Sep 22, 2022
Show Resources
Here were the resources we covered in the episode:
Campaign Quality Scores for Sponsored Content Help Article
LI's Privacy Centric Explanations - really interesting to understand LI's direction towards privacy
NEW LinkedIn Learning course about LinkedIn Ads by AJ Wilcox
Contact us at Podcast@B2Linked.com with ideas for what you'd like AJ to cover.
Show Transcript
I win auctions every day. No, I'm not an eBay addict. I'm a LinkedIn advertiser. We're talking about how to win the ad 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! LinkedIn's ad auction is confusing to advertisers because it's so opaque. And you don't quite know what's going on in the background. If you're looking to improve the performance of your ads, though, you'll definitely want to understand what's going on behind that curtain. So on today's episode, we're going to do a deep dive into the ad auction. First of all, some great stuff in the news. I found a resource that LinkedIn released called the professional identity resources for LinkedIn advertisers, I've gone ahead and link that down below in the show notes so you can check it out. But basically, it's a download of how LinkedIn is looking at, and thinking about privacy concerns. And it's really helpful to understand the direction that LinkedIn is moving with things like conversion tracking, and some of the things we've seen recently like with reach being taken away. So if that's interesting to you definitely go check that article out, there's quite a few things there. I was also able to attend the meeting for LinkedIn API partners and we got a bit of a roadmap update, which was really exciting. Usually, LinkedIn likes to keep their advances that are on the platform up to date with what's happening on the API as well. So if you happen to use LinkedIn API in any way, like if you're using a partner of LinkedIn, that helps you at all with what you're doing inside of campaign manager, then you might notice this kind of functionality coming out. What they talked about, that I got especially excited about were dynamic UTM parameters. And it looks like these are going to be available in LinkedIn API first, and then we'll eventually get them inside of the dashboard. So for those of you who are already advertising on Facebook, and you love how you can create the same ad once, but in every campaign, or every ad set that you run it in, you can have different UTM parameters running to it, we're eventually going to have that on LinkedIn as well. I'm a huge fan of this. I've been asking for it for years, they also talked about how offline conversions are coming. And I've got this inside of my campaign manager dashboard so it's available on the UI to me now, I don't know if it's available to everyone yet. But we know it's probably coming to the API soon. So you might be able to use a different partner program to help you with something like these tying offline conversions back to your adspend.
2:39
So let's talk about offline conversions. It's a little bit limited
in the way that LinkedIn is doing it. But you'll understand why.
Imagine that you're advertising and then one of your ads turns out
later to generate a sale. What you can then do is send that list of
email addresses of those people who've become your customer back to
LinkedIn. And then LinkedIn will take that and match it up to that
same user if they're able to find that same email address attached
to a user and figure out which campaign which ad they initially
clicked on. And then they'll be able to do that attribution. The
obvious weakness here, though, is that LinkedIn knows personal
email addresses, because that's how we log in and sign up for
LinkedIn. And if you close a customer off of LinkedIn Ads, they're
probably going to end up giving you their professional email. So
let's say if LinkedIn only understands 50% of email addresses,
that's still cool to have these offline conversions being tied up
in the platform. So I'm excited about the functionality, but just
realizing it's not going to be fully complete. There was also some
new development around DMP segments coming out with the API. So
that could end up being interesting. If there are those of you who
are using DMPs.
Going a different direction here. One of LinkedIn engineers reached
out to me with a really interesting ask. She said, she's working on
the accessibility features for campaign manager, and that she's
searching for some LinkedIn users who might actually utilize those
accessibility features. So maybe it's someone who is vision
impaired. And they're using things like a screen reader, or
utilizing the alt text of images, when either creating or consuming
different ads. She would love to have a brief chat with them over
zoom or over the phone and share some takeaways with their
engineers as they're trying to make their accessibility features
better. So I'm calling on all of you. Do you know anyone who builds
ads who uses any of those accessibility features? Maybe because
their vision impaired or for some other reason? If so, please reach
out to us at Podcast@B2Linked.com and I'd love to introduce you to
Carol.
We reported a few weeks ago that LinkedIn is now only reporting on
some averages for reach and frequency. And we're no longer getting
those accurate counts. It was done so suddenly that I think there
was a lot of advertiser blowback. And so LinkedIn appears to have
reverted this change, but the bad news is, it's likely still going
to be averaged again in the future. But at least for now we have
the raw numbers. So go ahead and use them for whatever they're
worth. But know that we're probably going to lose them again soon
here in the future.
5:13
I have a really cool announcement. I'm actually getting married
this week. And so I'm very excited. It's gonna be fun. We're gonna
go on a cruise for honeymoon. The bad news to you is that I may end
up skipping a couple of weeks of episodes, but I promise I'll be
back as soon as I can. I wanted to highlight a review that came in
from Maggie Mulholland one of our friends, actually, who works at
LinkedIn. She said, "Great resource! Such a time worthy listen for
anyone in the industry. AJ brings an honest and well rounded take
on all things LinkedIn." Maggie, thanks so much for sharing that. I
do try really hard to have it be an honest and well rounded take. I
do hope we're not ruffling any feathers at LinkedIn when we talk
about some of the products the way that we do. But I also hope that
the praise that we keep on the platform also comes across as
intended as well. So thanks, Maggie, great to have you as a
listener.
6:01
Alright, so now to the topic at hand about the LinkedIn ads
auction, let's hit it. First, we have to ask ourselves, what is an
auction. Put simply, LinkedIn has a limited number of ad
impressions that it can show to any given user. And there are
obviously quite a few companies who would love to show an ad to any
of these users. So LinkedIn, just like all the other major ad
platforms, especially like Google and Facebook, it holds an
internal auction to decide which advertisers ad to show to any
individual during the day. This was a concept that I believe was
pioneered by Google early on with Google AdWords, that's now Google
Ads. And it's a really genius way of maximizing the profit of any
individual user on the platform. So let's talk about how it works.
Let's say you and I both want to reach the same audience member on
LinkedIn. And let's say I'm willing to pay $8 for a click, but
you're willing to pay $10 per click. We would naturally assume that
LinkedIn would look at it and say, Oh, that person is willing to
pay $10 A click that's $2. More for a click than Aj is. Let's show
their ad. So that situation seems pretty easy at first blush. Well,
what about if you and I are both willing to pay $10 for a click?
How does LinkedIn then decide which of our ads they're going to
show? Google's answer to this was called the quality score. And the
essence of it was, if we're only paying for when someone clicks,
then LinkedIn can figure out which one of us is more likely to make
the network money when they show our ads. So for example, let's say
that my ads have a .5% click through rate, meaning that every 200
times they show my sponsored content ad, I'm going to get a click,
and LinkedIn is going to make $10. But in this example, you're also
bidding $10. But historically, your ads get a 1% click through
rate. So your ads get clicked on twice as often as mine. LinkedIn
looks at that and says, whoa, both of these advertisers are willing
to pay $10. But they make $10 for every 100 people they show your
ad to, but they only make $10 for every 200 people they show my ad
to. So now you can see how it's in Lincoln's best interest to show
the ads of the advertiser who tends to get the best engagement. So
the metric that judges how effective you are at getting people to
click on your ads, and how effective I am at getting people to
click on mine is called our relevancy score. Really similar concept
on Google, it's called quality score, really similar concept on
Facebook ads, it's called relevance score. And it's really cool.
But the challenge to it is we as advertisers, we don't know what
our relevancy score is. Google used to show us quality scores all
the way down to the ad level. But over time, they took that
visibility away, which makes a lot of sense, because it's in the
platform's best interest to hide your relevancy score or your
quality score from you. The reason why is just so someone doesn't
game it. So let's say your relevancy score is updated every single
day, and LinkedIn shows it to you. That means you can effectively
do one test per day on changing an ad or launching a new campaign.
And maybe over time, you can start to understand how your relevancy
score is affected by the changes that you make, effectively gaming
the system. Many of you know my background, I started out in Google
ads. And I know that the quality score algorithm was something that
was heavily debated by advertisers. We always wanted to try to
figure out more about what goes into it and how it's taken into
account. This dates me a little bit, but I remember back when
Google announced that they had 21 different factors that affected
their quality score. So when I started getting really heavy into
LinkedIn Ads, I was pleasantly surprised how simple the relevancy
is. Your calculation really was. At the time your relevancy score
was really just a combination of your historical click through
rates and your current click through rates. So if your campaign has
had really good click through rates over a long period of time, and
you launch a new ad into that campaign, you may start out with a
really good assumed relevancy score, which is so helpful.
LinkedIn's product team hasn't given me any sort of insight into
LinkedIn relevancy score, currently. But my guess is, it's now
gotten a lot more complex. But we'll of course talk about that a
little bit later. So your relevancy score is effectively a
normalized range from zero to 10. Zero meaning that your ads are
providing no value, and a 10 mins that people are clicking on it
like crazy, and really loving what you're putting out as an
advertiser. Now I say a normalized range, because if you are a
seven today, but all of a sudden, a competitor enters the auction
against you, and they have a much higher relevancy score, let's say
they have a nine, your relevancy score is a calculation of your
performance compared to those who are in the auction with you. And
so yours might sway. Yours might bumped down to a six, because
there's just so good, and it's all averaged. Okay, so you have this
relevancy score that somewhere between zero and 10, and you start
bidding in the auction, you don't know what those who are bidding
against you what they're bidding. And so it really is a blind
auction that way. So that example that I used before, where you're
willing to pay $10 for a click, and I'm only willing to pay eight.
And of course, the auction is going to give it to you. Well, that's
not the case. Thank goodness, it's not so simple. So the way it
works is that two parties enter the auction. And in reality,
there's a lot more than just two parties. But for simplicity's
sake, let's say it's just you and I who are bidding for an audience
member. Let's say I'm willing to pay $12 for a click, but you're
only willing to pay $6.50 per click. So you're bidding basically
half of what I'm bidding. So at first blush, you're now thinking,
ooh, it sure seems like LinkedIn is going to want to show AJs ad
over mine. But now when I tell you that behind the scenes, this
campaign only has a relevancy score of four, but yours has a
relevancy score of eight, what LinkedIn is doing behind the scenes,
they are multiplying my bid, times my relevancy score, and your bid
times your relevancy score, to get this combined score. So in this
case, my combined score would be a 48. It's my bid of $12 times my
relevancy score of four, and you're bidding $6.50. But you already
have a relevancy score of eight, which if you multiply that
together, you get a 52. So now what LinkedIn is doing is saying,
Ooh, whoever has the highest combined score is who actually wins
the auction for their ad to show in this exact impression that just
arose. Okay, so your combined score is higher than mine, which
means you're going to win the impression. But now LinkedIn has to
decide how much are you going to pay for that click, because we
were obviously bidding very different amounts. I was willing to pay
$12 for a click, and you were only willing to pay $6.50. So the way
that it does this is it takes the second place bidders That's mine,
my combined score, which in this case is a 48. And then they divide
that by your relevancy score, which is an eight. This simple
mathematical operation tells the system what the second place
bidder would have had to bid in order to become the first place.
And with this simple mathematical operation, it lets us see exactly
what you would have had to bid in order to beat me in the auction,
which would have been exactly $6. And because LinkedIn is a second
price auction, the same auction model that Google pioneered, we add
one cent to it. So basically, even though you're bidding $6.50, you
only have to pay $6.01 for that click, because that's all it took
for you to outbid me. If this is a little bit complex to hear about
math over a podcast, I totally get it. Down in the show notes, you
can click on a video that LinkedIn created that actually explains
their auction system, and they show it with a really cool
animation, I think you'll like it. So now you understand how the
auction system works behind the scenes. But that's not especially
helpful because all of this is visible to LinkedIn, those who are
hosting the auction, but it's totally invisible to you. All you see
is you bid a certain amount, and then get a certain amount of
impressions that turned into a certain number of clicks. So you can
change your bids frequently and try to understand how close am I to
be getting more advertisers in the auction and being able to win a
lot more impressions, or how low can I bid without losing the vast
majority of my impressions that I get. But LinkedIn is help article
will tell us exactly how they explain relevancy score. It says, ads
are assigned a relevancy score that measures how likely a member is
to take an action on the app. Relevancy scores include factors like
expected click through rate, comments, likes, and shares, your
relevancy score can change over time, as members interact with your
content. While you can't see your ads relevancy score, you can use
the campaign quality score as a proxy. So this gives us a really
important clue. When I go and run a campaign export from within
campaign manager, I noticed one of the columns says campaign
quality score. And that's a number from zero to 10, which looks a
lot like relevancy score. So if we read into what LinkedIn is
saying, we can look at campaign quality score, and it acts as a
proxy to what our relevancy score is. But we won't actually know
what our relevancy score is. So relevancy score is something that
applies to ads and to campaigns, but your campaign quality score is
just that same normalized range from zero to 10 that describes the
campaign. So they're essentially not giving you any information
about individual ads, but you do get a little bit about the
campaign itself. Alright, here's a quick sponsor break, and then
we'll dive back into the meat of it.
16:19
The LinkedIn Ads Show is proudly brought to you by B2Linked.com,
the LinkedIn Ads experts.
If you're a B2B company and care about getting more sales
opportunities with your ideal prospects, then chances are LinkedIn
ads are for you. But the platform isn't easy to use, and can be
painfully expensive on the front end, at B2Linked, we've cracked
the code to maximizing ROI while minimizing costs. Our methodology
includes building and executing LinkedIn Ads strategies that are
customized to your unique needs, and tailored to the way B2B
consumers buy today. Over the last 11 years, we've worked with some
of the largest LinkedIn advertisers in the world, we've spent over
$150 million dollars on the platform, and we're official LinkedIn
partners. If you want to generate more sales opportunities with
your ideal prospects, book a discovery call at B2linked.com/apply,
today. We'd absolutely love to get to work with you.
Alright, let's jump back into relevancy scores here. So when I used
to look at the campaign quality score column, in my campaigns
report, I honestly thought that oh, some engineer at LinkedIn who
used to work at Google, accidentally mislabeled it and wrote
quality score when they should have written relevancy score. But I
put a post out in June, asking for help and thoughts from the
LinkedIn masters out there. And one of my connections, Decker
Frasier., he turned me on to this. He pointed me towards that help
article, where it talks about how your quality score is a proxy for
your relevancy score, but they're not going to show you what the
actual relevancy score is. So Decker, thanks so much for turning me
on to that. That was new info for me. So then I got to dive into
the help article all about campaign quality scores. And I've linked
to that article below if you wanted to go and do your own research.
The article starts off "A campaign quality score is an estimate of
how likely a member is to act on a sponsored content ad in your
campaign. scores help indicate how relevant your campaigns are
compared to your peers campaigns targeting that same audience." So
immediately, we're understanding that this is totally based on the
competition around you. So you could have a terrible click through
rate and still have a great quality score if all of your
competitors are also getting terrible click through rates. Or
conversely, you can have an amazing click through rate, but if so
many of your competitors have even better than you're just stuck
with a normal or an average quality score. The article then goes on
to explain how campaign quality scores are based on the predicted
click through rate of the ads in your campaign, as well as the
predicted click through rate of your peers ads targeting the same
audience. So that raises the question, how does LinkedIn predict
what your click through rate is going to be? Here's another little
nugget to point out in this article, campaign quality scores and
predicted click through rate are only helpful for evaluating
campaigns using CPC bidding. So what that means is, the only ones
of us who are actually part of the auction are those who are
bidding by the click. So if you are bidding by the impression, or
if you're using LinkedIn's maximum delivery or auto bidding
options, you're effectively bypassing the auction entirely. You
don't have to worry about your campaign quality scores or your
relevancy scores. And that's because if you tell LinkedIn that
you're willing to pay a certain amount, regardless of who clicks
for every 1000 times they show your ad, LinkedIn can very easily
compare you to another advertiser who's saying the same thing. So
if you're willing to pay $120 for them to show your ad to 1000
users, and your competitor is only willing to pay $100 to reach
that same 1000 users, LinkedIn doesn't have to do much calculating
at all. It just says, oh, that advertiser is willing to pay me $20
more for the same traffic, I'm going to give them more impressions.
I have had several members of my team come to me and say, "Hey, has
campaign quality score gone away? Because when I run a report, I
don't see it in there." Well, LinkedIn is help article here says,
"If a campaign quality score isn't available, it's probably because
your campaign isn't active, or it's not using the sponsored content
ad format. Or maybe it's too early. And your campaigns ads haven't
competed in the minimum number of options for that score to be
calculated." So if that column is blank for your campaign, one of
those four reasons is going to be why.
20:47
So that then begs the question, how do we improve our quality
scores or our relevancy scores? Well, that's pretty simple. It's
improving our click through rate. But of course, that's much easier
said than done, check out Episode 59, where we talk all about how
to increase click through rates, and all the different controls we
have on them. And although our bid doesn't directly contribute to
your relevancy score, your bid is used in the calculation of your
combined score, which then decides if you win auctions or not.
Let's say you're bidding pretty low at $7 per click, you may see
that you get, let's say, 200 impressions per day. If after
increasing your bid to $12 per click, you might see your
impressions jump up to 1000 per day. And what that means is you
were only winning like 200 impressions per day with your lower bid,
but now that you're bidding higher, you're qualifying and winning a
lot more of these auctions. But of course, as you're bidding
higher, it means those auctions that you do win and when a user
actually clicks, you will pay quite a bit more for that, click, go
back and check out Episode Six, that was all about bidding, if you
want to become an absolute ninja Jedi Master or whatever, on the
whole topic of bidding. If we look at the other platforms,
especially like Google and Facebook, who are much much more
advanced in tech, it might afford us a bit of a glimpse into what
relevancy score will be in the future, or maybe what it's already
developing to be. Google, for instance, has so many advertisers and
so much competition, that just deciding someone's quality score
based off of the click through rate of their ads, it doesn't tell
the whole picture. So think about the other kinds of factors, which
might tell the platform, how successful you are as an advertiser.
Some of those things might be when you are sending traffic to a
landing page. If that landing page loads really slowly, you could
tell that people who are clicking on those ads probably are not
going to have the best experience and they may end up bouncing
before the page even loads. So it's much better for them to reward
the advertiser with a higher quality score, whose pages load near
immediately. With Google, it's really simple because they're
searching by keywords and so Google can take into account how
relevant the keyword is that someone clicked from, to the keywords
that are actually found on the page. Who knows if LinkedIn is using
something like this, but they certainly could. So my recommendation
to you is whether LinkedIn or using any other factors other than
your historical and current click through rates or not, I would go
and look at things like content of my landing page and how fast it
loads. Because you can guess that if Google and Facebook had been
doing something for lots of years, LinkedIn is sure to follow.
Something else we have to talk about is how LinkedIn reps talk
about pausing your campaigns or pausing your ads. There can be lots
of great reasons to pause a campaign or pause ads, but if you're
working tightly with a LinkedIn rep, you've probably heard them
say, "Don't pause your ads or your campaigns, because it will
affect your relevancy score, or it will reset your relevancy
score." And we've done a lot of testing because this sort of
sounded like an empty threat to us. And largely, I think it is,
we've had several reps tell us that if you pause your campaigns for
more than two weeks, then when you go to turn them back on, your
relevancy score is totally reset. And I definitely don't think that
his reps are lying. But I understand if they are bonused, based off
of how much their advertisers are spending, and we pause things
like over the weekend or at nights, then the campaign's will likely
not end up spending as much as they're budgeted for. And so of
course, the reps would want to caution us away from that. So in all
of our testing, we found that pausing campaigns does not seem to
affect us in the auction at all. That means if we pause a campaign
with its ads for a full two weeks, and then turn it back on, if the
relevancy score were reset, we should see action that was very much
like the campaign was newly launched, which means it would probably
enter a learning phase for the first one to one and a half days,
where we either got fewer impressions or more impressions, as the
system is just testing to see what relevancy score we actually
deserve. So we haven't seen this action take place. But I'd love it
if any of you have done this same test, if you're seeing anything
like this, that would show to you that your relevancy score is
being reset, then please do reach out to me and let me know. I'd be
really curious to hear about that test. So let's say that LinkedIn
is actually resetting our relevancy scores after pausing. If we're
not seeing a big effect come from that. One reason could be that if
LinkedIn shows your ads, and it's earned a certain relevancy score
and place in the auction, and then you pause, and then start again,
that audience is just as likely to interact with your ads the same
way that they did before. And in fact, maybe even slightly higher,
because they haven't seen the ads in two weeks. So those who have
already interacted and seeing them, they have probably forgotten
about seeing them, and it looks new to them. So that's a possible
reason why even if your relevancy score does get scrubbed, after
two weeks of pausing, why performance can still look good. But
again, that's just a hypothesis. I'd love to hear from you guys, if
you've seen any sort of effect like this. All right, I've got the
episode resources for you coming right up. So stick around.
26:28
Thank you for listening to the LinkedIn Ads Show. Hungry for more?
AJ Wilcox, take it away.
All right, I mentioned Episode 59 of the show. For those of you who are curious about how to improve click through rates, definitely check that one out. It is the key to getting lower costs on LinkedIn. There's also the LinkedIn Help Center article about LinkedIn's Ad auction, and how it's calculated. That is sincerely interesting and I would recommend it as a read. There's also the article that we sampled out of here, the campaign quality scores for sponsored content help article. There are some great insights there. It's really awesome when LinkedIn is so transparent about how their system works. It sure helps us advance advertisers not put on our tinfoil hats and assume the worst. There's also LinkedIn's privacy centric page that talks all about their new initiatives around privacy. Definitely worth checking out, especially after Episode 70 all about the cookie pocalypse, you'll probably understand that one quite a bit better. I mentioned Episode Six all about bidding. That that one's definitely worth going back to have a listen, if you've missed that one. Or it's always worth a read, listen, because it's honestly one of the most important things from your whole. So go back and have a listen of that. In case you missed it. Or even if you've already heard it, it's super important. It is the basics of advanced LinkedIn Ad strategy. So make sure you know it like the back of your hand. If you or anyone else, you know, is looking to learn more about LinkedIn Ads, check out the link in the show notes for the course that I did on LinkedIn Learning all about the basics of LinkedIn Ads. It's incredibly inexpensive, and really high quality. The LinkedIn Learning folks really know what they're doing. If this is your first time listening, welcome, we're excited to have you here. Make sure you hit that subscribe button. Of course, that's only if you liked what you heard. If this is not your first time listening, please pay us the fee of leaving a review on the podcast. It really, really helps. I'm not just saying that. And of course I'd love to shout you out for leaving a review. For any questions, feedback or suggestions on the show, reach out to us at Podcast@B2Linked.com. And with that being said, we'll see you back here next week. Cheering you on in your LinkedIn Ads initiatives!