The NEW Facebook Lookalike Audience Approach

0:00hi guys it's ben heath from lead guru and in this video i'm going to talk about a new approach to facebook lookalike audiences a lot of people have been complaining recently that since the ios 14.5 update was released and the changes that have accompanied that have been put into effect they've complained that they haven't seen as good results with their facebook ads look-alike audiences the look-alike audience specifically haven't performed as well as they used to for a lot of facebook advertisers look-alike audiences where our go-to has that changed all that sort of stuff and we've seen some of the same stuff with our own clients and we've

0:35worked out a way to mitigate some of that and help improve performance so i think if you've been advertising to facebook look-alike audiences anyway but particularly if you've seen a bit of a drop off after the ios 14.5 update and this is going to be really useful to um to help improve performance and to to get things back up and running because local accordances are great we don't want to have to stop using them before we're going to a very quickly ask you to smash that like button and of course subscribe to my channel if you haven't done so already um i release facebook advertising related content all the time if you are

1:04into facebook ads wanting good results facebook ads um i would strongly recommend you go ahead and subscribe anyway let's um let's get into it so i'm in an example facebook ad account here and there's a bunch of stuff in here that's just sort of created for demonstration purposes hence why everything's below a thousand um over there and the issue that we're having with lookalike audiences is really around the source audiences more than anything else so we know that the ios 14.5 update has reduced facebook's ability to track user activity that reduces the amount of data that facebook has that reduces the effectiveness of a lot of things that facebook do within

1:44their platform from a machine learning standpoint now look like audiences and if you don't know what lookalike audiences are how to set them up i'll include links to video videos in the description of this video to show you exactly how to do all that i've got tutorials and all that sort of stuff lookalike audiences have been fantastic and still can be i want to make that quite clear but they have been absolutely fantastic because facebook's machine learning around how to take a group of people and closely match those to other facebook and instagram users to help you get the best possible has been amazing so you uploaded a list

2:16a customer list previously and you say these are the people that have bought from me so far go out facebook and find me an audience of people that have bought that are just like the people that have bought from me because they're just the sort of prospects that i want to advertise to those people that are gonna give me great results and that's exactly what's happened and i've been an advocate for many years of creating look-alike audiences based off of in fact let's let's go ahead and and pop this in there so i'll start with the custom audience because that's really where it starts i've been a big fan it's probably my

2:45favorite look-alike audience to be great it's been a look-alike audience based off of this source audience so it's been website uh visitors and then specifically people who have triggered the purchase event okay and we'd normally increase that as much as we can 180 days right and that has been a fantastic um [Music] source audience this isn't a look-alike this is a source audience a custom audience that you create a look-alike audience off the back of this has been a fantastic source audience but as i just mentioned because of the reduced tracking capabilities who is being registered as a purchaser within your facebook ad account and the

3:25number of people that are being registered has decreased so we know that the people that haven't opted in to be tracked who are ios users and have updated to 14.5 which is most ios users at this point um they're not going to be included in this and that reduction in the size of the source audience has led to a less effective look-alike audience and we've known that that relationship has existed for a long time so if someone had a you know 500 people on their customer list and they created a look-alike audience from that or 500 purchases as this sort of custom audience would create and they created a look-alike

4:03audience from that and they completed that they competed with someone um who's got say 5 000 purchases registered and they're using that as their source audience to create their look-alike the person with 5000 purchases registered is nearly always going to perform better that audience the source audience has a lot more data that helps facebook create a much more accurate look-alike audience and that helps improve performance now for a long time there's been this debate around which sort of source audience you should use for your look-alikes like should you go for purchase as this is this would be a source audience for a purchase um based look-alike should you use

4:39something like i mean i haven't got it set up because an example that account but should you use add to cart or something lower down the facebook advertising funnel and what you might have is a situation where you say have you have 200 purchases registered but you have a thousand to add to cast and you're thinking well purchase is better because that's more representative exactly who i want to advertise to but add cart has more data which one's going to perform best and my advice in that sort of scenario and that's something i get asked a lot would be to test both usually thinking that purchase is going

5:07to win particularly as you get more and more conversions but i think that what we're seeing after the ios 14.5 update is that creating look-alike audiences based off of source audiences that are larger is it has increased priority so whereas before let's say six months ago if you had 200 purchases and a thousand add to carts and you're thinking which one should i create a lookalike audience from my recommendation then would have been to test both but probably the look-alike audience based off your 200 purchase is going to perform better now i would still recommend that you test both but the look-alike audience based off the thousand add to carts

5:45might now perform better and that's the that is what i want to discuss in this video and that is what i want to describe and what we're trying is to improve our look-alike audience performance we are increasing the size of the source audience if based on website data and that is proving to be quite effective now of course you could create a look-alike audience of something higher up in the funnel that has even more data perhaps initiate check out perhaps it's a pageview perhaps it's based on all website visitors i'm not just going to give you this is the source audience for your look-alikes because it will depend on your business

6:21the business with 10 000 purchase conversions registered or 10 000 leads registered who is optimizing for either purchases or leads still wants to go ahead with that look-alike audience that's going to perform best so the best one for you is going to depend on your circumstances others to have 150 leads might want to go with all website visitors but i think that the shift that's happened is you know along the continuum of going basing our look-alike audience of people that have done exactly what we want them to do versus basing our lookalike audience off people that have that haven't quite taken all the steps we want them to take but there's a lot

6:55more of these people okay we've just shifted from say here on that continuum back a notch with the ios 14.5 update so i think you want to be using larger source audiences where possible particularly when you're creating your look-alike audiences based off of custom audience website custom audiences so things like um purchase event or website listeners or people who visit specific pages and things like that and i know that the majority of look-alike audiences in my experience at least are based off of website custom audiences that's the most common the other thing that you can look to do and that's that's the first major point around

7:32a new way of creating a facebook lookalike audience is a new approach to it that should help improve performance particularly if you've seen a drop-off the other way that you can can do it is by changing the type of custom audience that you're using as your source audience for your look like audience so if for example you are using this setup so you've got um you're using it based on website custom data anyone who's triggered the purchase event um you've gone for the maximum days allowable which is definitely recommended again in making your source audiences larger you don't want to put this at 30 days or anything

8:07like that like make this as big as you can instead of using this setup instead in fact let's just go back you could upload a customer list and you would most likely have this data somewhere um and you would upload this customer list that would contain the information from the people that have bought from you previously perhaps you have that stored in an email list via mailchimp or whatever and that could be a a different way to target to basically end up with the same results so instead of creating a look-alike audience based off people who have triggered the purchase event pixel uh the purchase event via your um

8:45you know the purchase event that's part of your facebook pixel on your website hopefully i haven't confused you too much there but instead of using that option you could instead upload a list that facebook can then match the people on that list to facebook users to then create a lookalike audience based off the back of those should come out with a similar result now previously i was a fan of doing it the other way so using the website custom audience option because that updates automatically which is great um and saves a lot of um manual work to come in and re-update this it also meant that you would match at a

9:19higher percentage rate so when you upload a customer list which is what i'm now recommending by the way when you upload a customer list facebook has to take that data and match it to facebook users facebook and instagram users they can't always do that so you might see match rates of 40 50 something like that you're not going to match everyone on that list so it's better to do it or based off pixel data before but with these changes the ability to track using pixel data has been curtailed therefore we're leaning more and more towards customer lists and of course another advantage of uploading customer lists and this is more

9:47important if you've been going a while is that you might be able to upload three years worth of customer list data you can't go three years into the past with um facebook pixel data the maximum is just gonna be 180 days so you can potentially have a lot more people on that list and as we now know the thing i want to really get across here is that the larger your source audience so the larger your custom audience is for your source audience the better the resulting look-alike audience will perform because facebook has more data with which to create it and accuracy and data matter a lot to

10:21facebook's machine learning process and these audiences by the way so you can do it based off a customer list and there are some other options you know if i go back you could create um a look-alike audience based on people who have watched video of yours on facebook or interacting with lead forms or whatever you know however your business is set up you could do it based on facebook page engagement or instagram account engagement um which would be really sort of top of the funnel um source audiences for lookalikes always test this stuff you know don't just be like this is the one test a whole bunch test four or five see which

10:50one performs best for your business but what i'm getting across here is that the larger audiences that are higher up the funnel the larger source audiences that higher up the funnel are now performing better than they were previously after the ios 14.5 changes you want to be shifting your sources away from the end up towards higher up the funnel it's the big change that we're making we're seeing that's helpful i think that facebook's basically suffering from a bit of a it's almost like a bit of data it's like fasting it's like a little mini little famine for facebook or on the data stamp standpoint so whatever you can do to add more data

11:26into your setup to help them out is proven beneficial it's why larger retargeting audiences are working as well it's why um broader targeting is doing well for a lot of advertisers opening things up in general is helpful certainly applies to facebook look-alike audiences okay before we go a couple of things i want to quickly mention the first is our facebook advertising services my company lead guru is a facebook specialist facebook and instagram advertising agency we create manage and optimize facebook ad campaigns for our clients if that's something you would be interested in you can book a free call with one of my team members just click on the link in the video

11:58description below you go through to a page and a website book straight into a calendar super simple we do ask that you have at least a 3k uh facebook ad budget that's sort of our minimum budget that we can start working with a company but if you do meet that criteria you can book a call completely free no obligation we can just find out more about your business see if we can help you can find out more about our service um quite casual in in that sense um other thing i want to quickly mention is my facebook ads mastermind group so we've gone past a hundred thousand

12:26members 107 000 members right now my facebook ads mastermind group amazing community full of facebook advertisers just like you completely free to join you can ask questions get them answered i'm in there daily i do live trainings free live trainings links in the description go ahead and join if that sounds like something that you'd like to be part of and if this video has been useful let me know in the comments um i try to get as many comments as i can give this video a like thumbs up that really helps me out and of course subscribe to my youtube channel if you are new and haven't done so already

12:56best of luck with everything thanks a lot guys bye for now