Cost Caps Overspending On Low Spending Ad Account transcript

all right in this video I'm going to walk through an ad account where it appears that cost caps Have overspent and Have once again cost an Advertiser money that they were not willing to spend and therefore the advertiser turned off cost caps at least at some point because of it and this is a really common problem that I hear a lot that people go why why are cost caps spending when they're not performing and so I'm going to show you how I would think about this problem in this ad account so in this ad account um what you can see first is that um the uh date range here is May 1st through June 30th okay so about two months and the total spend this is all the spend um in the account is about 11.5k I'm also going to show you that I have a custom aov column here and the aov is about $211 which means over those two months within the um standard attribution Windows here this is these purchases are in in the attribution windows with the with the settings set by The Advertiser there's only 123 purchases over those two months and this is the first thing I noticed when I open this ad account is that the spend is pretty low and the aov is actually decently High um and so the top spending campaign has 77 purchases over the course of two months now of course if you know uh meta you have heard people talk about the learning phase requiring 50 purchases uh to get out of learning um and and even some people you the old advice at least was like 50 purchases per week is what you wanted at the um adset level now that's not the case for many advertisers and so really this brings up the question of what do you do if you're a smaller account a smaller spending account account how do you deal with this and I think the advice that I'm going to have here is going to be relevant even if you are spending more money but uh that's the first thing I flagged in this account the second thing I flagged is is here in the attribution setting right 7-Day click one day view um The Advertiser who who uh showed me this did tell me that uh that they pay specific attention to 7-Day click um only performance and maybe keep viewing view attribution in there um specifically because they get such a low purchase volume that they wanted to give meta more signal of purchases but they actually pay attention to to click that's reasonable to some degree to me I still think it's going to be confusing especially when you set things like cost caps and bid caps I would still move over to a click-based conversion um and and just run everything on 7-Day click okay now the other thing I'll notice is that over these couple months you have uh let's see how many active campaigns you've got uh let's see campaigns that spent you've got eight okay eight campaigns that have spent over that time and most of them have very few purch now some of that is because there's a Father's Day um bump in this in this for this Advertiser they have they they have some seasonality related to Father's Day and so that definitely is a reason why um they have such a um so many different campaigns they also had a Fourth of July sale um and we're prepping for that at the end of June so there's a couple outstanding circumstances here but this core campaign is an ASC campaign and and the reason I flagged the eight campaigns is because there's too many there's too many campaigns if you are running so few purchases one of the things you need to understand understand Baseline and this is going to be crucial advice for this particular account and for really this is where I think this advice goes broader um Baseline what you need to understand is that um if you want purchase data to dictate future purchase probabilistic forecasting which is what conversion which is what bidc caps cost caps tros campaigns do then you need to consolidate the purchase data into less places okay um they they're at a lower spending account the number one piece of advice I would have for somebody is to try to get uh more purchases and less total adsets or less total campaigns this is where ASC can be a really good tool actually but being able to put 150 ads into one campaign um you you can uh consolidate your learning in one place now there's some downsides to ASC for example it can be hard to separate out different products at different aov levels if you're running lowest cost that's a problem um I think you really want to keep that existing customer cap at 0% with ASC but otherwise it's a really good tool if you actually have um ads that are all for the same product producing basically the same aov it's a really good way to keep the learnings Consolidated into less total places because when you're running a smaller campaign that can be really really helpful so it can be a useful tool for doing those things um okay so um so uh so that's the first thing I'll notice is that you've got this situation here where that's going on now in these in this time period actually um they're they're hitting Target or at least close enough to Target and feeling pretty good okay but then after June July comes around and now in July you've got another 4,400 in total spend all right 11 purchases over the course of the month in July and our CAC has gone way up 276 up from let's see $49 in that campaign okay so that's a really really big difference and if I showed you the cost caps along that time period what you would see is that over time the advertisers lowering the Caps lowering the Caps trying to get meta to spend less without blowing through this but what I want you to realize is that this means that over the course of these 30 day 31 days this Advertiser um is getting one purchase every few days within this window that um that meta has selected that means there's basically no way for meta to realize something really important which is that we are now past this core sale period the same thing is true in August um because it's just not getting enough data okay so if you go to August same thing you've got five purchases in here it's really low CAC goes up even more um and so and now your spend goes down again so the the the the fundamental challenge here is that if anything has changed in customer behavior from Father's Day to later meta cannot know that now this is where I want to back up and think about what how meta's forecasting works okay so there are basically two things that are powering cost caps and bid caps and and that is expected click-through rate which meta can learn from really fast because gets a lot of Impressions and a lot of clicks even on a $2,000 spend in this campaign how many clicks do we have let's see we've got 2,900 clicks so it can learn a lot on 2900 clicks and it can update an expected click-through rate really really quickly so that number is nearly always going to be basically accurate in meta's probabilistic forecasting machine okay but it is also working off of and this is documented by meta an expected conversion rate okay so it's going to convert uh it's going to assume a certain conversion rate and then forecast your cost cap based on that because the only way it can forecast CAC which is what is powering a cost cap right is uh is by forecasting uh the price of a click okay and the price of a purchase and to forecast the price of a purchase it has to forecast how many of those click convert okay so spend click uh CPM clicks and um and purchases are what it forecasts now if you have a moment where your uh conversion rate drastically changes then meta probably can't know that until the conversions come through so it's expected conversion rate is going to be too low or too high based on that for example if you have as I mentioned a Father's Day moment and this brand does and so because this brand has a Father's Day moment and the spend is so low what meta is almost certainly doing here is saying hey we're going to convert plenty of traffic because it saw in June you know 50 some odd purchases or 77 purchases in that time window and it's and it is updating its expected conversion rate as it gets all those purchases in that campaign uh and has some sense of what the conversion rate is going to be but in the two months since then uh the expected conversion rate couldn't have changed that much because it is getting so few purchases to update against right So Meta isn't does can't quote unquote realize that the conversion rate has changed all right so this is the hardest thing in my opinion about buying with bidc caps cost caps tros is when when what you know individually is actually different than what is actually happening for the brand now many people think that's happening all the time and most times it's actually not right so people will say well see Meta doesn't realize this ad is a winner and so it's not scaling it fast enough or meta something like that right um that's actually not what I'm talking about what I'm talking about is something has meaningfully changed based on the offers on your website the seasonal moment or something like that um that is keeping that is uh that is going to not fit with a statistical sample uh that meta has to work from okay that's that is the thing so Father's Day for example you launch a sale you expect suddenly your conversion rate doubles well there's no way based on the previous week or month or however many purchases worth of data meta will know that the next 10 clicks are likely to convert much differently than the previous 10 clicks okay um and but it has to know that for it for it to forecast this so how do you handle that on either side of this I think you do two things one of them is you adjust your budget proactively okay so um if you come out of a sale and you know your conversion rate is going to be worse because you're on the back end of a sale you just bring your budget way down especially if you're very worried about losing money okay if your budget's at 10 grand a day and you're like oh man I think we can't probably spend that much money and have it work right just drop it down to two grand uh and and force meta to do that that will make the budget the pacing mechanism instead of the cost cap okay or the bid cap or the TRS Target secondly the other thing I would do is I would um proactively over adjust the Caps or the or the TRS targets and so um if you think your conversion rate is going to cut in half and you can go study this in your account what happens the day after a sale then go adjust your cost cap uh down by 50% okay so now make force meta to do that now after it gets enough purchases at that point it will learn over over more time what's going on um but it takes some time now this account is going to be really hard to run cost caps with either way because you've got higher aov lower volume of spend it's just going to really tough I would mega super consolidate I would set my cost caps um in this case I'd run bid caps uh click-based only and I would set them extremely you know conservatively if I really couldn't afford to lose money it can be really tricky uh whether it's sale or not just to make sure that everything is is working correctly but um but yeah that's that's the way I would think about this for a smaller account and even for a bigger account um proactively adjust your caps proactively adjust your budgets that's probably the most important thing to do here