Depesh Mandalia argues that the single most important metric Facebook uses to evaluate ads is click-related signal quality (link clicks, CTR, outbound clicks). Facebook fully observes on-platform click behavior and user history, but has partial visibility after users leave (ad blockers, pixel issues), so advertisers must send rich post-click signals (low bounce rate, add-to-cart, form starts, bookings, purchases) to demonstrate user value. He describes a grading-like system (higher-status pixel events = better advertiser grade) and recommends optimizing landing-page first impressions, tagging events, and prioritizing signal quality over purely business metrics like AOV or LTV.
Chapters
[DEPESH MANDALIA ]: The number one reason why your Facebook ads are not working might surprise you, but it all comes down to clicks. I know you're going to tell me, but it's about conversions, it's about targeting, it's about creative, it's about funnels, it's about conversion rates. But let me tell you something. As someone who studied the algorithm for 15 years. Facebook relies on click data to determine where your ranked as an advertiser against competitors. It also looks at click data to see if people actually click on an ad and have a good experience. Here's the thing. When you run ads on Facebook, Facebook can track the click. So you have click through rate, you have link clicks, you have outbound clicks. Facebook can see what people are doing with a hundred percent accuracy. So they know who the user is. They what the user's history is, what they've done for different advertisers and different offers, and how they're reacting to your ads. So that's one part of the equation and Facebook can see all of that. The other side is what happens when that user is taken away from Facebook. Facebook doesn't have 100% visibility here because you have ad blockers, you have pixel tracking problems and different things that can impact how Facebook sees your data. They are the two key ingredients that determine your ad costs and the quality of the person you are targeting. So let's say for example, you're after purchases. How does Facebook know whether one purchaser is better for you than another? It looks at that purchaser's history to determine whether the person that's clicking is a good user and also whether you are a good advertiser. Combine those two things together and you get the perfect mix of an advertising campaign that is absolutely going to crush. Now you can look at and say I need to look at my average order value, I need to look at lifetime value, I need to look at cost per click and things like that. Sure they're important, but the click data is the thing that the algorithm looks at to see whether you're serving Facebook's users in the right way. Because they're not yours, they're not your users. You're paying Facebook a fee in order to put your ad in front of that user and in order to get that user to click and visit your site. So Facebook wants to make sure that user having a great experience and they do that by tracking what's happening after they leave Facebook. So this is why it's really important to make sure things like your bounce rate are as low as possible. Bounce rate is someone visiting a website and then exiting. That means they've bounced. They've spent less than 30 seconds or less than 10 seconds, depending on how your analytics are configured. And they've not taken a positive action beyond landing on the page. And that's why it's so important. This is why I spend so much time just on the headline and just on the first screen they see when they land on the landing page. It doesn't matter if it's a long form, landing page, short form, I will spend probably far longer than most people making sure my headline and my visual side of whatever they see when they land on the landing page is as good as possible because I want people to progress. Now, not everyone that's going to click is going to be a perfect customer. So not everyone that's going to click is going to stay on your website and that's absolutely fine. So there's a margin of error here. 20 to 30% of people bouncing is average. It's normal. But I want the 60 or 70% that staying to not just read the page, not just stay longer than 30 seconds, but I want them to take a positive action on that page. And I want to let Facebook know, hey, Facebook, this user had a good experience. Look at me, I'm such a good advertiser. How do we do that? If it's e commerce, getting them to add to car, that's a really crucial step because then that fires another pixel. Facebook can see how long it took someone and how many clicks they've made per session, per day, per week. Think things like that for call booking funnel. It could be someone actually applying to book a call or just starting the application form. This is why I often tag things and my clients are like, why are you tagging that for? Partly it's for testing and that's something I would cover in a separate video, but it's partly to give Facebook more and more signals. Signals are everything when it comes to Facebook ad performance. I know for you, you want revenue, you want profit, you want lifetime value, you want average order value. Facebook honestly doesn't care about that. All they care about is has this user had a good experience. And that's determined by them clicking your ad, which means your ad was relevant, and then doing multiple conversion points on your website. Pixel events, which are maybe not conversion events, like I said, starting a Typeform survey, starting a survey, starting an application form, even getting to the call schedule page. They're things that I'm sending back to Facebook as data because Facebook can Stitch the journey together and then my important conversion points. It could be a purchase, so I'm optimizing my ad sets for purchase, or it could be completing an application, or it could be booking a call. They're super important points back to Facebook. They're like the highest status. So if I say I'm optimizing for purchase and a purchase happens and Facebook can see it, great. Facebook is going to give you an A plus grade. If Facebook sees that they added to cart, they went to checkout, but they didn't purchase, maybe you get a B, maybe a B minus. If they got to add to cart, maybe you get a C. If they went to the product page and didn't do anything, you'll get an E or a U. So there's a grading system that goes on. So when we're optimizing our Facebook ads, we are concerned about clicks and we're concerned about the quality of the clicks. And that's where I then look at all those events being passed back and make sure that's post performing as well as it can. Then secondary is average order value, is lifetime value, it's revenue, it's profitability. There are things that matter to you that matter less to Facebook. Sure, Facebook want you to make a profit. Of course you make profit. You can spend more with Facebook, but the algorithm, the thing that you're advertising against and with, doesn't necessarily always take that into account. Now, I know some of you will say, but we can optimize for roas, we can optimize for value and things like that. You can, but the underlying data beyond that is click data. So this is why I think about five or six years ago, Facebook were heavy on signals. Signals was all they were talking about. If you spoke to a Facebook rep, if you downloaded a white paper, you're attending an online or in person event, they were heavily talking about signals. And I think most people ignored that. Or maybe with iOS 14 and some of the problems of tracking, just thought that was less relevant. It's not. It's still as relevant as ever. So if you want to make Facebook ads work for any niche, focus on click data and quality signals, back to Facebook.
