>2) Companies want to know that their advertising dollars are actually producing results, and traditional advertising completely sucks at this. If you've never run a traditional ad campaign, I suggest giving it a try and then decide how much value you think you got out of it. You can spend thousands to millions of dollars and have absolutely no idea if your money impacted sales even the smallest amount. Compare that to adtech, where you get real data on outcomes on a regular basis. Even if it's crap data, it's often better than "well we think the brand is stronger now".
The data to do this already exists in many cases.
For example, Google, Facebook - both know when people visit a location, and could cross-reference this with store visits now. If advertising exists in a physical location, and is effective, there is the potential for either of these companies to determine whether the customer visited a location where that physical inventory exists.
Traditional advertising only sucks at quantifying results because "new media" has no incentive to share the loads of data they are now collecting.
But how do you know if they were exposed to the advertising? You don't. Even if you did, how do you know they actually saw it/were paying attention? You don't. And even if you knew that, how do you know that it was because of that particular ad that they ventured to your location?
This is my current hesitancy with "view through conversions" and the industry's rather, optimistic, definition of an impression (especially when running remarketing with VTC... essentially paying to target people who have quite likely found us via organic means and aren't being influenced by our ads). Currently trying lots of experiments with direct, tracked clicks from specific ads through to the CRM. Seeing only a handful of ads get any kind of direct click > conversion response with a reasonable attribution window.
So a larger scale programmatic campaign is either a much earlier part of the pipeline than a lot of the adtech companies would have you believe, or is fairly pointless.
The data to do this already exists in many cases.
For example, Google, Facebook - both know when people visit a location, and could cross-reference this with store visits now. If advertising exists in a physical location, and is effective, there is the potential for either of these companies to determine whether the customer visited a location where that physical inventory exists.
Traditional advertising only sucks at quantifying results because "new media" has no incentive to share the loads of data they are now collecting.