Meta Ads New Strategy: L0 Approach

[00:00:00] so Amazon did better new subscribers did better enac did better new sales on new customers at chopi did better then Google and then have all of our inapp me inapp metrics according to Wicked reports and then all the detail in Wicked reports so we look at everything so we don't look inapp inapp you know doesn't matter the L zeros this one that you're seeing here this is our creative testing this actually has now been migrated over to our main campaign as of the first and this l0o campaign essentially is a CBO open Target customers excluded no bidder cost cap maximize conversions and it is using five duplications of an ad at $100 per day each and what we are identifying is the average of the first click a cruel 7-Day look back window andca in comparisons and that's how we're judging them so we're looking at the last seven days for example in Facebook uh we see lzos LZ spent 16 Grand you'll see that we have you know bar a damn it

[00:01:22] you see we have Barber Barber Barber barar barbar we have five barbaras uh let me pull up the spreadsheets that we're using uh in order to um review this give me one second Google Sheets there we are um I remember which one which okay yeah so here we go so it's all the basic same one so this is being built from a mean this morning but this is after four days so what we're looking at is a collective Claudia has a $300. 199 endak with five

[00:02:08] different ads running at $100 a day each it's only been running four days as you can see here um so we we basically are kind of jumping the gun a little bit this one only has you know one two three four and today's the fifth day so after four days we have it's above a400 $300 endak it gets paused anything that is low at $300 enat gets cloned so cloned meaning that five Dian turn into 10 Dian five lores turn into 10 lores five suzan turned into 10 suzan and that is we're using one week uh per campaign so we have let's say we have five new ads we start a campaign it gets dated the date that it launches and then those five are then now stuck in that one campaign forever and as they progress we find that there is you know three winners one loser you can see this this is a new campaign that just started yesterday sorry yesterday so on day one we have $500 and zero sales we have $500 and two sales for an ncac of $245 $500

[00:03:11] one sale for 503 $500 one sale and 509 so once obviously a week goes by if these are still above 300 they get paused if these are below 300 they get cloned and then we put five new creatives and a new lzo and what that's doing is allowing us let me grab the other report here here history funnels so what we've been measuring is we've been doing this for like the last three weeks and it's been kind of you know hit or miss with our with our consistency we really haven't hammered out now what we want to do but our spend went from uh we have 164 up from 145 so we added about $16,000 this last three weeks in spend our Revenue went up to 341 it was 315 so scaled Revenue our media efficiency ratio is at 6.22 Collective it was at 6.5 so it did go down a media efficiency ratio however we had $13 uh CPA and a 98 so only $5 higher CPA for 15K more in spend okay good so what we're then looking at is the

[00:04:16] difference between L Zer and uh cost for new customer Top Line so this is our inapp um L zero enca and how they're going to be correlated right now there's really no correlation just yet that's what we're going to be tracking a 342 and a 34 2 is the difference between 2166 and 227 however that went down and that went down you know compar of that but not compared of that so it's kind of like weird like you can see they're all kind of hovering so what we're trying to do is okay get our L Zer enac below 300 and that's where we're going to be trying to push to to see does that get this down to under a $200 Global endak because under $200 Global endak everything looks really good $30 CPA 2.4 M Etc um Glenn what's up hi John long time listen to first time caller oh very nice to meet you Glenn where you calling from uh Australia austral I really thought that you were actually from Antarctica but go

[00:05:06] on um this is something that crossed my mind on Friday as well you're doing five yeah and then you're scaling at any point in time let's say you've got your five ad groups just say two are going really well do you keep at any point in time or do all five collectively have to be performing or if two are doing well would you kill three and scale the two that were doing well nope it's an average of five okay it's always an average of five yeah because average of five yep so you're doing groups of five so the five collectively have to perform yeah and if they're not you kill them right and that's that's the reason why we're doing this is because if you look at like the uh this one here the 915 you'll find that in the 915 one which is this one right there you have like Claudia 399 and when you look at Claudia oh oops CL Claudia you have one good one three at 133 and then you have fail fail fail fail yeah okay so it's like okay

[00:06:22] they're all 62 to 65% cold traffic so you see this is why this is paused and that's pause there and they're all exactly the same there's no exactly no ai's turn AI Creations turned off like very everything is completely identical because what's funny is once you find something that doesn't work it really doesn't work then you look at Diane Diane check this out Diane has 17 z34 so it's a lose massive win lose win win if I clone these five more times I'm expecting the average to also pass but if I clone you know five of the failing ones I'm also expecting that to to fail so we use a collective five of five so we get basically five data points of seven days of $100 per day each so it's a massive spend are you finding that doing this duplication of five is and it makes sense that it's increasing your frequency on these ads it's increasing the frequency Collective to an audience both new and returning so

[00:07:39] and they're all the same audience aren't they it's all broad yes yeah yeah yeah our l1s were only 40 percentile cold these have been in the mid-60s percentile cold so this is actually also having effect on our Amazon um and that's what you have to look at too is like our Shopify we're seeing new customer 703 this one 607 so Shopify new customer count went up from a 607 up to 73 so it went up almost 100 and then we looked at recharge and our active subscriptions are at 7470 before there were 7191 so we have 300 new subscribers too and then we look at Amazon here and our sales are 2564 they were 2313 so they're up 200 so Amazon did better new subscribers did better enac did better new sales on new customers at shopy did better than Google and then we have all of our inapp me inapp metrics according to Wicked reports and then all the detail in Wicked reports so we look at everything so we don't look inapp inaps you know

[00:08:47] doesn't matter if our Top Line is heading in the right direction which is yes there's more Revenue there's more sales there's better CPA Shopify is doing better with new customers recharge is doing better with new subscriptions Amazon is doing better with new sales and what we identified is that if it's at you know 300 or below everything looks the best so that's why we then chose a $300 like action steps today this one is pause and clone clone clone so we did is we turned we went from 510 1520 now we have neg five plus five plus five plus five all right so so we had 20 then we went down to 15 and then we went from 15 to 20 to 25 to 30 so went from 20 to 30 in our spend so we actually increase spend by 50% you're just doing this on a 7day rolling window yes sir sexy right oh my God I'm glowing you're on the face of the sun I'm ascending John you're going really old school horizontal scaling kind of thing

[00:09:57] isn't it in terms of like to a person found a whole bunch of gray hairs and being bald like I am old so that's why I'm going in Old School you want to compare who more gray hair these days I would yeah actually I would spank you but no I like it I like it it's very yeah it's pretty cool right yeah it's definitely going really old school like we're going like 2018 status but it's working here's what's funny it's like used to run ads really really well with high touch good strategies and then meta took over I was like don't worry I got it from here and then they pmax the thing this is just the standard shopping version of meta yeah so question because you're doing horizontal where you're just stuff have you tried increasing budgets on an at like just atet level what you're doing or is it have you done oh yeah we tried everything you could possibly imagine increasing spend fail cloning it into L1 while keeping it into

[00:10:49] L zero failed L1 succeeded L zero everything that we've tried has failed except for cloning the existing one the previous test that we ran we're cloning it at 225 per day five at $225 it did well but it was way too hard on the system so it it it took a while to come back down each time so we found that five at 100 was a good amount for this account to get a sale immediately within the first day at a $500 endak and then it either got worse or better after that okay so we also have a few rules if it's between if it after four days if the group of five are collectively above a 450 it gets paused if it's between three and 400 in the first four days you give it more time because it was going to either get to get stay the same get worse or get better by day seven but then day seven comes around and anything below $300 that has been running for seven days media gets cloned anything above $300 after 7 Days gets immediately pause

[00:11:47] because it could have been like 425 down to 350 that's okay I have 15 other ads I need to clone so I need your ad spend that's anytime we add new ad spend or sorry anytime we add new ads as creative testing it's always incremental our paused ads are that ad spend is used for our cloning and then we add on new you add it in the same campaign or do you build out a new campaign when you're testing like if you're trying to switch up a headline or a description to test is it in the same campaign every week new campaign every week new campaign every week new campaign five ads clone five times $100 a day each okay and then are you just testing one thing at a time in terms of like Crea what that means or something like that you just sorry broke up there say say over time um are you just testing one thing at a time and when I like are you testing like one headline and then or like testing different headlines same

[00:12:44] same image something like that are you just duplicating whatever is working and just contining contining are you test are you testing new creative how are you do that yeah so the creative is is it's always like kind of creative from scratch but we're not changing any one of our ads um so for example like when we take an ad here this is a new ad so it's uh like single image multi- Advertiser ads this is the media this is the primary text headline description I mean it's really old school yeah no what so you're keeping all that constant you're not changing that any AD you're just going for whatever ad is working duplicate duplicate duplicate you're not testing new creative or different ad copies and stuff like that no if it works it gets cloned and if uh if it doesn't we add new ones right okay it's only 60 it's only 64% cold John explained it to me instead of scaling like vertically scaling

[00:14:00] yeah horizontal scale I get it I get it because I know I was watching a training last no a couple weeks ago where they're like they do the same thing you're doing but what they do is they they'll take like one headline and then they'll play adjust a headline and they'll only change the headline and they'll keep everything else constant and then they'll change like the description like the ad copy up top and they'll keep everything else constant and then they find essentially what headline works the best in for this um audience what the description Works type of description description work the best what typee of images works the best and then they create variants of that so they'll have like if they have like two um like let's just say two headlines two ad description and two images they have six variations of it in total because they get all the variations combined and then it and depending on how it performs

[00:14:46] they duplicate increase decrease something like that that's yeah the problem is I don't know what their warm or cold traffic is and that's that's the bad part so the way that we're looking at this is one day look back window click attributed a cruel first click so this is and the reason being is that I was we everything inside of meta shifts every other day things that were suck are now amazing things that are amazing Now suck and that's because it's difference between engage views and then click and then it's first click and second click and third click and fourth click and y y so the only way that I know to clone this in a growth ecosystem that's been proven for last three weeks is looking at it on a or sorry a sday look back window only click attributed only measured by NCL enak and only measured by new customers um if you take anything else in consideration I can't tell you if that was a second third fifth or 12th

[00:15:40] click from a person that's new or returning and that's that's part where we did that for about a month and we end up just chasing our tail in circles because we're like well that one used to do good what happened I don't know went from 180 to 740 yeah but it didn't affect top and we were literally just spinning in circles and so we were like pause you revive now pause and revive the old ones and so I'm like we have to look at this in a if I was to take this ad and triple it with 25 you know $25,000 more per month ad spend I have to know that if it's going to be cloned to cold traffic it's going to perform as good as it did before and that's why we had five uh five kind of you know cold traffic you know low frequency it's funny they're all going to like 65% cult trit they're all getting like 1.11 frequency like everything looks very even so when we cloned it it got better so like okay good at least I can prove that it's it's

[00:16:32] static to cold traffic fair enough I like it so it depends on how you measure it yep yeah measure it by how much money in my bank account that's how I measure it well it's funny because like Amazon all of a sudden had a shitty week but everything else looked good so we're like okay let's just keep going and all a su the next week it came right back we're like okay good like that returning and what meta says all wasn't true like we have to do this by Top by metrics period but the inapp is actually not measured inapp is measured by third party and so what we're trying to do here I don't know if my settings are still up but yeah so this is Click date first click only not conversion date this is you know conversion by by click this conversion by time so it defaults the conversion date so I change it to click date first click only and that's when I can look at okay you know at least at least this is if this is 70% accurate

[00:17:33] if I have five data sets Okay I'm averaging out I started looking at first clicks only too pretty much now I don't even look at the other ones because I'm just trying to get traffic in and see how much that cost me anything else just is a blur to me in terms of like I don't trust it at all first place there's still some truth to it like it's someone's first time out the website but even then fre assault yeah yeah and that's the thing is like we're never going to be able to get you know completely uh accurate but you know as least is if what I'm trying to do is just say like if I can hand this off to you know an employee of pet wellness and just say like here's your rules here's your structure you know add $25,000 a monthly Aspen every week this is when you pause this is when you starts so if I can establish these rules at scale oh it's Off to the Races yeah I like it about 11 months ago that's when I posted

[00:18:35] I'm like why I'm moving almost all my clients off pmax um and then it's funny is like then like nine months ago um that was like what we shot out pmax which standard shopping that was n month I mean almost on the same timeline as you're were like hey this thing starting to become less valuable the same thing that I saw