The educational center conducts online courses and webinars in psychology areas


Collect participants for a free-promotional course, which at the end sells a paid one

Marketing budget: RUB 1,200,000
Target: 20,000 registrations in two weeks
Project revenue in the first month: RUB 2,100,000
Project revenue in the first quarter: RUB 3,150,000
Project revenue in the first half of the year: RUB 4,200,000

Solution steps

  1. Researching customer audience by affinity index and finding distinctive attributes for targeting.
  2. Creation of creatives aimed at target audience mirror neurons.
  3. Creation of advertising campaigns on Instagram,, Facebook, Google, Yandex,
  4. Warming up the audience before the course.
  5. Implementing Data Science to Analyze Results.

Mirror neurons

We have abandoned the standard decision to use the speaker's photo for the banner. A brand is created by interacting with a quality product, and the face of an unknown person in the feed causes only irritation and negative recognition.

On social networks, the user is in a very unconscious state when viewing the news feed. Banners were created aimed at capturing attention and transferring it to the desired mental state using mirror neurons.

{Contrast of title and photo, Main emotion is pity}

Results of some campaigns on Instagram and Facebook

{Registration price: ∽55 rubles}

facebook ads

Data science

We implemented tracking of all actions of participants based on the cookie. The resulting data was sent to a machine-learning algorithm (CatBoost) to determine what factors (site pages, events, sources, video views) influenced the purchase.

Such data can optimize the sales funnel and even make a smart website that adapts to the user experience and only offers the right content.

The machine-learning model was able to predict user purchases with 76% accuracy (F1 score)

Sales cohort analysis, LTV and RFM analysis

Sales cohort analysis

Work with the customer base

We conducted a cohort analysis to find quarterly ARPU and calculated the optimal cost of attracting a customer, taking into account future income after the first sale.

We made an RFM analysis to segment users by age, frequency and volume of purchases. This made it possible to find segments requiring special interaction. For example, those who bought a lot in the past, and now have stopped. The received segments were entered into the CRM system.

After analyzing the geography of buyers, we made a decision that resulted in increased revenue of the next project by 800,000 rubles in the first month.




CRM class