- Test new offers
- Improve segmentation via headline rotation
- Personalise quizzes
- Reclaim lost leads
- Pilot the Master Campaigns tool
Project passport
HFE — Master Campaigns
Cost-per-lead comparison between Master Campaigns and standard campaigns
Key outcomes
- Lead share from Master Campaigns grew from 31% to 80%
- CPA dropped by 55% in Master Campaigns
Section 01
Contextual Advertising for a Medical Clinic
Section 02
Challenges of Advertising in the Medical Niche (2022)
Lack of trust in Master Campaigns among specialists
Marketers don't trust automation, preferring manual management due to fear of losing control.
Need for sufficient data for algorithm training
Algorithms need large volumes of data. Without statistics, the system cannot optimize campaigns.
Mixed results from past launches
Previous experience shows mixed results, creating doubts about the system's reliability.
Choosing the right target action was critical
Incorrectly chosen optimization goal can reduce overall advertising campaign effectiveness.
Section 03
Take the first step towards business growth - leave a request!
Book a Free AI Marketing Strategy Call
Filling out the form takes less than 1 minute and helps us prepare a tailored proposal specifically for your business.
1 Information
2 Contact
Website
Annual revenue of your business
1-year average revenue per customer (ARPU/ARPC)
Location
Next
Get Free Strategy in 24h1
The form takes less than 1 minutes to complete and helps us prepare a personalized offer for you and discuss it
1 Information
2 Contact
Your name
Your WhatsApp Number
Your email
How soon would you like to start working with us?
—Please choose an option—We have an urgent taskWithin 1-3 monthsWe are just gathering proposals, no urgencyOther
I agree with the privacy policy.
Back
Section 04
Setting Up Master Campaigns: Key Stages
Stage 1. Setup & Preparation
Created 3 campaigns using the principle:1 campaign = 1 target action for optimisationCampaigns & Target Actions
Campaign (Name is conditional)
Target Action
«Quiz»
User submitted a form after completing the quiz
«Call»
User called the clinic number
«All CRM Inflows»
All submissions and status updates tracked via Client ID
Set targeting across all of 🇷🇺 Russia
Added high-frequency keywords
Included 4 headline variations + 4 image options (to be clarified)
Optimised ad display for both male and female audiences
Set target cost per action at 50% of the maximum allowed by KPI
To help the algorithm learn quickly, we set the weekly budget equal to ten target actions — per Yandex recommendations
Important Note: With this setup, we only pay for conversions explicitly defined as target actions in each campaign. For example, if a campaign targets calls but users submit forms instead — those are free! We only pay for the set goal.
Stage 2. Launch Results
First 2 weeks:
Master Campaigns accounted for 31% of all leads
CPL was on average 38% higher than in standard campaigns
By Week 3:
CPA dropped by 3% compared to other campaigns
Share of leads from Master Campaigns rose to 55% of total from Yandex.Direct
Over the next 6 weeks:
Master Campaigns delivered 58–80% of all Direct leads (67% average)
- CPA was on average 55% lower than non-Master Campaigns
Final Lead Distribution Across Campaign Types:
Campaign (conditional name)
Lead Share
«Quiz»
66%
«Call»
17%
«CRM»
17%
The «Quiz» campaign performed best , likely because it generated large volumes of data quickly, giving the algorithm enough to learn from.
Section 05
The Impact of Introducing Master Campaigns
After launching, we confirmed that Master Campaigns are a powerful tool for generating high-quality leads at a lower cost.
Key Outcomes:
10x increase
in the share of leads from "Campaign Master" out of total leads from "Direct"
55% reduction
in customer acquisition cost (CPA) from "Campaign Master"
66%
was the share of leads from the most successful "Quiz" combination
End of archived case
Want a modern case? Hermes data, not Google Analytics.
The modern cases at /cases use a different proof model: per-engine AI mention rate, citation tracking across the answer engines, and delta reports after every scan. Different category, different proof, different era.