older campaign UTM label. Anonymous HTTP readback returns 200 and normalizes the duplicate slash while preserving the query parameters. This warrants attribution naming review. It does not prove a broken landing page. The KPC bundle's Planner requests have no configured language or geography, so their estimates should not be read as Hungary-specific demand.

The inherited field `phone_click_conversions` sums all Google Ads conversion actions. This account's eight 60-day events are form events, not a verified phone-call count. A separate provider field-presence query confirms that 148 raw Quality Score zeros represent an absent optional field, not a score of zero. The review presentation displays those scores as unavailable, corrects the event label and the shared template's obsolete column-help dates, and preserves the canonical embedded dataset. No tracking start date was inferred from the new action's name. CRM leads, qualified leads, bookings, revenue, margin, ROI and ROAS remain unverified in this KPC assignment.

## Reviewable workflow

[Keyword dashboard](./) · [Reusable Analit skill](workflow/SKILL.md) · [Period helper](workflow/period_metrics.py) · [Presentation helper](workflow/presentation.py) · [Advice validator](workflow/validate_advice.py)

The Sol xhigh pass covers all 185 criteria with 628 exact citations and zero rejected rows. All 69 flagged rows have two practical options. The imported advice has three account actions and a complete partition of 2 planning investigations, 6 next investigations and 177 waits.