Optimization of advertising campaigns in directories and catalogs for a network of medical clinics - audit, tracking, anti-fraud and increase in requests

Initial situation: a network of 6 private medical clinics is promoted in online directories and catalogs (specialized medical directories, city directories, service aggregators). Now there are paid placements and advertising packages in 8-10 sites, but applications are counted manually, some calls are not associated with a specific site, and the statistics contain many suspicious clicks and “empty” calls. Over the past 2 months, CPL has increased, but the quality of leads has decreased.

Technical environment and data: access to personal accounts of sites, Yandex Metrica and Google Analytics 4, call tracking (if there is no current one, connect via Calltouch or UIS with dedicated numbers), CRM (amoCRM or Bitrix24), advertising packages of sites without the ability to install arbitrary scripts on the directory side. Landing pages on Tilda and parts on WordPress.

What needs to be done (several steps):
1) Inventory of all placements in catalogs - cards, tariffs, categories, geo, display schedule, contact information, UTM and current links, moderation conditions, restrictions on content and medicine.
2) Setting up end-to-end tracking of requests from catalogs - a single UTM standard and naming rules, separate numbers/pools for key sites in call tracking, goals in Metrica/GA4, events based on clicks on the phone/messengers, forwarding the source to CRM (fields, rules, lead deduplication).
3) Analysis of traffic quality and anti-fraud hypotheses - comparison of metrics by site (CR per lead, bounce rate, depth, duration, repeat visits), identifying anomalies by time, geo, device, frequency; reconciliation with the logic of displays on sites. The result is a list of suspicious sites/locations and an argument as to what exactly is considered low-quality.
4) Budget and card optimization plan - redistribution of costs between sites and tariffs based on data, recommendations on the structure of cards (photos, USP, services, price list, schedule, answers, trust blocks) taking into account the restrictions of medical advertising and catalog rules, proposals for A-B checks where possible (options of offers/phone numbers/links).
5) Implementation of changes - making changes to cards and links, setting up reports and control dashboards (Looker Studio or Yandex DataLens) for hits, CPL, share of targeted leads and quality.

Specific result: a working hit attribution scheme for each catalog site (calls and requests), a dashboard for daily monitoring, a list of placements that can be disabled/restarted, and a set of implemented changes to cards and settings that allow you to manage CPL and lead quality.

Restrictions: you cannot change the code on the directory side, only link parameters, phone numbers/messengers and card contents within the framework of the site rules. Medical topics - without prohibited wording and taking into account the site's requirements for licenses/disclaimers. Access is granted in guest/marketer mode, without administrator rights to CRM (edits are approved and carried out through the responsible employee).

Measurable acceptance criteria: 1) at least 90% of calls and at least 80% of application forms from catalogs must fall into the CRM with the source/site/campaign fields correctly filled in; 2) a single documented UTM standard and a correspondence table: site - link - number - target; 3) dashboard with sections by site: costs, impressions/clicks (if available), visits, leads, targeted leads, CPL, share of suspicious ones; 4) a list of specific changes implemented for at least 6 sites (what happened - what happened) and a budget optimization plan for the next month with a forecast of the effect on CPL/number of leads.