Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

How can we know if paid search advertising works?

causalinf.substack.com

Post reply on HN
Print view View original

101–110 of 112 posts

Re: How can we know if paid search advertising works?

#101

Post by dotancohen »

somedude895 wrote: post #88
vgeek wrote: post #2

The article fails to mention that shortly after this made the rounds in the marketing communities, eBay suddenly got a very large SEO penalty. Conspiracy theories aside, I once inherited a large Adwords account with nearly 25% of spend dedicated to branded KW bidding. Simply dialing back bids by 90%+ dropped impression share from 95% to 80%, but the paid branded clicks just shifted to organic branded clicks-- confirm…

> the majority of brands are simply paying protection money Exactly this. Google should be forced to not allow bidding on other companies‘ brand kw.

Should Basecamp and Trello not be allowed to advertise on searches for "Jira alternatives"?

Re: How can we know if paid search advertising works?

#102

Post by dotancohen »

rndphs wrote: post #87
RubyRidgeRandy wrote: post #79

Earlier quoted context omitted.

Except your storefront is on Google's land...and people find your website using their services...and you are free to not ever do any kind of ppc or seo work on your site at all.

It just so happens that Google's land includes all of the main streets where people shop. But, you are free to set up your store on a backstreet alleyway a few miles from the town centre. That way you won't have to pay the protection fee. That's some real freedom right there.

In your analogy, are not stores on main streets taxed higher than stores on back alleys?

Re: How can we know if paid search advertising works?

#103

Post by notreallyserio »

dazc wrote: post #84
bigger_cheese wrote: post #43

Where search advertising works well for me is 'service' related advertising. As an example I have a burst pipe and I need to call a plumber out to my house. 10 years ago I'd go to the physical phone book and look up a plumber to call. I don't do that anymore now I type "my suburb name + plumber" into my browser. Where it does not work well is because I searched for the term 'plumber' every site I visit in my browser…

Maybe 'plumber' isn't the best example since any plumber currently generating business via paid search is probably the kind of plumber you would want to avoid?

I'm not so sure. They probably have a dispatch line and a scheduler that can get you on their calendar inside a few weeks. That would put them ahead of most plumbing businesses, IME.

Re: How can we know if paid search advertising works?

#104

Post by sweezyjeezy »

RubyRidgeRandy wrote: post #80
sweezyjeezy wrote: post #69

Earlier quoted context omitted.

> Does that mean that major brands with entire departments filled with analysts running their paid search efforts can be debunked in a couple thousand words The old "all the successful companies are doing it, so it must be doing something" argument. Even within companies, incentives are skewed - do you really think a CMO would ever say, “maybe online ads don’t work, I don't need so many staff reporting to me anymore,…

You can talk to pretty much any ecommerce site or ppc specialist - open up linkedin and message away. I had clients triple their yearly revenue with well implemented ad campaigns. Do you really think these things are implemented without any kind of tracking in place? We have people who only run ads with a targeted ROAS of 10x. That's not something you can do or track effectively with a magazine, billboard, or radio a…

Well that's pretty far from what I asked for, more appealing to authority and unpublished internal analyses (presumably run by marketing departments). I used to work for an ecommerce A/B testing company, and have seen first hand how untrue it can be. Google doesn't make money because it's damn effective, they make money because people believe it. There is published research that that show that it doesn't work as well as Google or internal marketing teams want people to believe it does (did you read the article, or just count how many words it was?)

Re: How can we know if paid search advertising works?

#105

Post by toto444 »

withinboredom wrote: post #100
toto444 wrote: post #99

Earlier quoted context omitted.

That would mean you can just buy 'hotelsincornwall.com' and no one will be allowed to bid for 'hotels in corwnwall'. You are now facing a different problem.

I don’t think you’d get a trademark for that name…

Why not? 'Apple' is a trademark right?

Re: How can we know if paid search advertising works?

#106

Post by RubyRidgeRandy »

sweezyjeezy wrote: post #104
RubyRidgeRandy wrote: post #80

Earlier quoted context omitted.

You can talk to pretty much any ecommerce site or ppc specialist - open up linkedin and message away. I had clients triple their yearly revenue with well implemented ad campaigns. Do you really think these things are implemented without any kind of tracking in place? We have people who only run ads with a targeted ROAS of 10x. That's not something you can do or track effectively with a magazine, billboard, or radio a…

Well that's pretty far from what I asked for, more appealing to authority and unpublished internal analyses (presumably run by marketing departments). I used to work for an ecommerce A/B testing company, and have seen first hand how untrue it can be. Google doesn't make money because it's damn effective, they make money because people believe it. There is published research that that show that it doesn't work as well…

I suggested you get literal direct experience from people who have used ads in their marketing mix. I gave you my direct experience. This is not an appeal to authority.

You then immediately appeal to authority by your own definition by bringing up your experience. Working in an A/B testing company is a lot different than doing A/B testing. I create and implement digital marketing strategies for companies. We do A/B testing as well. So when a client is making over 6x return on their ad spend, how is that 'not working' when the baseline has already been well established?

Re: How can we know if paid search advertising works?

#107

Post by RubyRidgeRandy »

rndphs wrote: post #87
RubyRidgeRandy wrote: post #79

Earlier quoted context omitted.

Except your storefront is on Google's land...and people find your website using their services...and you are free to not ever do any kind of ppc or seo work on your site at all.

It just so happens that Google's land includes all of the main streets where people shop. But, you are free to set up your store on a backstreet alleyway a few miles from the town centre. That way you won't have to pay the protection fee. That's some real freedom right there.

That's interesting, so google ads don't work yet simultaneously if you don't use them your business will just break down into nothing. How does that work?

Re: How can we know if paid search advertising works?

#108

Post by sweezyjeezy »

RubyRidgeRandy wrote: post #106
sweezyjeezy wrote: post #104

Earlier quoted context omitted.

Well that's pretty far from what I asked for, more appealing to authority and unpublished internal analyses (presumably run by marketing departments). I used to work for an ecommerce A/B testing company, and have seen first hand how untrue it can be. Google doesn't make money because it's damn effective, they make money because people believe it. There is published research that that show that it doesn't work as well…

I suggested you get literal direct experience from people who have used ads in their marketing mix. I gave you my direct experience. This is not an appeal to authority. You then immediately appeal to authority by your own definition by bringing up your experience. Working in an A/B testing company is a lot different than doing A/B testing. I create and implement digital marketing strategies for companies. We do A/B t…

I'm just trying to say to you - yes I also have direct experience with this, so don't try and give me some anecdotes. We ran meta-analyses across hundreds of clients / experiments using advertising, both retargeting customers who abandoned their baskets, and consulted on more traditional experiments as an independent auditor. These were mostly in the fashion and travel verticals. The revenue uplifts we saw as a whole across the sectors were poor, expecially compared to other things that these sites were doing (social proof, scarcity etc.) I never once saw a value like 6x ROI. We could not publish this publicly unfortunately, as it more generally showed that stuff people were A/B testing doesn't really do much, which would have been bad for business.

This does not mean that search advertising is useless as a concept. But the idea that it 'just works' is also clearly false.

Re: How can we know if paid search advertising works?

#109

Post by withinboredom »

toto444 wrote: post #105
withinboredom wrote: post #100

Earlier quoted context omitted.

I don’t think you’d get a trademark for that name…

Why not? 'Apple' is a trademark right?

Apple is a distinct name in that industry. You can call yourself Apple and sell toddler toys without violating any trademark. You can’t call yourself “X in Y city” because it’s not distinct in that industry unless you literally are the only unique one in that city.

Re: How can we know if paid search advertising works?

#110

Post by cwkoss »

nradov wrote: post #75
cwkoss wrote: post #61

Earlier quoted context omitted.

https://veritasbusinesslaw.com/trademark-and-google-adwords/ Looks like several companies have tried to sue Google for this, but Google threw a lot of money at the problems to prevent them from winning.

Trademark litigation can be expensive, but you can't win every trademark case just by throwing more money at the problem. If there was a legitimate IP violation then someone would have won a case against Google by now.

I think that's a very reductive take.

The idea that I can buy placement in searches for a competitor is not clearly a "legitimate IP violation" but I think it's also not clearly "not a legitimate IP violation".

Is it good for society that I can redirect users who intend to be looking for a competitors' products? Its a complex subject that I'm not sure the courts have examined the repercussions of properly. The counterfeit and deception arguments in some of the suits certainly seem like they have some merit to me.

Post reply on HN