Half the ads I see on the internet are for pickup trucks, something I'm clearly not going to click through & buy.
Interestingly enough, I did buy a truck I don't need recently...
Much as I find the practice annoying, a large stadium or other facility endorsement programme has a sticking power for brand awareness which a microtargeted, microsecond-auction online placement doesn't. Much the same as investing in streetfront retail space speaks to a certain confidence in your own products or services.
Don Marti and Doc Searls have both written on this many times and at length.
Assumedly that measurement can lead to more impactful ads, but we are awfully new to programmatic advertising for that to be proveable.
I do know that advertisers believe targeting is valuable.
Better IMO is to look at broader measures of response. Multiple campaigns with various metrics -- coupons, response codes, promotions, differing Web URLs, etc. See what has a better measured response as an aggregate on desired outcomes.
The problem with ever-more-intrusive-and-objectionable advertising is that it turns people off.
I've largely abandoned commercial media, will consciously avoid stores whose advertising (or products) are annoying, and keep winnowing down the list of mobile service providers who are even vaguely acceptable to me (if I can possible swing it, I don't carry a phone), due in part to consideration shown their own customers.
I'm hoping that eventually those negatives will be large enough to sway practices. Google are actually in a position to have an impact, I've encouraged them on this before.
The question wasn't "What would you do if you were looking for something to buy".
The question is: "What would you do if you were a business and had something to sell?"
Make a high quality product.
A trusted third-party reviewer (Consumer's Union, CarFAX, Kelly Blue Book) helps.
Anecdote: the last really big-ticket item I bought (automobile), I'd had positive experiences with two manufacturer's products previously, but really didn't care for their current offerings within the product class / price range I was seeking. The vehicle I did buy was one I considered based on its Edmunds review -- best in class, multiple years running. Some years on, I've been happy with the purchase (though of course, post hoc rationalisation is a common fallacy).
For other products, other signifiers matter.
On G+ I detailed frustrations looking for a decent LED cabinet-lighting system. I'd exhausted local, online (Amazon and major hardware/lighting store) sources. A recommendation from my G+ contacts was for Ikea. The specific product didn't work, but Ikea had another set of lighting components (separate cord, transformer, remote switch, and lighting elements) which has worked excellently. Took a few weeks longer than intended to find what I was looking for.
Otherwise, I rely strongly on what local retailers will carry and stand behind. Ikea, for some products, turns out to be a good source. Factors depend heavily on what it is I'm purchasing and just how it will have to function.
High-quality product is nothing without leads (i.e, potential users, buyers, downloaders, etc), and one way to get consistent, reliable way to get leads is through promotion.
Then there's branding and awareness, which is probably what most ads these days are meant for.
If I'm looking for something to buy:
1. I'll search for it directly. In particular, I'm interested in solid reviews. This quite often means not giving much credence to review mills -- CNET, Yelp, etc., -- where the nature and intentions of reviewers are quite often suspect. Amazon have this problem as well, though to a very slightly lesser extent.
2. Map-based systems (OpenStreetMap, Google Maps) are often what I use when seeking a specific option locally. I suspect thes are underappreciated.
3. Classifieds systems. There's a reason Craigslist took off like it did, though its quality/relevance have been lagging for some time. It's still ground zero for housing ads.
High quality recommendations mean a lot. I happened to be reading a blog through one of my rare unfiltered browsers, and realised it included book recommendations on the sidebar, for which the blog editor was highly qualified to offer assessments and endorsements. The state of advertising is so bad that I had missed these previously. The blog would be better off putting those recommendations into posts or other content less likely to be filtered. First time in quite literally years that I'd had the least sense of missing something.