Making better quality lookalikes on Facebook!5 min read
What are lookalikes?
Facebook marketing has grown immensely, with multitudinous advertisements taking the scope to millions of users. Marketers have dwelled on the concept of facebook lookalikes for the purpose of better ad campaigns and with good reason. What’s the use of filling up the feed of irrelevant users with ad information when we can efficiently sift the masses for the right people. This what lookalikes are all about–getting to the right people! Marketers must select a custom audience based on their page activity by looking into the hits on pages similar to the one they plan to host.
Now, custom audiences are those groups of people who have interacted with your business previously. Use this data to generate lookalike audiences who are similar to your current customers but have interacted with other businesses. In doing so, you basically seek to attract this group of people.
Lookalikes on Facebook:
What Facebook does is it takes your seed audiences (custom audiences, here) and tries to find common characteristics in them. Now, once these characteristics have been discerned, it tries to find a pool of audience comprising of people who have similar traits as per the marketing requirements. The measure of the quality of your seed audience is key to a good lookalike.
Before venturing into the details of using lookalikes in the market, let us look into the essence of a good lookalike:
When we have a ready custom audience set, Facebook allows us to choose how close or similar we want our lookalike to be from it. In fact, there is a scale from 1% to 10% that allows us to choose. Now, suppose our custom audience is from the same region where we wish to find our lookalike, we can afford to close in our lookalike to 1% or 2% similarity owing to the precedence of our custom audience. However, if we take custom statistics from a foreign land, we need to ensure sufficient leeway of up to 10% to ensure a feasible lookalike.
Following is an in-depth approach to make better quality lookalikes on Facebook with the available resources. Keys to a good lookalike-following are the quintessential questions that need to be asked.
1. Where are your viewers located?
The basic custom audience that caters to the necessities of the marketers includes people from different countries. In order to increase your reach, you must be careful while selecting the range of audiences: for local audiences, the range of closeness for the lookalike from your desired product can be more relaxed. However, if one takes a broad sample set, say, from a variety of countries, you need to stick to being as close to your requirements as possible, which means less divergence. This means you need to take the pick of the lot when it comes to custom audiences.
2. Are they spending enough time on the website?
Now, the art of making good lookalikes lies in incorporating the right parameters while selecting the custom audiences. We take into account select crowds based on their activity on various websites similar to the one we wish to promote. This activity, however, varies. We have people who casually browse through websites, especially during particular free hours of the day. This is when the website traffic is on the rise. If you bank solely on this, it would make for a poor lookalike as the trivial users will not be filtered out. Hence, you must first impose the ‘time-spent’ indicator so as to accurately deduce and segregate the interested users from the browsers. This is key to discern the lookalike audience.
A common practice is to select the visitors over a short span of time if the hits on the page are high. However, you would be amiss big time if you do not filter the time spent by people on the page. The casual browsers are only as good as the no-shows. Hence, they must be filtered out.
3. How to apply our indicators?
Checking all the indicators in the same sample audience is another flaw. Separate indicators must be used against different qualifiers. For instance, the time-spending, money-spending, and the population density filters can be applied at the same time, but that will not help you gauge the success of any one indicator. Hence, to understand which feature creates a better-suited lookalike audience, it is important we check all indicators against separate sample sets. Alterations can be made with the reach of the advertisement under each sample set, and hence, it is important we do not merge all indicators into one sample set so that we can keep track of the various changes.
Hence, finding the people who spend a lot of time on your website, people who actually spend on lookalike products, and generating a suitable audience to create a group of lookalike audiences is the secret of a healthy Facebook marketing campaign.
Customer Lifetime Value (CLV or LTV)
Among the indicators, should you wish your lookalike to be value-based, be sure to include customer value as a working factor. This helps you estimate your net gains out a single customer by looking into simple figures (or their estimates) such as the frequency of a specific customer’s purchases, net quantity they spend, and the expected duration for which your relation runs. Hence, a sufficiently clear idea of whether a customer is profitable for your business or not.
Best Practices for formatting customer data:
Create separate columns for uploading the phone numbers and email addresses so that Facebook considers both as separate functions which gets more data for sifting. Including the country code along with the phone number is helpful. More data types benefit the cause of the marketer. Similarly, since we are using multiple data types, writing first names and last names in separate columns helps. Also, it is best to include the name of the country even if you have taken your custom from the same country.
NOTE: With these pointers in mind, remember, a local custom audience based on which a 1% lookalike is made is likely to yield healthy results. Hence, try to stick to domestic populations in order to gauge the viability of your audience in the best possible way.