At scale? Margin of error at best. I will tell you what my interests are by vising web properties that cater to them in that specific time. You deciding to keep showing me ads for a Nespresso machine when I'm reading about circular saws is idiotic.
That doesn’t pass the sniff test. Would you say that in general, you wouldn’t expect a difference in results if you sold Taylor swift albums to white suburban women in Iowa, versus black urban men in Atlanta?
You know what would blow both of those out of the water? Selling it to those that have indicated that they are interested in Taylor Swift as white suburban women in Iowa do not buy Taylor Swift albums. They buy mom jeans. Instead they are getting ads for Taylor Swift.
How do you increase your market past people who have already liked Ms. Swift?
The argument still stands - just replace TS with mom jeans. My point wasn't actually about Taylor Swift marketing - it was that demographic targeting is a reasonable way to identify an audience. Those mom jeans are not going to sell very well in a younger, urban demographic compared to Iowa.
Ebay didn't stop buying targeted ads, afaik they stopped buying a specific type of ad in google search that had the keyword 'ebay' in it. Most companies bid up their own searches with the theory that they have to or else their competitor would, but ebay showed that people who search 'ebay magic cards' would most likely skip the search ads and go straight to ebay.
No they completely stopped buying ads. The brand keyword experiment gave them the confidence to run an experiment to completely turn everything off.
>TADELIS: Yes. So, for non-branded search, we actually had no idea what the results are going to be. Because here, if I am searching for, example, a studio microphone I’m sure that on eBay I might find a variety of used ones. But if I’m not thinking about eBay, and I just search for “studio microphone,” if eBay doesn’t pay an ad, they might not even show up on the first page. And by the way, the automated machines at eBay that were doing the online bidding, they had a basic library of close to 100 million different combinations of keywords, because eBay has practically everything you could imagine for sale on the site. So, we really had no idea what the returns for the non-branded searches would be.
>TADELIS: And we took a third of these D.M.A.’s, and we turned off all paid-search advertising. This was an extremely blunt experiment where we’re saying, “What would happen if we didn’t advertise at all?” And to our surprise the impact on average was pretty much zero.
Given that ebay is a worldwide company that is relevant to pretty much everyone on the plant, it seems like they are exactly the case for targeted ads being least useful.
Nearly every company in the world does not fit that description, and I would bet that the vast majority of them would benefit from targeted advertising. One example being local stores targeted only to local people.