Lookalike audience

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search

A lookalike audience is a group of social network members who are determined as sharing characteristics with another group of members.[1] In the digital advertising age, it refers to a new targeting tool for digital marketing, first initiated by Facebook, which helps to reach potential customers online who are likely to share similar interests and behaviors with existing customers.[2] Since Facebook debuted this feature in 2013, additional advertising platforms have followed suit, including Google Ads,[3] Outbrain,[4] Taboola,[5] LinkedIn Ads[6] and others.

Considerations[edit]

Lookalike audiences anatomize existing customers and their user profiles to find the commonalities between the existing audience. This helps to find highly-qualified customers who previously would have been difficult to identify and reach.[7] This expands the potential audience in different countries and applies to new differentiated audience segments;[8] This approach saves time and lowers advertising costs for the acquisition of a new audience.

In order to be effective,[9] a lookalike audience seed needs to be homogeneous. This is commonly achieved using a consistent behavioral pattern. The homogeneity of the lookalike seed has a greater influence on the audience's effectiveness than the size of this sample group. In Facebook, the minimal lookalike seed size is 100 users from the same country.[10] Facebook generally recommends creating a seed from an audience of 1,000 to 50,000 users.[10]

Lookalike audiences might have limited effects on small companies or startups because of the small sample size of their existing audience, which would inevitably lead to insufficient data drawn from the current audience and interference from outliers. Namely, there would be no high bounce rate with these companies' websites.[11]

Examples of seeds[edit]

Marketers use many data sources to create lookalike seeds. Some examples of eCommerce lookalike seeds include:[12]

  • CRM-based – A seed based on an email or phone number list of customers who have had a past interaction with the business. This can be further segmented, for example customers with the highest lifetime value or past purchases of a specific product.
  • Conversion-based – A seed based on users that have performed an action such as a Purchase or Lead form submission on the website.
  • Engagement-based – A seed based on users segmented by their engagement, such as pages viewed, time spent on the site, video views, etc.[13]

Methodology[edit]

Facebook, as an example, takes three steps to build a lookalike audience:[14]

  • Choose the audience seed to build a lookalike audience from. This can range from page fans, visitors to the website, and customer lists etc. Generally the base audience should be composed of a minimum of 500 people. Larger pools will increase the accuracy of the lookalike audience.
  • Choose the specific location (country or region) to find a similar audience in.
  • Customize the audience size. Facebook offers a range of percentiles from 1% to 10%, indicating the size of the combined population of the locations selected. Larger audiences provide a wider reach, but a smaller lookalike audience is more targeted, which means ads are seen by fewer people, but they are likely to be better aligned to the features of the audience's seed.

Debate[edit]

One study has shown that the tool of lookalike audiences, to some degrees, does well in generally advertising results.[15] It is also listed as an important trend of pay-per-click (PPC) by Delhi School of Internet Marketing.[16] However, debates over such a third party behavioral targeting being used for digital marketing hasn't stopped either, because using the data of customers is against online privacy settings.[17]

In 2019, limitations were put in place by Facebook to stop discriminatory targeting of audiences according to zip code, income levels and demographics (age and gender).[18]

References[edit]

  1. ^ "CUSTOMER ON FACEBOOK - NCMA" (PDF). google.com.hk. Retrieved 18 March 2018.
  2. ^ "How to Use Facebook Lookalike Audiences | WordStream". www.wordstream.com. Retrieved 18 March 2018.
  3. ^ "About similar audiences for Search". Retrieved August 8, 2019.
  4. ^ "Look-a-like Audiences Solution For Advertisers". Retrieved August 8, 2019.
  5. ^ "Lookalike Targeting". Retrieved August 8, 2019.
  6. ^ "Targeting with LinkedIn Lookalike Audiences – Overview". Retrieved August 8, 2019.
  7. ^ "What's a Facebook lookalike audience and why is it important?". Bigcommerce. Retrieved 18 March 2018.
  8. ^ "The Power of Lookalike Audiences". Online Advertising School. 5 December 2016. Retrieved 18 March 2018.
  9. ^ Shpivak, Etgar (June 6, 2019). "How to Best Scale Lookalike Audiences". Kenshoo. Retrieved August 8, 2019.
  10. ^ a b "About Lookalike Audiences". Retrieved August 8, 2019.
  11. ^ "Lookalike Audiences: Why You Can't Ignore Them - South Agency". South Agency. 24 April 2017. Retrieved 18 March 2018.
  12. ^ Levy, Elad (April 25, 2019). "7 eCommerce Lookalike Audiences That Are Worth Testing". Ladder. Retrieved August 8, 2019.
  13. ^ Basis, Ehud (October 29, 2018). "Scaling Paid Campaigns via User Engagement Signals". Outbrain. Retrieved August 8, 2019.
  14. ^ "Ticketfly Community". community.ticketfly.com. Retrieved 18 March 2018.
  15. ^ "Advantages of WCA Facebook advertising with analysis and comparison of efficiency to classic Facebook advertising" (PDF). google scholar. Retrieved 18 March 2018.
  16. ^ "DSIM- Digital Marketing Blog". Digital Marketing Blog - DSIM. Retrieved 18 March 2018.
  17. ^ "Facebook Custom Audience Terms of Service: Are You Breaking the Rules? - Jon Loomer Digital". Jon Loomer Digital. 31 October 2013. Retrieved 18 March 2018.
  18. ^ "Facebook removes age, gender and ZIP code targeting for housing, employment, credit ads". Retrieved August 8, 2019.