11 comments

[ 2.7 ms ] story [ 36.3 ms ] thread
I worked for this industry some time ago, and, to my knowledge, you would be right to assume that "95%+ of Adtech Companies Are Full Of Shit".

Even in big Corporations like Criteo, someone will occasionally omit a portion of samples during testing, tweak slightly output results, or flat out invent numbers from the bottom of their heart in order to please clients and to give them an illusion of something that works.

Criteo is very full of it. They mislead customers, pointing to last click attribution which makes them it seem that Criteo is doing WAAAY more than they are actually doing.

There a lot of other stuff they are guilty of too. And it's not just them. Every retargeting vendor is guilty of the same stuff. They are like the A/B testing vendors. Tell a beautiful story, hide the fine print and make it sound like you are the best thing since sliced bread.

> invent numbers from the bottom of their heart in order to please clients and to give them an illusion of something that works.

Think (some) customers are part of the problem here - that they are happy to accept this illusion because it makes their lives easier.

In your traditional bricks n' mortar company, it used to be that the marketing manager could just hand over big chunks of budget to an agency and watch them splash it on a few big campaigns (TV ads, print etc.) - easy work with the added bonus of free lunches and other perks from the agency to keep their client happy.

But online advertising is a different beast. The legitimate players in this space, like Google and Facebook, require you to actively manage your campaigns and provide all kinds of reporting to show exactly how you're performing. That means work for the poor marketing manager who's a bit too used to long lunches; there's effectively an audit trail for your CEO to see exactly what their marketing budget is buying them, rather than some vague "brand recognition" metrics which are easy to fuzz.

Spending the ad budget becomes like Brewster's Millions ( https://www.youtube.com/watch?v=-9kMHiqS6Kg ) so anyone who can step up with a way to blow large chunks of it quickly is going to do well with these guys.

>In your traditional bricks n' mortar company, it used to be that the marketing manager could just hand over big chunks of budget to an agency and watch them splash it

This is the story that online marketing gurus love to tell but it's not the case at all. It fits their narrative or they just don't know how it is measured. The reality is that digital has a lot of data but it's not actionable at all. People in digital like to think it is, but it just cannot compare to downstream POS data, for example. And for a LARGE portion of the business models, digital just could not provide the marketing-measurement feed back loop for a long time. It is just now starting to come of age and has a lot of catching up to do.

And there are plenty of lunches in digital too.

Another interesting discussion [1] about the notorious VPAID that so many people here were outraged about.

Turns out, HN is full of it too.

[1] https://www.reddit.com/r/adops/comments/4o0v2b/the_cancer_of...

https://news.ycombinator.com/item?id=11900453

Does geocar's comment in this link - https://news.ycombinator.com/item?id=10271348#10273922 , talks about the same thing.

I was earlier reading the above link, and also the links you posted, and was wondering whether geocar was talking about VPAID ads.

Edit: corrected the link

Title needs a (2015). Old thread.
As someone who is working on an attribution platform, I'm always amazed at how few marketers use real performance measuring tools (like attribution modeling).

Fraud should be irrelevant (well, a network problem) if you're measuring against your actual returns. It's the immaturity of the industry (brands especially) that allow fraud to be rampant and sometimes network-sponsored.

Ah, attribution. Another great idea that sounds sooo good, it must be true!