One of my favorite pandemic obsessions has been watching Bon Appetit videos. Specifically, Claire Saffitz’s many attempts at gourmet recreations of classic junk food. This segment that I just made up is going to be my attempt to recreate (or pull apart) stuff I see marketed in adtech to see if it’s a real thing or just well-polished bullshit.
Speaking of, the world of audience targeting is up to its armpits with bullshit. Call me jaded but as far as I can tell there’s nothing new under the sun and hasn’t been for quite some time. To their credit, NYTimes has been very good at marketing their various audience targeting products. Note, I said marketing, not creating.
Ever since the NYT got credit for it’s “motivation” targeting, despite the fact that ESPN and USA Today announced the same capability on the same day, I’ve been very skeptical. I assume all three just white-labeled the same/similar product.
So when Sara Fischer posted her newest exclusive that NYT was phasing out all 3rd-party ad data by 2021, my ears (eyes?) perked up. Then I read the list of first-party audience segments that will be available and, honestly, had a little panic attack.
Beginning in July, The Times will begin to offer clients 45 new proprietary first-party audience segments to target ads.
Those segments are broken up into 6 categories: age (age ranges, generation), income (HHI, investable assets, etc.), business (level, industry, retirement, etc.), demo (gender, education, marital status, etc.) and interest (fashion, etc.)
How TF did they do it?
How would the NYT know, using only first-party data, my household income? My entire NYT profile is my email and billing address. And even if they did ask me my HHI or investable assets there’s absolutely no way that data would stay accurate beyond 90 days of entry… assuming it was accurate at all.
Adding to my panic, Ari Paparo doesn’t seem to think this announcement means much. Now it’s starting to feel like everyone is in on this joke, except me. WTF?
My imposter syndrome is now firing on all cylinders. So in an attempt to calm myself and also write a thing today, I’m going to brain blurt all the ways I think NYT is performing this black magic. I would also like to implore (beg?) of you der reader, if you know how this works or have a better guess than me… email me at firstname.lastname@example.org.
How To Get Third-Party Data Segments In A First Party World?
Ok so let’s think this through. For the sake of time and sanity, I’m going to focus on the audience segments that most closely replace third-party segments (income, business, and demo).
First, to get it out of the way, do we think NYT has some sort of data science magic that no other company in this space has? Do they have the silver bullet? Did they invent the magic revenue growing beans? No. No there are no proprietary miracles to be found here. To be clear, there are some VERY smart people working at the NYT likely doing some very creative bleeding-edge shit. I just don’t believe they’ve turned water into wine.
So if I had to recreate this where would I start?
1) Geographic data
The simplest answer is often the most likely. None are simpler than taking some geographic data (HHI by zip let’s say) and by taking a user’s IP and zipcode, applying some guesswork and presto. You’ve got third party segments using just first-party data. There are some very obvious issues with this methodology (especially with demographic data) but for HHI and investable assets, this does feel like the easiest way to get there.
Ok, boom. One guess and one of the three groups down.
2) Panel / Survey Data
Ok, so this one is tougher. It requires a lot more investment but as far as I can tell the only way you get a good understanding of demographic data is by surveying a decent-sized portion of your audience and then using that data to build LAL segments.
The first problem is enticing a reader to complete the survey. Next is building the methodology to cover the majority of your readers, both subscriber, and drive-by variety. Luckily NYT is a large org with very smart data people so I think it’s safe to assume they can handle this.
Two methods down and I’ve been able to recreate income levels and age/demo targeting. But what about c-suite?
3) Email Lists?
So this is the one where I’m most stuck. Could you figure out business level (c-suite and business decision-makers) from a survey? Sure. But something tells me that’s not robust enough for the NYT’s clients.
My gut tells me this is where they use their HUGE amount of user email data to build out the business level lists. How? I HAVE NO IDEA. I’m already so far over my skis here that I can only say catch-phrases and keywords with hope that they make sense.
If this email does anything I hope it exposes how little I understand about the first-party data enrichment world.
I know for a fact some of you readers know this world way better than I do and are probably chuckling to yourself at my expense. That’s fine! I don’t mind being the butt of a joke. Please reach out to me and explain how wrong I am or where I’ve gone off track. That way I can report back to everyone else.
Why should NYT be the only one able to build out a product like this? Shouldn’t this be something we want every quality publisher to be able to do?
Shameless Self Promotion
Check out Erik and I’s newest episode of What Happens in Adtech. This week Erik and I talk to Will Doherty. Will talks through the process he and his team use to build IX’s Marketplace pulse email.
We also offer an audio-only version of WHIAT on Spotify and iTunes. Subscribe!
If you like the words I wrote above please share it with other smart people:
Ok, that’s all for me. Have a great day, or don’t. I’m not the boss of you!