Ep 108 - Demystifying the LinkedIn Ads Auction and How to Get Better Performance
SHOW RESOURCES
Here were the resources we covered in the episode:
Contact us at Podcast@B2Linked.com with any questions, suggestions, corrections!
A great no-cost way to support us: Rate/Review!
SHOW TRANSCRIPT
$8.75. Do I hear $8.75? $8.75? Sold to the advertiser with a high relevancy score. You guessed it, we’re talking about the ads auction on this week’s episode of the LinkedIn Ads Show.
Welcome to the LinkedIn Ads Show. Here’s your host, AJ Wilcox.
Hey there LinkedIn Ads fanatics! So I’m not going to leave my day job to become an auctioneer. You’re welcome. Alright, longtime listeners will know that we talk about the ad auction on LinkedIn a lot. It’s a bit more technical of a topic. But a deep understanding of the underpinnings on how the platform serves your ad content and how it charges you is absolutely crucial to becoming an expert. So by the end of this episode, you should understand why we make the recommendations we so often make things like bidding manually first, or bidding lower than the suggested bids, or only bidding maximum delivery when your CTR is significantly higher than the benchmark. And for the key takeaways here. I just want to reinforce the strategies and claims that you’re probably using that you’ve gotten from us. And these are claims that we’ve made for as long as we’ve been in business, We want you to be able to decrease your costs on LinkedIn ads, which is a notoriously expensive platform to begin with. And we want to help you optimize and get better performance on a platform that is really quite difficult to master. First, I want to highlight a review that came in BitE777 from South Africa says, “Best podcast for LinkedIn practitioners. AJ puts a lot of effort in educating hard to understand concepts. He asks his guests, who make impact themselves, the right questions to really push value. I enjoy working through these podcasts. And it’s an absolute no brainer for those who like to run better ads on LinkedIn for their clients or companies. Thanks, AJ and team”. With a little fire emoji. BitE777, I don’t know who you are, but thank you so much for leaving such an awesome review. That’s exactly what we’re trying to do here on the LinkedIn ads show podcast. And you regular listeners, yes, you! I want to feature you here as well. So please go and pay me the biggest favor that you possibly can and go and leave a review specifically on Apple podcasts right now. And of course, I’ll shout you out live as well.
Targeting
First off, we need to talk about how the ad auction works. Now the resource that I’m actually sharing from in order to create this episode is an internal LinkedIn resource, but I haven’t been given permission to share it externally. And this is an awesome resource. And it’s one that I haven’t seen before. And it really enlightened a lot for me. So the ad auction works in four different steps. The first step is targeting. The way it works is when a member on LinkedIn opens up their newsfeed, LinkedIn take stock of all the advertisers who are wanting to target that individual by their targeting facets, because it knows they’re about to be competing for being able to show to this individual. That’s step one.
Bidding
Step two is bidding. So going into this I had a question that was if you are bidding manually. Are you in a different auction from those who are bidding automated? Or put another way? If you’re bidding by CPM? Does that give you an advantage over those that are bidding by CPC? And what I came to understand here, and I specifically wanted to know how do both of these engage in the auction together because we know that CPC advertisers are really held to a standard of needing a high relevancy score in order to win auctions. But it didn’t seem like CPM advertisers were held to the same accountability. In fact, LinkedIn is even told us through official representatives, that CPM advertisers don’t actually use relevancy score in the auction. So I needed to understand how CPC and CPM is actually handled in the auction. So here’s the beautiful thing. LinkedIn says in order for all bid values to operate on the same value system within the auction, all bids from all types of bidding that you can do are converted to a metric called ECPI, or, or estimated cost per impression. And this makes perfect sense to me. If you have CPC advertisers and CPM advertisers running on different systems, at some point LinkedIn has to synchronize those and decide who comes up first, and who comes up second. But by using simple formulas. Of course, if you’re good at algebra, these are simple. You can convert both your CPC bid and your CPM and your cost per view. If you’re doing video advertising to this one single metric called estimated cost per impression. We’re not going to dwell a whole ton on formulas because quite honestly, it’s really difficult to listen to a formula over audio and synthesize it but I am going to tell you what makes up all three of these formulas. The first is in order to convert cost per click into estimated cost per impression. You take your predicted click through rate, which if you’ve already been running something you could say it’s your average click through rate and you multiply it by your bid. And it’s just that simple. Click Through Rate times bid that That’s gonna give you your effective, that’s gonna give you your estimated cost per impression, then because we’re trying to get it into a cost per impression, you can imagine that bidding by cost per impression or cost per 1000 impressions is actually really easy to convert. And you would be right. The formula to convert cost per impression or cost per 1000 impressions is your CPM bid divided by 1000. Done, that’s it. It really is, you take your cost per 1000 impressions, but you want to see what your cost per single impression is. So you divide that by 1000. Easy peasy. Cost per view, if you’re doing video ads is a little bit different. You take your predicted cost per view, or your average cost per view, and you multiply that times your bid. And that gives you an estimated cost per impression there. And what I found fascinating, I’d never heard this ever before, when LinkedIn holds an auction for a member, the auction actually determined the top 20 slots. I think I was thinking in my mind that was like, oh, they’re likely to see, let’s say, one to five ads. So maybe you do an auction for the top three. And then as they keep scrolling, maybe it holds more auctions, but no. LinkedIn does 20 for the newsfeed, I don’t know what the auction is like for the other ad formats. But I did notice as I scrolled every 10 posts in the newsfeed as I scroll down, I would notice that the right rail updated and showed something new. So I wouldn’t be surprised if right rail was something like I don’t know, maybe it’s 20 as well. But I could see it being something more like 10. And then as someone scrolls every 10 posts, it refreshes and includes whoever came next in the auction. And of course, as we’re talking about the auction, you need to understand what kind of auction this is. LinkedIn claims it’s a second price auction, meaning you don’t actually pay what it is you bid, you only pay one cent more than the person you beat out to get there. Or in other words, you only pay one cent more than what the second place bidder would have paid if they would have won. Now there is an exception to this. And that is floor pricing. If you’ve been listening for a while, you know that the newsfeed doesn’t have a floor anymore, or at least it’s not one that is stated anywhere. But it used to enforce it. When you put a bid that was too low. LinkedIn used to say, hey, here is the amount that you have to pay. If you want to compete in this auction, that was the floor price. But we do know that text ads still have a floor that’s $2. Or if you’re in a different currency, it’s the $2 equivalent. And conversation ads may have a floor as well. So this is why if you bid the floor price, you’re going to pay the floor price. But if you bid anywhere above the floor, like let’s say I bid $12 per click, if I look at my actual costs per click, they might be like $11.60, or $11.80. That’s the second price auction at work. So what’s happening in this bidding section is LinkedIn is taking all of the different kinds of bids. And instead of comparing them as apples to oranges, it’s making it so all of them are compared apples to apples. It’s a brilliant system. It really is. So what I did is I just inside of Excel, I just made simple graphs based off different click through rates and different bids, and I wanted to understand how the auction would deal with all these.
8:07
Alright, here’s a quick sponsor break and then we’ll dive right back into the process.
The LinkedIn Ads Show is proudly brought to you by B2linked.com, the LinkedIn Ads experts.
Managing LinkedIn ads is a massive time and money investment. Do you want to return? Consider booking a discovery call with B2linked, the original LinkedIn Ads performance agency. We’ve worked with some of the largest accounts over the past 12 years. And our unique scientific approach to ads management, combined with our proprietary tools, allow us to confidently optimize and scale your LinkedIn Ads faster and more efficiently than any other agency, in house team, or digital ads hire. Plus, we’re official LinkedIn partners. Reach out to us at B2linked.com/apply and we’d absolutely love the chance to get to work together.
Let’s jump right back in. I don’t want to read you items off of a chart. So I’m just going to try to give you the takeaways. I ran this imaginary auction based off of average values that we oftentimes see from advertising on LinkedIn. And I pretended we had a point 5% click through rate bidding CPC as well as a 1% and then a 1.5% percent. And as I calculated this, I saw that as my click through rate went up, my estimated cost per impression also went up. When I had a point 5% click through rate it was 6 cents per impression, at 1% it was 12 cents, at 1.5% percent it was 18 cents. Now this isn’t 100% how this works on LinkedIn because we know that the higher the click through rate we have the less of a bid it’s going to take to actually win that auction. But the mechanics here are the same. It’s as your click through rate goes up. Your estimated cost per impression also goes up because you’re paying pretty close to what it is you’re going to bid but as your click through rate goes up, you don’t actually pay too much less than that amount. You start getting some discount, but it’s not huge like if you are bidding by cost per impression. All right, so let’s turn the tables here, let’s run the same scenario for cost per impression bidders. And I was basing this off of $130 cost per impression bid, which is like not super low and not super high, kind of somewhere in between. And I ran the same scenario, whether you have a .5%, click through rate, a 1%, or a 1.5 percent, your estimated cost per impression is 13 cents in this case. Regardless of your click through rate, you’re paying the same cost per impression, because that’s your bid. So just comparing our cost per click versus our cost per impression bids. If we look at just our E CPI here, reminding you the CPC bidder was around 12 cents ECPI when their click through rate was at 1%. And the CPM bidder was really close at 13 cents, just one cent higher. And that was across the whole range, no matter what our click through rate was. What that tells us is the break even point between bidding cost per click and cost per impression happens when you’re around the 1% click through rate mark. We’ve talked about this before, especially on the bidding episode, where you want to only bid cost per impression, when your click through rate is over about 1%. That’s when it actually becomes cheaper per click or per whatever action you’re talking about. It’s cheaper on a per impression basis and a per lead basis to be bidding cost per impression when you have a really high click through rate. And remembering that LinkedIn is benchmark click through rate what most people are getting is around .44%. So we see here very cleanly in the auction, this mechanic happening where it doesn’t make sense to bid by the cost per impression until your click through rate is over two times the benchmark. When I get the same thing on a cost per view basis, it was about 25 cents per view, which is a higher ECPI than any of the other bid models, which we’ve long said bidding for video on LinkedIn is just a little bit more expensive than other kinds of bids that you can do. And that’s okay. The value of video advertising is usually not in the clique. As Eric Jones talked about in our video ads episode here not too long ago, the real value in bidding for video is a cost per 50% video view. Okay, that’s as technical as we’re gonna get.
Relevancy
So we move on to the step three of the auction, which is relevancy. So in step two in the bidding, this is where all the advertisers are submitting their bids trying to get stacked up. And in the third step, here, relevancy. This is where LinkedIn actually orders them. LinkedIn calculates this score, we call it a combined score, from their predictive models. And the way that it does this, and if you remember our relevancy score episode, which was episode 72, you’ll know that LinkedIn is trying to order people by how likely a member is to take action on their ads. Now, if you’re bidding by cost per click, or by cost per impression, LinkedIn is using this predicted click through rate. But if you’re running a lead generation campaign, then the metric that LinkedIn looks at is a predicted lead through rate, which if I understand properly, that leads over impressions. And it might even be leads over clicks, just determining how likely someone is to submit the lead form. Video campaigns use something similar, they use a predicted view through rate. That’s the number of people who stick around for at least two seconds of video viewed on their screen, at least when it’s 50% on their screen, divided by the number of impressions. So LinkedIn is taking all of this into account, trying to understand how good of creative how how engaging they are, and how probable they are to actually get a visitor to take action.
Ranking
Now the fourth and final step is actually ranking, deciding which advertiser is above another. And if you listen to the relevancy score episode, you’ll know that this is done by multiplying your bid times your relevancy score. Your relevancy score is a number from zero to 10, whichever one has the highest number, with their bid times their relevancy score gets the top of the auction. They’re number one. So the question I have is, now that everyone is bidding exactly the same way, they’re all being compared properly, apples to apples. They’re all being compared in this ranking algorithm based off of their relevancy score. But in the past, we’ve been told by LinkedIn reps that CPM advertisers aren’t calculated the same way in for relevancy score, as CPC advertisers are. So how does this work? I actually don’t know. So if you’re listening in, you know, please do reach out to us and provide some clarity here. But I would imagine what’s happening here is, even CPM bids are still using a relevancy score. But most CPM bidders are using maximum delivery, which is dynamic, LinkedIn can change your bid however much it wants. So if you’re bidding by maximum delivery, you’re a lot more likely to win more auctions because LinkedIn can bid you up. But you don’t necessarily want that if your click through rate is lower than about 1%. Because then LinkedIn can really bid high and then you’re spending a lot more per click or per action than you should be.
15:02
Alright, so now we need to talk about campaign quality score, because we’ve been talking about relevancy score, but you can’t actually see what your relevancy score is on LinkedIn. But they do give us a metric. And that is if you go and run a campaign level report, within campaign manager, you’ll notice one of your columns says quality score. And this is LinkedIn trying to make this metric available to us without actually making it available to us. So your relevancy score isn’t something you can see. But your campaign quality score is kind of a proxy for it. It represents a lot of the same thing, and it should be able to be used directionally. And the way your campaign quality score is calculated, it takes your predicted click through rate for when you’re in the first slot and divides it by your campaigns predicted click through rate somewhere else. And what this is doing is it’s normalizing whether you’re in the first slot or somewhere else, it allows your quality score to be normalized, so that they can still compare apples to apples. And I love this because we get this question a lot. When we launch a new campaign, we tend to bid lower than what LinkedIn recommends. And a lot of people will ask us, why wouldn’t you want to bid higher, my LinkedIn rep is suggesting bid really high. And then we’re gonna get a better click through rate because we’re bidding high. What this campaign quality score calculation tells me is it actually doesn’t matter where you’re showing up, LinkedIn is still going to normalize you for your click through rate as if you were in the first slot, kind of cool. In the past, LinkedIn told me that your relevancy score is reset every night, probably at midnight, UTC time. And so if you want to see changes to your campaign quality score, we would recommend experimenting with different bid types, different bid values, try new creative, and maybe even adjust your targeting. But once LinkedIn has been through this process of ranking, now they’re ready to choose the actual campaign winner. Now they’re ready to choose the actual winner for this auction, who’s going to show up in the first ad slot, which is usually the second newsfeed item. So there you have it. Now we have actual proof through formulas on exactly why we recommend starting by bidding cost per click and bidding low. And then once you see your click through rate over about 1%, that’s when it makes sense to switch to a CPM bid, usually maximum delivery. That’s also why we recommend launching ads that get high click through rates. Because if you can get a click through rate over 1%, you’re going to start saving money by bidding CPM. And you probably know that when you’re bidding max delivery, it’s a lot easier to budget. So this is why we preach what we do. It’s why we push for the strategies that we do. And when you learn to work with the system, you can get better performance than when you’re working against it. Alright, I’ve got the episode resources for you coming right up, so stick around
17:54
Thank you for listening to the LinkedIn Ads Show. Hungry for more? AJ Wilcox, take it away.
The two episodes that I would highly recommend going to catch up on if you haven’t listened to them recently, the bidding and budgeting Episode Episode 89. And then the relevancy score Episode 72. If you’re looking to become an absolute LinkedIn Ads expert, go check out the LinkedIn Ads fanatics community that’s fanatics.B2linked.com. Here you’ll find all four of our courses taking you from absolute beginner to absolute expert. Plus, you’ll be a member of the community being able to ask questions get answers from other LinkedIn ads experts. It’s a low $79 a month to be a member of the community. But for an extra $200 a month, you get access to weekly group coaching sessions with me leading the charge. So if you want to ask specific questions about your campaigns, on screen share, and get my direct attention and feedback, come and join as a super fanatic. Now if this is your first time listening, we’re excited to have you here. Welcome! Hit that subscribe button on your favorite podcast player. But if this is not your first episode, please go do me the grand favor of rating and reviewing this podcast especially on Apple podcasts. That is by far the best way you can help us out with any questions, suggestions or corrections, reach out to us at the email Podcast@B2linked.com. And with that being said, we’ll see you back here next week. I’m cheering you on in your LinkedIn Ads initiatives.
Copyright © 2014 - 2023 B2Linked.com. All Rights Reserved.