Originally posted by d000hg
View Post
Part of my job is to take customer lists of 50k and score the customers to pick out maybe 500-1000 of perfect ones to build the perfect seed audience. Which is an ideal job for a judgmental fooker like me.
If you dont have a customer list use page views to build it from engagement, if you dont have that run a "brand awareness" campaign. Build 5 lookalikes from the one custom audience. 0-2,2-4,4-6,6-8,8-10. Then you can stack them up to get a decent audience size after you have set the radius.
Ideally what you want is a website to have a pixel on, then you can run conversion campaigns, with the "learning event" on "page view". Then the advert will just train itself when people hit the site. As conversion campaigns only ever target 25% of the audience you give it so you go broad with maybe 10% lookalike and let the training take care of the rest.
Always a lot of debate about lookalike size...
The $1,500 Facebook Audience Experiment: 1% vs. 5% vs. 10% Lookalike
For a tenner a week you will get 1500 impressions. Now maybe after 26 weeks you dont get one customer which would be highly unlikely, but by that point 10-15k of people in your area now think you are the Don Corleone of child care as they have seen your adverts for 6 months. So it is all about long term brand building on Facebook which is great for business that rely upon trust.
Leave a comment: