Look-alike

Look-alike is a similar audience found by the advertising platform algorithm based on the profile of your existing clients. The algorithm analyzes behavioral, demographic and psychographic characteristics and finds “similar” ones in the general database of the platform.

The source for look-alike is usually: - List of clients from CRM (emails, phone numbers) - Pixel events (Pixel.Purchase, Pixel.Lead) - Subscribers of the TG channel or VK community - Site visitors for the last 30/90 days

The size of look-alikes is usually 1–10% of the total platform base. In VK Ads I use 1% - the most “dense” audience, more expensive, but of better quality. For scaling - 3–5%.

Look-alike not magic. It depends on the quality of the original list. If your clients list contains 50% random people (for example, they downloaded free material but did not buy), look-alike will be “dirty”. I always clear the source: only paying customers + active in the product for 30+ days.

In my projects, look-alike gives a CPL that is 30–50% lower than “cold” targeting based on interests. This is the most powerful tool in Russian performance after the departure of Google.