Lookalike audiences strategy for Indian markets

Lookalike audiences are one of the most consistently effective tools in Indian digital advertising. The concept is straightforward: you give Meta (or Google) a list of your best customers, and the platform finds new people who look like them. In practice, the effectiveness of a lookalike audience depends entirely on the quality of the seed audience you start with.

Building the right seed audience

The seed audience is the list of people your lookalike is based on. A lookalike of your purchasers is better than a lookalike of your website visitors, because purchasers are more qualified. A lookalike of your high-value purchasers (people who bought above a certain value) is better still.

For Indian e-commerce businesses, the best seed audiences are: customers who have placed orders in the last 90 to 180 days, customers who placed repeat orders, and customers above a specific order value threshold (e.g., orders above ₹2,000 if your average order value is ₹800). Upload phone numbers for Indian audiences rather than emails; phone number match rates are higher in India because mobile numbers are more consistently used as primary identifiers.

For lead-generation businesses (education, real estate, financial services), use converted leads (people who became paying customers) as your seed, not all leads. Using all leads in your seed produces a lookalike of people who fill in forms, not people who buy.

Minimum seed size for a meaningful lookalike is around 100 to 1,000 people on Meta. More is better. Below 100, Meta does not have enough signal to create an accurate lookalike. Between 1,000 and 10,000 is a good range for most Indian businesses.

Choosing the right lookalike percentage

Meta allows you to create lookalikes ranging from 1% to 10% of the target country's population. In India, the total Facebook/Instagram audience is roughly 300 to 400 million people. A 1% lookalike in India represents 3 to 4 million people. A 5% lookalike represents 15 to 20 million.

1% lookalikes are the most similar to your seed audience but have smaller reach. They tend to perform best for direct-response campaigns where conversion is the goal. Start here.

3% to 5% lookalikes give you more scale at slightly lower precision. Use them when you have exhausted 1% audience capacity or are running brand awareness campaigns.

Above 5%, the similarity to your seed audience weakens significantly. I rarely recommend 7% to 10% lookalikes for Indian campaigns unless the goal is very broad awareness.

Layering lookalikes with other targeting

Do not run lookalike audiences completely unconstrained. Adding a few guardrails improves quality.

Age and gender filters are appropriate if your product has a clear customer profile. A fitness supplement primarily bought by men aged 22 to 40 should have that filter applied to the lookalike campaign.

Location layering is important in India. A 1% India lookalike of your best customers might still include a broad geographic spread. If your fulfilment covers only certain states, or if your product is specifically relevant to certain regions, add location targeting.

Interest overlays can tighten a broad lookalike, but be careful not to over-restrict. If a lookalike of 3 million people is narrowed to 200,000 with heavy interest layering, you limit delivery and may hurt performance. Test with and without interest overlays.

Testing and refreshing lookalikes

Seed audiences age. Customers from 18 months ago may not represent your current best customers. Refresh your lookalike seed audiences every 60 to 90 days by uploading an updated customer list. This keeps your lookalike aligned with your current customer profile.

Test multiple seed audiences simultaneously in separate ad sets. A lookalike of all purchasers versus a lookalike of repeat purchasers versus a lookalike of high-value purchasers might perform very differently. The data from these tests tells you which customer type Meta's algorithm finds most efficiently.

Frequently asked questions