Hi Quo Vadis! I have a question about one of the Quo Vadis portfolio unicorn club: Criteo!
We are in the process of setting up a simple dynamic remarketing head2head. Our custom dynamic creative templates vs. Criteo.
So, initially, my hypothesis was:
If we have access to the same product feed, and the same inventory as Criteo (the open exchange) but we create custom assets that are aesthetically more engaging (sounds subjective, but the difference is so significant that I think we can assume this is objective) than generic Criteo templates, we will drive a stronger CPA than Criteo.
One obvious assumption being that we do, in fact, have access to the same inventory, and on an even playing field.
However, it seems that there is at least one technical disadvantage that means we won't be able to compete:
Upon diving in to the Criteo data, their CPMs clear at circa $1. Since I first visited the advertiser's site, Criteo ads have accounted for almost 100% of the ads I have seen. This seems to imply they are winning an extremely high percentage of auctions (perhaps a win rate +70% at least), whilst bidding at around the $1 mark.
- If I compare this to our remarketing activity that we have run this year, our average CPM is circa $8 and our win rate is just 18%. Even if Criteo's 70% win rate is not an exact science, there is undoubtedly a big difference here.
How is this possible?!
- Do Criteo have some kind of publisher/SSP partnerships, that enable them a first look at auctions across certain publishers?
- Or, what other variables can feed in to such a significant disparity between bids and win rates? And can the answer be generalised for other specialist programmatic partners/platforms?
Thanks for your time reading this. I hope that makes sense!
1. " Do Criteo have some kind of publisher/SSP partnerships, that enable them a first look at auctions across certain publishers?". Yes, these direct deals used to represent 50% of clicks/revenue a few years ago, I'm not sure how significant it is today, but I'd guess it's still a big factor.
2. They charge advertisers a CPC and buy inventory on a CPM basis. That means the eCPM they bid = CPC x predicted CTR of users x 1000.
If the prediction is accurate, they have a tremendous advantage. And Criteo's error rate is likely sub 1%, so it's very accurate.
But keep in mind it's not that hard to get high ROAS by serving ads to users that have a high chance of converting anyway. Kind of silly when you think about it, but that's advertising.
3. Having many direct pub 1st look deals with SMB pubs is important because with retargeting you need 2 ingredients:
1) client appetite for vanity ROAS
2) an ability to win impressions "first and fast" because click probabilities decay rapidly over the first 24 hours since the user shops on a site.
Hi Quo Vadis! I have a question about one of the Quo Vadis portfolio unicorn club: Criteo!
We are in the process of setting up a simple dynamic remarketing head2head. Our custom dynamic creative templates vs. Criteo.
So, initially, my hypothesis was:
If we have access to the same product feed, and the same inventory as Criteo (the open exchange) but we create custom assets that are aesthetically more engaging (sounds subjective, but the difference is so significant that I think we can assume this is objective) than generic Criteo templates, we will drive a stronger CPA than Criteo.
One obvious assumption being that we do, in fact, have access to the same inventory, and on an even playing field.
However, it seems that there is at least one technical disadvantage that means we won't be able to compete:
Upon diving in to the Criteo data, their CPMs clear at circa $1. Since I first visited the advertiser's site, Criteo ads have accounted for almost 100% of the ads I have seen. This seems to imply they are winning an extremely high percentage of auctions (perhaps a win rate +70% at least), whilst bidding at around the $1 mark.
- If I compare this to our remarketing activity that we have run this year, our average CPM is circa $8 and our win rate is just 18%. Even if Criteo's 70% win rate is not an exact science, there is undoubtedly a big difference here.
How is this possible?!
- Do Criteo have some kind of publisher/SSP partnerships, that enable them a first look at auctions across certain publishers?
- Or, what other variables can feed in to such a significant disparity between bids and win rates? And can the answer be generalised for other specialist programmatic partners/platforms?
Thanks for your time reading this. I hope that makes sense!
Hi Omri — it comes down to 3 advantages:
1. " Do Criteo have some kind of publisher/SSP partnerships, that enable them a first look at auctions across certain publishers?". Yes, these direct deals used to represent 50% of clicks/revenue a few years ago, I'm not sure how significant it is today, but I'd guess it's still a big factor.
2. They charge advertisers a CPC and buy inventory on a CPM basis. That means the eCPM they bid = CPC x predicted CTR of users x 1000.
If the prediction is accurate, they have a tremendous advantage. And Criteo's error rate is likely sub 1%, so it's very accurate.
But keep in mind it's not that hard to get high ROAS by serving ads to users that have a high chance of converting anyway. Kind of silly when you think about it, but that's advertising.
3. Having many direct pub 1st look deals with SMB pubs is important because with retargeting you need 2 ingredients:
1) client appetite for vanity ROAS
2) an ability to win impressions "first and fast" because click probabilities decay rapidly over the first 24 hours since the user shops on a site.
Does that help you out?
This was incredible! I'd love to build a partnership, please email me at chris.hartley@britepool.com