Eric Deeran explains why most marketers kill ads too early. Spend 2-3 times your cost-per-lead before calling an ad a failure, not $30-100. Launch with $500 daily budget to get a clear answer in two days, then scale the winner into nine different formats.
I was talking to a client yesterday and he was asking me if he should kill a certain ad set that he has. And so I say, okay, great, let's pull it up. Let's take a look at it. And the ad set that he had had only spent like $30 in it. And he had launched it, I think yesterday. And so I was walking him through this framework, which is pretty well known, but I think it's a good reminder for everybody. When you're running ads or you're running some sort of campaign is you have to spend 2 to 3x KPI, what your average KPI is before you can deem an ad not a winner. And I see so many people waste a ton of money money. And then also just kind of do these half-baked tests on their ads. And they never actually find out what's working and what's not working because they're simply running it for like $30, $40, getting nervous, killing it, starting a new ad. And it's like sometimes more ads is not actually the answer. It's not what you should do. Depending on how much you're spending every single day, I like to launch ads at like $500 a day because then I'll know in 2 days whether or not they're working or not. Right. Um, and when you're looking at your ad campaigns, you have to make sure that you have enough budget. You don't have too many ads, and then you're also not touching things. So if you launch an ad set, my rule of thumb is I always launch it. I let it run. I let it, like, I let it do its thing. And then if I want to change something, I'll duplicate that ad set and then I'll relaunch it with new clean data if I'm changing something, but you have to let something run to 2 to 3x KPI. So like if your average cost per, um, cost per lead is $100, you have to let the ad run, a new ad run for $200 to $300. That's 2 to 3x KPI. And if not, you're just burning the test and you're wasting money because you'll— what people will do is they'll let it run for like $100, not get a lead, kill it, do another one for $100, not get a lead, kill it, do another one for $100, not get a lead. And then you just burned $300 and you have no conclusive data at all. And then when you do find a winning ad, you need to double down on it and then make new formats of that same ad. And so we have 9 different formats we use, but it's like, If you have a talking head ad like this, you need to make that into like a sketchbook format. You need to make that into a green screen format. You need to make that into a podcast format, into a text format, into a carousel format. It's like, I don't think you realize how valuable it is when you find a winning angle to double down on it, because if you don't, then you're just testing a new thing all the time, and you have all these variables running. It's such a sensitive— the algorithm is so sensitive, and like having a funnel that's working is so sensitive. Like, don't mess with it once you find something something that's working. So hopefully, hopefully, um, hopefully it helps. So
