Perfect Audience (YC S11) outperforms Adroll in Retargeting Test | Hacker News Reader

Perfect Audience (YC S11) outperforms Adroll in Retargeting Test

pearanalytics.com

50 pointsbrandnewlow8 comments

I work for AdRoll. I left a comment on this blog post but it hasn't been approved yet.

We’d really appreciate the opportunity to do a real comparison. In this post, they are comparing a campaign without Facebook Exchange (AdRoll) to a campaign with Facebook Exchange access and using two different definitions for conversions.

It’s a little odd that FBX wasn't included with the AdRoll campaign since unlike PerfectAudience, AdRoll is actually is one of few PMDs that have a seat on FBX ( http://www.facebook-pmdcenter.com/fbx ) and has the most clients running FBX campaigns of any FBX PMD.

PerfectAudience’s CPCs and CPMs are likely lower because Facebook retargeting is cheaper in this regard, and if they ran an AdRoll FBX would likely be comparable.

Also looking at your charts, they never setup conversion tracking in AdRoll. Without this they are comparing (PerfectAudience) view through conversion CPA to last touch Google analytics click through conversion CPA. To analytics and online marketing expert, it should be obvious this is not a fair comparison.

A lot of the UI callouts are valid and AdRoll is working on new features and launching a new dashboard very shortly. If anyone wants to setup a real test and do real analysis of performance, we’d be happy to help.

-Tom

Can you go into a bit more detail on the difference between these two types of conversion?

I suppose view-through conversion might be clicking on ad, user goes to site, completes purchase and then some kind of callback gets fired. On the other hand, Google analytics one would be user completes purchase, GA looks at which referrer came last and registers that as the source of conversion?

View through conversions are conversions that take place after a visitor has just viewed an ad and not necessarily clicked on the ad. In the case of traditional display, the ad may not have even been visible on the screen it was just served somewhere on the page.

Google Analytics conversions are last touch click through conversions, where all marketing touches compete and the last that occured gets 100% of the credit.

Yep. And just going by Google analytics tracking is a rough way to assess marketing channels, agreed.

The ubiquity of Google Analytics and thus, ubiquity of "last touch" conversion tracking, has really helped Google make the case for Adwords. If someone sees an ad from a display advertising company, retarteted or otherwise, it's not uncommon for them to head to Google to look up more information about that product or service. If they then sign up or convert, the advertiser will see Google analytics listed as the sole source of the conversion. This is great for Google and less great for everyone else higher up in the funnel.

This article is so specious. I am actually shocked that something this bad would come out of a company that purports to understand internet marketing and analytics.

First, its comparing two completely different types of retargeting. It compares one campaign where Facebook retargeting makes up 3/4 of the impressions to one that is just general retargeting on the ad exchanges. These perform completely differently.

Second, the timeframes are different. This makes a big difference especially if they have been running the same ads over a long period of time as ads (especially retargeted ads) tend to get ignored after a while so the falloff in performance can be quite rapid.

Third, comparing post-click conversion across retargeting partners doesn't really make much sense. You're targeting the audience to begin with, you should probably care mostly what percentage of your audience they are reaching and what the CPM is more than the CPA. Unless these guys are doing fancy stuff (e.g. more aggressively retargeting people who have clicked but then did not convert) it actually does not make any sense that CPA would be any different at scale.

Fourth, if you're running two different retargeting platforms on the same audience its possible that you're bidding up your own retargeting cost as the two platforms are bidding against each other. e.g. this shows a CTR of around 0.13% which is actually pretty good for FB retargeting, but they're paying ~$1 CPC which is really bad for FB retargeting.

They got 18% conversion on their clicks from Perfect Audience? Seems dubious.

With such a small sample of data 109 sessions vs 68 sessions... I wouldn't call this a clear winner.

This could actually just be a case of misattribution of sales.

For example, if the customer clicks on a Perfect Audience ad, they could be tagging that customer as "theirs" for the next 30 days, or forever. So imagine someone clicks on this ad, browses the site again, and leaves. Then they come back 5 days later of their own accord and make a purchase. Perfect Audience may be (incorrectly) attributing that sale as their own.

I'd like to see the Ecommerce breakdown that GA is reporting. (this data: http://www.pearanalytics.com/wp-content/uploads/2012/12/2012...)

Hi Tom, Ryan from Pear here. AdRoll did not have the Facebook Exchange option when it was available, and we saw Perfect Audience had it first. Also, I do have the same conversion goals set up in AdRoll as I do in PA. (/checkout_thankyou.asp), and they are not firing properly, even though in Analytics I do see some conversions.

I'm happy to devise a better experiment with you and run our next customer on both platforms with the exact same criteria. Email me if you're interested ryan at pearanalytics dot com.

Hey Ryan. We're happy to help you get setup and use the AdRoll platform with all of its features. I'll email you and have someone on our account and support team follow up with you.

Reply on news.ycombinator.com