Just left a data engineering job in adtech. Speaking from personal experience, marketers make terrible ROI analysts because they always want to spend the same or more. Agencies are especially bad.
Most incremental lift studies I've seen are valid, looking at CTR and A/B testing as you mentioned. But as soon as you start looking at revenue and brand lift, all studies seem to come to the same conclusion - the higher the spend, the better.
Don't get me started on multi-channel attribution... garbage.
Marketing is still very much spray and pray with diminishing returns. And in most cases, marketers will claim otherwise.
I got tired of drinking the koolaid. Modern marketing is just as fraudulent as ever. Agencies are just getting better at lying.
Agencies are thieves. At WPP Google wouldn't play ball with kickbacks, but you can't be a credible digital agency and not do search. So Google got their taste, just no GDN. Facebook wouldn't do kickbacks, so they were heavily discouraged.
Any DSP that could advance client goals but wasn't part of the inner circle would be overruled by senior management. Total gaslighting.
Everything goes to AppNexus or AppNexus partners. Because the parent company owns 20% or so of AppNexus. I got downvoted yesterday for this but I'll say it again, the mafia is in the tech, the tech is in the mafia. I have so many stories.
edit: The funniest bit was when I would take thoughtful, younger clients out to dinner who understood the business model. They would laugh and say "don't tell me too much, just hit the artificial KPIs and make me look good for my Christmas bonus.".
Speaking from personal experience, marketers make terrible ROI analysts because they always want to spend the same or more.
Yeah, principal-agent problem. At a previous job there was a case where we sent partial refunds to advertisers because we noticed we hadn't delivered as many impressions as they had paid for. Some of them were upset that by that, and would have preferred that we had just kept the full amount.
Most incremental lift studies I've seen are valid, looking at CTR and A/B testing as you mentioned. But as soon as you start looking at revenue and brand lift, all studies seem to come to the same conclusion - the higher the spend, the better.
Don't get me started on multi-channel attribution... garbage.
Marketing is still very much spray and pray with diminishing returns. And in most cases, marketers will claim otherwise.
I got tired of drinking the koolaid. Modern marketing is just as fraudulent as ever. Agencies are just getting better at lying.