Hacker News new | past | comments | ask | show | jobs | submit login
Google AdWords charged me for clicks in Istanbul while location-targeting the US (medium.com/siggi)
419 points by siggi on Nov 29, 2019 | hide | past | favorite | 172 comments

Google changed it lately without any special notification. Unfortunately. There are more things to watch out.

For example if you set a budget of 100 dollars per day, Google can, and most likely, will, spend 200 dollars in short campaign per day.

Another thing: in display campaign, on default, most of the budget will be drained by mobile fraud applications that generate fake clicks. And also by video clicks made by babies on YouTube. And Google removed an option to disable mobile apps altogether lately.

Keyword targeting - on default Google uses broad matching including variants etc. Eventually you can pay for clicks for non relevant keywords that will also drain most of the budget.

Another thing: Google advises to put ads on your company name. It is a neat strategy, because you pay for something you had for free. And campaign seems to have better statistics.

Another thing. You wont believe it, untill you see it. Aprox up to 80% of ad traffic are bots. And advertising platforms does not recognize it or refund.

And so on... I am working on Google Ads campaigns for several years now. Always say to my client two things:

1) Google and Facebook will spent every penny you want to spend

2) Spend on ads only such amount on money, you can live without, and keep it running for a longer time without interruption.

3) Optimise often

I am a Google certified premier partner managing hundreds of thousands of dollars in spend a month for Google AdWords... Google ads have become a giant scam over the last few years. Be extremely careful, they are constantly changing things in ways that border on outright fraud. Google used to be a company I respected and loved, I am disappointed to see what's happened recently. It would be great to see someone with integrity step up and set the company back on the right path.

Oh yes, those baby clicks on YouTube. Annyoing as hell. Stay away from broad affinity / in-market groups or you'll end up spending 20%-40% of your budget on channels for toddlers.

By time though you'll build up a nice list that you can exclude, but it's going to cost you in the start.

> Oh yes, those baby clicks on YouTube.

A friend's toddler was watching a cartoon with songs on YouTube, and they were deeply focused on it when an ad popped in their face. They didn't stop crying for an hour.

By excluding channels for kids from your ad campaign you can both save money and avoid scaring children, so go for it. :)

This killed my budget on a display campaign I setup last year and there seems to be no easy way to mass exclude these channels. I even tried setting the content of my ads as not family friendly but still found this didn’t solve the problem for retargeting campaigns.

That is exactly why i dont use display. Eventually you can set reach with frequency set to 1 and lowest cost per view possible. And each our ban all junk. There is a lots of it, so it is impossible to remove all. But after two, three days, you get a nice list of websites where actually you want to display ads. Than you can show ads only on these pages.

Other way is to use keyword setting. Not as topic or observation, but page containing a keyword. With good set of keywords it is possible to rule out baby clicks

> if you set a budget of 100 dollars per day, Google can, and most likely, will, spend 200 dollars in short campaign per day.

How is this possible? It sounds like a bug, but it's probably not, so what's their justification?

Launching an ad campaign feels like rocket science now mixed with a good amount of woodoo magic. 80% down the drain by default? That's a number I could believe, but freaks me out how to market my side project to get actual people on my service. And I'm on a ridiculous budget, I can't afford the "throw money on the wall and see what sticks" method with 500$.

As I understand it the reasoning given tends to be that ad auctions happen extremely quickly, and therefore too quickly to check some single source of truth for your current budget.

Therefore, the budget is only "eventually consistent" - i.e. at the end of the campaign or after a day or so, the number you're being charged is accurate. However during the campaign itself it's not possible to guarantee that things won't go slightly over budget as each individual ad auction cannot feasibly check the central budget.

That said, it definitely feels like there should be a way to implement this such that it backs off as the budget is approached, so that overshoots are likely to be minor, rather than 100% of the budget as apparently happens on a regular basis.

I'm pretty sure fixing bugs that make money is on the very bottom of the list of priorities....

They don't need a bug fix as a stopgap. The fair thing to do would be refund the excess. If I set a limit and they display ads and exceed it, why should it be my problem?

Excess is calculated at end of month not daily

Back when I was in the ads org, there was no month without a big important future launch, that was supposed to make us make less money in the short term.

I suspect a factor that makes this trickier is cost per click. You're bidding for a click not a particular spend, so if your cost per click is ~$0.01 then overrunning by 100 clicks only costs $1 over budget. If you're more like $1-5 a click (not uncommon) then you can easily get to $100 over budget.

My guess is that this is worse the lower your budget, and $100 a day is a low budget for Google ads when you're considering the whole industry.

I would expect that with a budget of, say, $5000 a day, you might still be ~$100 out.

No. They overshoot because there’s a small-letter asterisk within the budged definition that allows then to overshoot at twice the amount.

I wouldn't use the word likely. It might be because of who your target group is etc and I therefore have another experience. But yes, it does happen. The way it works is that Google allow themself to spend up to 2x your daily budget. If they exceed your daily x2, you'll be credited any overcharged amount.

Eg. Your daily budget is $200 and they spend $420, you’ll get a $20 credit.

Then there is the monthly limit, which is your daily budget x 30.4. Also here you'll be credited any amount they exceed.

So basically it sounds like the "budget" is actually more of a "target spend", that they will try to approximately hit in the short run, and hit very accurately on average over a month.

what's the point of that? knowing that, why wouldn't i just set the budget to $50 and avoid the problem that way?

Most likely the algorithms that chooses whose ad gets shown are optimized to not have to look up each persons spend for that day, as syncing that between all servers on every page request would be more expensive than all other steps combined.

I'm not sure how they could let you overrun by several times your budget for a day however, syncing these every hour shouldn't be a problem.

I have explained it in other comment above. Its not technical issue. Its by design

Google explains that on the long run, like,30 daya of campaign, it will be proper on average. The explanationis, that Google will spend more when it sees possibilities to give more value in specific time. For example, there is a special day when people buy more than usual, but later they dont. So there is more budget spend on this special day, and less on others to make it even.

So there is a reason behind it. But its something well hidden. You have 500 dollars for a one week campaign, and it occurs you pay 1000 dollars for it. It does not seem to be right.

There is no reason to that aside from Google making more money. They are not your friend.

Actually yes. Google wants to have advertisers who spend fixed amount each month without going much into details. They dont care much about anything more

For example keyword targetting gets broader.and broader. If you target.for keyword you want to target it precisely or with synonyms but you want to have control over it. But Google changed it lately to make it harder to narrow down keywords.

The result is you spend budget on bad keywords. And it seems like ok,if i get money from it overall.

But actually not only you loose budget but others too, because they need to bid higher.

They need.to bid higher for keywords you dont care about but they care. For Google its additional money.

It's not a bug per se and is designed this way. The ad budgets are calculated on 30 days basis. So if you set the budget to $100 per day, Google will display ads with monthly budget of 100*30 or $3000. The objective is to suck out as much money as possible from the advertiser very quickly.

The only assurance is that your monthly budget won't exceed $3000

I've been seeing 80% of clicks as fraud/nonsense/fake/whatever for a decade.

> And Google removed an option to disable mobile apps altogether lately.

can you adjust bid to -100% for mobile?

It excludes websites. And nowadays you dont want to block ads on websites viewed on mobile devices. You want to block Android ad fraud apps. Google blocked the possibility to do it. Essencially making fraud apps creators extremely easy and luxurious.

And as you can expect, Google has 10 times stricter requirements for websites showing their ads, than for Android apps...

Minimum modifier is -90%, so you only can penalize mobile.

I can personally confirm another f'd up Google Display Ad scheme... When looking at the list of mostly gambling sites that their almighty algo decided we would have most success in for a niche UX platform!

So why even use a platform that so easily throws your money away for nothing? With those points you laid out I would be better off keeping my money.

Facebook is the same way with all of this crap too. I’ve seen more success with the hard way, networking and getting content made about me and distributed in magazines or online blogs from people who get to use my service for free.

Because the money it doesn't throw away for nothing more than covers the expense.

Sales aren't some abstract thing that can't be quantified.

Site gets no traffic -> Spend $X on google or facebook pointing traffic to site -> Y Amount of traffic -> $Z amount of sales.

You now know how much revenue $X worth of ads buys you. You also know how much each amount of revenue is worth to you in profit. As long as spending that $X makes profit, you're gonna do it. No matter what crap is mixed in by the ad company.

If something you want is bundled with something you don't want and is priced such that you're willing to buy the thing you want at that price, you don't turn it down. You buy it and ignore the thing you don't want.

And it is best you can do. Essencially Ads should be the last resort and used with extreme care to not waste a dime. And this takes a lots of focus unfortunately.

Keep the budget low, so you wont bother about it, and use money for anything else than Google Ads.

And its not like i am against Google or something. Actually i have a company that runs Google Ads for clients. So i dont have any hidden interest to criticize Google Ads, contrary. But its the truth and its just something i think everyone should be aware of.

I think it best to think about the ads in the same way you would in any other medium. Unless you have a special offer or something, it's just branding. The original hypothesis upon which Google was founded, that people would be interested in buying things if they were searching the internet for something relatable, seems to have failed- at least in the near-term sense. Just because I am looking for campsites in the Rockies doesn't mean I am in the market for a sleeping bag. But to put the ad in front of me then to make me aware of your brand might be useful. That's the game played by advertisers since they began with media advertising.

Given this, all of the data collection/tracking/etc. seems like a round-off error in effectiveness, and doesn't seem to be worth the loss of individual privacy.

Yes, for near terms its branding. In theory it should make you consider the company when you move to buying stuff stage. But the theory also tells you would need to be exposed to the company with multiple channels. Search counts as one. So it does not work solo.

Given their monopoly power, this is beginning to sound increasingly like a tax on businesses.

Google bundles bad clicks like a pile of stinking, rotten subprime mortgages and then peppers them with some good clicks and sells them to you like JP Morgan in 2005.

Did you read the fine print? Doesn’t matter, you can’t redline it anyway so it’s take it or leave it.

Their whole ad empire is desperate for a massive audit. Google had to be sued to deliver a $400,000 refund they owed but later claimed couldn't be delivered, claiming 3rd parties had been paid the funds when it was their own platforms. It turned out Google had accidentally kept $75 million.


As someone who has adblocked everything since my teenage years, the whole ad industry seems to be making orders of magnitude more money than it should. I just don't understand how anyone comes to buy things shown in ads.

Are ads really effective? I don't buy advertised products.

Do we really trust ad tech accounting? They've been caught inflating numbers over and over again.

The whole house of cards seems like it could fall over if actually investigated. Google, Facebook, Twitter - every single one of them.

Influencer and direct social referral advertising seem much more "legitimate" than injected ad impressions.

How can we get the thing to implode? Google has destroyed the web to prop up its ad business. I'd be happy if they couldn't cause any more trouble.

To my surprise yes ads are very effective. Without getting too specific, my close relatives run their own business as home improvement contractors. In the early days, before they had their reputation built, I setup a google adwords campaign for them — having never run an ad campaign before.

It was a simple PPC campaign that included keywords in google search results and geofenced it to the SFBay. Here are some numbers and findings:

- PPC costs were about $1-2.50/click with a ~$15/day max spend

- Selected cities by hand to fit the service area

- We did in fact get clicks outside of the zones we selected. Many if them were B2B trying to sell their services.

- Customer acquisition costs specifically for the google ads campaign was ~$70 (clicks/conversion)

$70/customer sounds like a lot except these home improvement contracts were typically thousands of dollars, so the campaign was a fantastic ROI. Furthermore, we dialed down ad spend as the reputation was built, and now most leads come via referrals and review sites.

In the end all that mattered was that customers were signing contracts so even if there was some funny business going on it was still worth it for our particular case.

That's not exactly a shining endorsement...

"Yeah, Google was robbing us blind and kind of scamming us with their geofencing shenanigans, but we were making enough money that it wasn't worth fighting, and eventually we figured out how to stop buying ads from them."

And I think that's exactly how adtech works, both internally and externally. Everyone is scamming everyone else with fudged and pretty opaque statistics, but there's enough ROI at the bottom that the whole thing hasn't collapsed yet.

So what you are saying is that all that is needed is for a newcomer to come and do things a little bit better and it will eat everybody's lunch. Just like google did to search engines when they started.

Sounds like a great business opportunity.

The core innovation that would be needed here is providing such good transparency and attribution that such newcomer could turn around and prove that all their competitors are scamming their customers. Without that, fudging numbers will prevail simply because it can make money without having to do as much work.

But in order to run a business you have to get customers from somewhere. And referrals aren't an option until you already have customers.

I don't think it was supposed to be an endorsement.

If I could pay just 70$ per customer bringing in 1000+$, I would spend all I could and borrow heavily to run those ads. From a business perspective, it's a great endorsement.

In this case they were selling skilled labor though so at some point their capacity would be reached. I'd guess that's why they only spent $15 a day.

This is how almost any business deal works. If you get more money out than you put in you do it.

Sure, the partner may be hard to work with, but so is almost any big tech platform provider (... Apple ...) and it can be quite profitable to put up with.

It only works if you sell a hugely priced item/service.

If you want to sell everyday consumption or end consumers items, this scheme will eat all your income/benefit.

After reading about all the flaws in automatic online advertising through clicks on this thread, it looks like a big scam prospering more or less in favor of opacity, lack of fine tuning or targeting and fraud.

Referral systems and social influencers seems to be a better option, as long as your product is good and would not shame the ones endorsing it. Betting on viral marketing and endorsement seems to be more natural and a better bet, but require maybe more work and preparation than blind automatic system.

Yes. As an advertiser, who traces each ad, I can say that advertising has demonstrably made me thousands upon thousands of dollars.

However, 80% of ads are trash peddled by blind men leading blind men.

There are very specific advertising techniques that work consistently, but most "marketing types" optimize for flashy artistic ads that have literally ZERO selling power.

In fact, in some studies, ads actually REDUCE sales -- by, for example -- accidentally highlighting something that is an unexpected turn off to customers.

But, the problem is, most companies don't trace their ads. And that's why overwhelmingly a lot of advertising money is wasted.

But this state of affairs has been going on since the 60's.

So I wouldn't underestimate the number of gullible humans to keep paying for advertising.

As long as you have eyeballs you'll probably always have an excess of ad spend.

Basically you just rattled off everything I, as someone who never worked in that industry, assumed was going on.

There's just a shit ton of money to be made pretending it's 1955. I went into the wrong business.

just out of curiosity, what are some specific advertising techniques that work consistently?

No specific tactic works forever. There is no "1 weird trick." You need to set up your advertising and sales funnels to track as prospects become customers, and continually experiment on the message and the channels. This is where most startups start and a lot of them get it wrong. If you have boatloads of VC money, you can run a lot of bad experiments until you hopefully hit paydirt. Paydirt is +ROI ad campaigns on a large targeted demographic. Then you treat this audience like a goldmine. At first it's easy to mine and pick up gold nuggets in river beds. As the easy pickings disappear, you'll find yourself investing in increasingly more expensive machinery and infrastructure to continue mining operations. As you've extracted most the gold, look for byproduct like copper, silver, etc (in startup terms, introduce ancillary products and services). As you hit one gold mine, keep prospecting other areas (distribution channels like display ads, native ads, social media, etc).

If you end up a multi billion dollar unicorn with tens of millions in ad budget, targeted direct ads no longer move the needle like they did when you were a startup. Now for the Director of Marketing to justify increasingly larger compensation, he needs to increase his ad spend and starts whispering about spending money on ads that have no direct correlation to the bottom line. This is called brand advertising. Hint: If an early stage startup spends more budget on brand advertising than direct marketing, run.

Effective advertising is matching the product/service with some problem, on a specific channel of eyeballs, with a specific message. There are tons of new channels popping up every few years (think google, facebook, instagram, youtube, twitch, pinterest, wechat, stackexchange, hn, reddit, etc). There are millions of permutations but you have a finite ad budget. The magic is deciding how to show what to who.

For the "how" and the "what", in the future, I bet ads platforms will ask you to write an long essay about what you sell, and they will automatically generate many different ads with GPT-2 (probably in real-time, to mix some content of whatever the viewer is browsing at the same time), and everything will be automatically A/B tested, segmented, adjusted, etc with reinforcement learning.

Even better, no need to write an essay about your services or products, Google already know everything about it by scraping your website, so just point to the URL, and done

So, there are no techniques that work consistently.

Correct, because everything is dynamic, including the search terms, eyeballs, and competition for keywords. It's pretty simple economics - once I find a way to spend $50 to make $100, it's only a matter of time before other people discover something close enough to make me spend $75 to earn $100. Then $90. Then (conjecture), winners curse takes account and people are spending $110 to make $100 (once Neil Patel's audience hears about it).

With the tactics being gobbled up by the Red Queen effect, you have to move to strategies to win. The OP writes about a pretty good one.

So... using voluum?

Brand awareness ads. Pretty much it’s irrelevant what you say about brand ‘X’ as long as you bombard gullible observer with ‘X’. Then buying ‘X’ would be more familiar choice for millions than buying ‘Y’ or ‘Z’.

I'm not saying you are wrong but brand awareness ads are, if I'm reading correctly, exactly the kind of wasted ad money the parent was talking about. They are the cluster bombs of the advertising world.

I think the Parent was actually advocating surgical PPC management of the style patio11 has written about extensively.

Branding is a moat. There's a reason Coke and Pepsi spend hundreds of millions on branding ads and I am sure they know there is a ROI there.

Brand campaigns are good if you use vast multichannel campaign to make people remember your product. I dont need to say it is extremely expensive but works.

Other reason to use it is when you dont want to be screwed by fake clicks, and are 100% sure people will find your product anyways and buy it.

And that was half of this post, the blog was as much about using their new product to track this problem as it was about finding the problem. Kudo's to solving a problem you find while making the product.

Yeah, it's funny to see a bunch of people complaining about both ads being ineffective AND conversion tracking being pervasive. But the reality is that unless there's perfect conversion tracking, garbage ads will be inevitable. This is why I have a (very unpopular) opinion of putting precise conversion tracking into the web standard (No, not the garbage one that Apple recently proposed).

Are they impulse buyers, or highly specialized? From time to time I see a cool thing that seems targeted to me, do click it, see the outrageous price, then do research and find the best price/quality ratio on Amazon. I assumed most people did the same really. An example from years back: I thought the Little Giant ladder was an awesome idea, but ridiculously priced. I bought a gorilla knockoff type ladder instead.

I'm not sure, but my girlfriend clicks on ads on instagram and buys things directly from there on a regular basis. I think the people who go on HN might be in a little bit of a bubble about the efficacy of advertisements.

HN is a bubble.

I've been running PPC ads for highly targeted keywords for the last 5 years which have generated a ridiculous return. In general ppc is steady and reliable once you have you're keywords and ads running for a while and are well optimized.

On the other hand, I've recently been trying dropshipping, advertising the product on Facebook.

I did a small test for a unique product, I sourced it from China for around 5 usd and sold them for 25 usd.

People actually bought them on site even though Amazon has them at a much lower price.

Facebook literally knows people who are high spenders / frequent purchasers and will put your ads in front of them. It's actually stupidly powerful.

Why haven’t you scaled out the dropshipping concept further? It sounds pretty solid.

I'm not in the industry so please forgive my ignorance, but what is ad tracing? Is it the same as ad tracking, if it is then I really don't understand as there seems to be a huge amount of tracking going on!

First, yes they are effective for effective marketing professionals. Digital advertising is a vehicle. The marketing creative is still the payload that does the work, and it has to be carefully and strategically created to maximize return. You can easily lose money if you don't know what you're doing.

Second, any marketing professional worth their salary is measuring with a number of tools to verify that the information they're getting from Google/Facebook/etc. is accurate. There are many metrics that only they can provide, but there are enough independently measured data points that fraud would be obvious.

Finally, you're blaming Google, but the advertising industry's current state is a product of extreme corporate greed (on the part of advertisers). It doesn't matter whether Google goes away, greed will continue to corrupt in other ways. Look what happened with the rise of ad blockers - native advertising, which is just a nice word for the paid corruption of editorial space, has become so prominent that you can't trust most of the articles you read on sites that were once considered reputable. The "influencer" marketing you mentioned is a perfect example. People with social presence and influence are being paid to lie about a product. You think this is somehow better or more "legitimate"? I would love to hear more about that. At least with Google, I can clearly see where the ads are.

So you can try to implode Google or any other facilitator of digital advertising, but money will find a way. It always does. You're trying to change human nature.

The old saw was that 50% of advertising is useless, it's just that nobody knows which 50%.

I guess that's why people are so excited about click tracking, analytics, etc, etc.

I think it's more like 99%.

When I was running advertising for a small business we put vouchers in to track advertising. The only people who brought in vouchers were people who were already customers who said they were already coming. So, we spent money to reduce our revenue. (I do the same, hunt for vouchers after making a buying/visiting decision.)

I was on national TV and national radio (UK) - none of that blipped our visitors/revenue.

Only thing that ever definitely worked was handing out flyers and chatting to people in the street (and being top of Google SERPs).

Funny you should mention vouchers - it reminded me of the article discussed here[1], which likened advertising with Google to handing out coupons to people who are already in your store


What were you selling?

Ads are massively effective.

My friends and I launched a physical product over the weekend last week. I put together a basic landing page and bought some Instagram ads. We spent about INR 250 (around $4) before I stopped the ads to do some optimization.

That $4 in spending brought us 1 customer who placed 2 orders worth $35. He also referred 2 other customers who placed an additional order of $35. $70 in revenue for the simplest of funnels is wildly lucrative.

The problem starts when you want to scale actually

> I just don't understand how anyone comes to buy things shown in ads.

What is so different about seeing something you are interested in from an ad or from a website you clicked through to yourself?

> Are ads really effective? I don't buy advertised products.

Do you buy many products online at all? If you don't, then they probably won't be useful as you aren't someone who spends much money online for anything. However, there are many people who do. For example, my girlfriend and many other women I know spend tons of money on clothes/shoes/jewelry online and a decent amount of the products they find out about and like come from advertisements.

To give you an example of the scale, I spent over $1,200 dollars online yesterday during black friday on clothes, shoes, and jewelry for her. While I don't know how many of those products got sourced from advertisements, I'm sure some of them were because most were from a doc she has been using to keep track of things she wants and she likely saw them first on an advertisement.

They are hugely effective.

Even if only a small percentage (e.g. <2%) click on the and, and then a similar small percentage actually buy anything after clicking, you are still making money so long as the cost-per-conversion (i.e. per sale) comes out low enough that you still make a profit. Spending $3-5 to make a sale in say a pair of shoes that retails at $100 with a profit margin of $30-40 is good business.

People are making huge amounts of money/doing huge amounts of business directly as a result of the ads, and that is not just Google and Facebook, but everyone all the way down to little mom and pop businesses.

When you track things down to the level of sales it very often works (not always - depends on products/business models etc), and that is why people are using online ads. AdWords and Facebook ads would have fizzled out years ago if people were not benefiting from advertising their business

The problem with this kind of accounting is that it doesn't consider whether the ad was really necessary to make a sale in the first place.

With a lot of brand keyword advertising, companies are paying for people to click an ad when they were already looking for your website anyway. So spending $3 on that customer is not a $27 profit but rather a $3 loss....

While buying your product name (or your competitor) is sometimes the case, this isn't the main source of volume in your pipeline.

And it's still a $27 profit, even if it is possible you could have saved money.

Most of the time a well-optimised PPC pipe looks for specific (non-brand) keywords and buys them, not just the product name.

Lets say it costs $3 to acquire a customer on average. And your margin is $30. That sounds good.

Now do an experiment. Stop your ads.

Do your sales go down 100%? In that cost the $3 cost was definitely justified.

Do your sales go down 50%? That means you actually paid $6 to acquire a customer, not $3. Still sounds like a reasonable deal since you still make $24 profit, but other channels might be more profitable.

Do your sales go down 10%? Then only one in 10 customers actually converted because of the ad, and you are paying $30 to acquire a customer and it would be better to stop the campaign. Google is just showing ads to people that would buy your stuff anyway. (Side note: I think this is a big problem with retargeting. You are paying to show ads to people that are already very likely to buy your stuff -- or even worse, who have bought it already. But conversion rates are really high!)

(To be fair, this is speculation and not based on experience with adwords. My only experience with adwords was a $100 campaign I made a few years back with promotional credit that resulted in no conversions)

While I agree that ads often are very effective, no it is just a $3 cost in some cases. Your ads can cannibalize on tge revenue from other, possibly cheaper, channels.

I think most marketers are aware that a lot (maybe half) of clicks are bots/fraud.

At the end of the day, if your ROAS is good, you don't care that much.

If Google’s ad business dries up, what do you think they will pivot to? It scares me to think about what a desperate company with that much data would do.

They have lots of cash but a pretty high burn rate too. Google has lots and lots of fat they could cut. Guess it would really matter how quickly the advertising business died as to whether or not they'd survive long enough to do something monumentally evil.

Google has been trying to diversify from just adtech for over a decade - unsuccessfully.

Still 90% of their profit comes from ads. They are much less diversified than Apple, Microsoft, and Amazon.

Maybe they're going to blackmail everyone. $10 a month or all your past Google searches get published.

Are ads effective? Yes. Are highly personalized and targeted ads more effective than, say, location+context? Dubious

It's well known in business that ads are generally dubious at best and likely mostly a fraud. Unfortunately, there isn't an alternative than doing all the work in-house. And, if you're good enough to do it in-house to make it a business, you'll easily figure out that you can cheat other people. IMHO, ads should be treated as a national resource with open accounting and a non-profit goal. This would instantly force companies to have to charge for services, pushing things back into competition.

Honestly, everything after your first sentence is completely incoherent. I'll stick to the first sentence.

>It's well known in business that ads are generally dubious at best and likely mostly a fraud.

You're using the phrase "it's well known" in an attempt to support a consensus that you completely fabricated. I spend over $100,000 every year on digital advertising, and I can assure you there is a very measurable return. Colleagues who manage much larger (and some smaller) marketing campaigns are also making measurable returns. You clearly have no experience in what you're talking about. So why talk about it?

There are lots of unsophisticated business owners and marketing "professionals" who think they can just throw money at digital and it will magically return. These same people don't understand how to effectively measure, so they may not even realize they're burning money. Like any failed venture, that is THEIR fault. In no way does that make advertising "generally dubious at best and likely mostly a fraud".

I have worked for a company that spent a multiple of your amount in a single day on a PPC advertising.

The fact that it works for you doesn’t mean that if that industry wasn’t as monopolised and unregulated you wouldn’t pay 1/10th of what you paid for the same.

google is just taking advantage of the old "Half of my ad money is wasted". They sell pretty dashboards to advertisers so they can pretend they 're working , and have a huge network effect among agencies now. very tough for a newcomer to find a crack

without competition, it's only going to get worse

The guy didn't read any print, not just the fine print. From Google AdWords help docs it says "By default, location targeting includes both physical locations and locations of interest"

When it says "location of interest" someone in Turkey can be interested products or services in the US.

The article even says itself that he didn't click on the advanced features until he asked his friend, after the clicks from Turkey. This is totally the user not reading the docs... and he used to work there... in that group for 2 years?

Also, Google and Google Adwords is heavily biased to be about relevancy, not mechanical-like billboard. Shouldn't be a surprise.

Replace 'Google' with any other large digital advertising platform and your comment would still make sense.

I've seen funded dating startup from US, which released their ad campaign on Facebook targeting 'Worldwide' for maximum clicks (thinking the clicks would be distributed), 90% of their users were from Philippines and they had to eventually change their product to Philippines only (no offence to Philippines, it's just not what the startup intended to do).

Then the marketing person lacks fundamental understanding of the platform.

Also, how do you spend so much cash and ignore the stats for so long that you have to change the target demographic for the app?

That actually sounds like ads helped them find a market, they didn’t come up with themselves.

a fine example of making lemonade

imagine if they said "lets pay double as much to try to target US users"

a) How do we know these clicks are bad? A sample size of 3 is pretty early to make any conclusions. b) If clicks are not delivering value, advertisers should lower their bids. As Google encourages advertisers to switch to bids set on hitting ROI targets, what really matters to advertisers is that they achieve their targets. If the clicks are "bad" and conversion rates are poor, then bids will be adjusted down, presumably by everyone.

The fact that prices are set by an auction makes the system inherently sustainable unless you think many of the bidders are irrational.

Define "sustainable".

Seems like it encourages companies to get in a bidding war against each other in a race to zero profit margins, while Google pockets what could have been their revenue.

Profit is already in the ROI target.

And good luck if you want to try and resolve a problem. There is no customer service to speak of.

I really wish there was a competitor to Google. They're one of the most opaque organizations around as far as their primary revenue stream - AdSense - is concerned.

Barriers to entry (network effects) :/

There's no obvious network effect in search.

Google does have brand and distribution though.

On targeted advertising the barrier is that you must already have a significant fraction of the personal data Google has before you can start to compete.

Not in search, in ads. The ad network itself. Many many many sites integrated Google Ads (AdWords, AdSense, Adwhatever).

Plus the sheer volume and reach that Google has. (It pays millions - if not billions summed up for every little channel they pay for - every year to be the de facto "internet" for many users, many kinds of devices.)

It's basically unimaginably hard to compete with this for newcomers. Sure, they can carve out a niche, and so on, but that's just that.

This is exactly the right comparison. Thanks for it!

I'm somewhat confused here. I live in the US but regularly visit Turkey to see family (say, once a year). I also regularly Google things related to the US when I'm there.

Clicks outside the US are not necessarily meaningless -- the screenshot in the article shows that they are targeting people "regularly in the US". Just because I left for a couple of weeks does not mean I abandoned everything related to where I live. There's still value in targeting US ads to me, independent of my current location.

I’m the opposite. An American living abroad. US targeted ads still are applicable to me. I can understand why someone might want to disable this feature, but it’s not a scam by any means.

> US targeted ads still are applicable to me.

Not necessarily applicable to the advertiser, though. If there's a US restaurant chain hoping for regular customers, they would be wasting their ad spend if you're on the receiving end of the ads.

The default option says that it targets "people in, or show interest in , your targeted location." The author tightened the settings only after he realized the clicks came from outside the US. Also, it's unnatural that multiple clicks from the same location were made in a quick succession.

The author tightened the settings only after he realized the clicks came from outside the US

It is worse than that, the country targeting setting was already at maximum tightness. This is a very bad dark pattern by Google, there is no setting to strictly target a country.


My guess is that they're getting "other Google affiliates".

What happens is a "publisher" combines CSE with Adsense then drives (often robotic) traffic to the ads that come back in the feed. They then identify high-paying keywords and make money.

Sophisticated attackers will observe that Google cares if the IP address matches the CSE request, but they don't care if the cookie and IP make sense (or if that cookie is being used elsewhere in the world). This means a sneaky extension can lift the (advertising) cookies of Google users elsewhere in the world and deliver them to places (e.g. Turkey) where it is convenient to run the click farm.

Sounds like excluding all other countries is the only solution.

Google does provide a CSV with all the AdWords recognized countries: https://developers.google.com/adwords/api/docs/appendix/geot...

You should be able to copy the list to the clipboard and paste into the exclusion list.

This is the answer.

I can’t tell you the times I have had to tell clients and other search marketing professionals that they need to have a negative list.

Budgets are not infinite and they need to be managed to get the most out of them.

I've mentioned this previously as a dark pattern that trips most people up.

It's probably fair to assume setting a target location is it, not that you additionally have to exclude the rest of the world. And it's buried in the advanced settings, so anyone looking for a quick default setup just do the minimum required setting presented to them.

Never rely on Google Ads to be honest. They are unlikely to respond at all.

After spending ~100k, our ads randomly stopped working last month without changing anything. Support took 10 days to reply saying the account was up for a review (no notifications sent, no indicators). Yet, it still wasn't working.

Booked a consultation with one of their so-called "Experts", he couldn't find anything wrong and suggested to contact support again.

After the initial 10 days response, still waiting for a reply from so-called support 20 days later. I'd assume support only exists for companies that spend millions a year.

Is it possible that your experience is an outlier, and not generalizable as advice?

I have read similar stories all over the community in general. The support simply doesn't exist or takes 2 weeks for a simple, useless response, generally.

Given your vested interest in Ads and Google in specific, I am sure you know this already.

What percentage of serious inquiries are responded to outside the SLA? What percentage of those fail to resolve the user's problem?

Without that kind of information, reading stories after having a bad experience will turn your brain into a petri dish for confirmation bias.

Ignoring the red herring, and the assumption I read these stories after a bad experience, let me address, yet, as a consumer, all that matters is the vocal user experience.

If there are 100 bad stories out there, that's enough of a failure of a company irrelevant of its size. And I have personally conducted the support experiment multiple time over the years to have the first-hand experience.

Either way, since you clearly have a vested interest in Ad companies from your HN comment history, I shall engage no more.

I mean, this all sounds sort of like a reasonable complaint... until you get to the sample size: 3 clicks.

It appears arguably reasonable that geotargeting isn't exact, nor is it clear that these clicks weren't actually real and valuable. If he's trying to make the point that the trend is disturbing, well, then: make that point with something akin to a real sample (at least 1000 clicks, spread over more than a week?)

Well... some people based in the US travel outside of the US sometimes.

That's the Google AdWords way ... by default, include people "interested in" the United States with no end date. So, lots of junk clicks at high CPCs from India, former Soviet Republics, and other places where the local language may not even be English. Turning these settings off requires drilling down into some obscure Google Ads interface ... it used to be AdWords Dimensions > Geographic, but not sure what it is now.

If you call them on it, the Google phone reps will say with a straight face, “Let's say people with family in Moldova want to look for something to tell their relatives in the U.S.”

The ethical way for Google to handle this would be to make all U.S. campaigns by default target users in the U.S., Indian campaigns target users in India, Moldovan campaigns target users in Moldova, etc. If businesses want to expand their campaigns to people living in other countries for whatever reason, give them the option but don't make it the default.

I feel especially sorry for the local small business owners trying to counter the SEO garbage and Yelp reviews that clutter the results. They try to set up the campaigns trusting Google's guidance, and they're lucky if 10% of clicks end up being target customers living in the same area. Often they have no clue about what's going on, and wonder why their campaigns are costing so much for only a trickle of real customers.

I had to do this for my midwestern US shed manufacturing business. Since we deliver up to 350 miles away, I set the geographic target to a radius around my business location.

Seemed to eliminate most of the junk clicks.

This type of behavior allows for a known ad arbitrage strategy. Bing’s PPC network, for instance, use to not have the same geo coverage as AdWords, and when they received traffic from a geo they weren’t serving ads in they would send the search to their US ad network. The strategy then is

1. Buy AdWords ads in countries out of Bing’s coverage,

2. Send these clicks to a page serving your Bing feed’s ads for same/similar keyword search,

3. Realize Bing US CPCs on international AdWords ad buys

But how much is your conversion rate for the transferred users. It is not like it is 100%. If it is 5%, you'll need x20 times the CPC to break-even.

no, because the Google ads would be served in low CPM countries (eg Turkey) whilst the Bing ads would be served as if the user was from the US (high CPM).

Of course, CTR is apart of the profit formula and cannot ignored as you correctly note. It’s not a trivial task to find those verticals or geos where this will be a successful strategy.

I mean it says right there that they may also be people regularly in the US. I don't see why I'm less of a marketable audience for a SaaS when I'm abroad.

I manage campaigns for a local, independent contractor. The business runs ads with a meager budget, <$500 a month. At the same time, clicks average around $5. The conversion rate hovers around 15% for those within the service area. That's a cost per conversion of ~$30.

For those who are outside of the service area it's less than <1%. While it may only receive a handful of clicks from outside the area a month, it eats into the already minuscule budget for the long shot of a person who decides to look up the service while on vacation or visiting relatives. It's not efficient ad spend for a business like this.

A SaaS business is a different case, but this issue is persistent and getting worse.

The problem is, it also includes anyone who has ever logged into a US VPN. There is no doubt that this will catch some Americans who are abroad, but if you are getting 3-4 non-US clicks to get 1 American abroad click, you are effectively paying 4-5x for that one useful click.

So is everyone else, so in theory the auctioned price per click should be less to compensate.

People don’t pay for clicks, they pay for results. Clicks are an intermediate measure. If I make $100 from $50 spend via 1000 or 100 clicks it’s pretty much the same thing.

A bit OT, but does anyone have any advice on improving page relevancy quality scores?

My site is focused only on relevant content in a particular infosec niche, and is well written - yet quality scores across the site are 2-4, and have been for the entire site history, which spans almost 20 years. Which means we pay more for ads.

I've been through the usual guides, and I can't find anything actionable - they all basically boil down to "have relevant content" - but I do, dammit!

I've also taken up Google's frequent offers of free help from an expert on 2 occasions. Both times all they did was try to get me to increase my spend (e.g. enabling the search network etc), and parroted about "writing relevant content", even though both agreed that the content seemed very relevant to them.

The site has been rewritten from scratch a couple of times over the years - no change.

I even paid a copywriter a couple of years back, but the tweaked content still resulted in the same 2-4 quality scores.

I'm considering switching to a different domain, closing the AdWords account and opening a new one, but I've no idea if that would actually change anything.

I would try several different approaches in your content. Don't get hung up on what you think is relevant -- what matters is what google thinks is relevant. Since they don't tell you precisely, try a few different variations -- going for maximum diversity, since you're in a search optimization grid like the board game Battleship.

You need to try things that are as different as possible. Don't rewrite it -- explicitly take different approaches with the goal of making them very different from each other. Then run an experiment.

If I am in Turkey and I google “chicken USA”, I should get shown chicken ads from the USA. And when I click, those chicken producers should pay.

If you hired an ad agency to run radio commercials in Texas and then get a bill saying "oh by the way we found cheap ad space in Tokyo so ran some there too, here's the bill", you'd probably question your relationship with them.

I understand it's not the same thing, but it sure smells the same

The blog author didn't read the docs. He signed up for a relationship and just assumed it would be different and now he's complaining about it. As I mentioned in another comment: From Google AdWords help docs it says "By default, location targeting includes both physical locations and locations of interest".

It seems logical that the chance of you buying their chicken is much lower than someone actually in the US. Why should they pay?

My guess would be that you're making the assumption that I'm looking to buy chicken. It might be the case that someone is just researching "US chicken market"?

And why would most chicken ad publishers want to show ads to those people? I imagine most of those publishers are selling chicken, not information about the US chicken market.

If I don't do business in Turkey, and I pay for the clicks, I should easily be able to exclude Turkey (without having to write a blog post and consult an inside source that most could never reach)

Only if the advertisers choose to have their ads served to you. They don't owe you ads.

There’s a difference between publishing ads for a product that is related to US chicken and publishing ads in the US for chicken.

I believe it’s clear in the Google publisher dashboard that you are targeting your ads to appear to people in the US, not that your product itself is relevant to the topic of the United States.

I think Google has some knob: point of interest vs physical location somewhere inside UI.

Only if the person who was paying for the ads wanted to show them to you in the first place?

I'm really skeptical of the narrative that Alphabet et. al. have brilliant, super-precise targeting of ads. I'm an American who only speaks English, but I'm currently living in a non-English-speaking country. Youtube constantly serves me videos in a language I don't understand. Mostly, I don't even understand what product is trying to be sold to me, and for the minority where the product is clear, I don't understand the arguments, and I'll never buy these things, anyways.

I've never watched a non-English video or ever interacted with the non-English speaking web. How on Earth can Alphabet screw up so much? The only thing I can fathom is that Alphabet is optimizing for getting more advertising revenue, long-term effectiveness of their ads be damned.

Maybe optimizing their ad impressions for folks like you doesn’t justify the engineering costs to implement it.

I’m not sure where you are, but I imagine that folks like you are a very tiny percentage of the market.

I recently created ad campaigns on Twitter/Facebook and Google for our startup. The dark patterns like this were shocking. They really taste of short term metric driven growth that doesn't consider "now I'm paying for clicks that will never convert, I'm going to churn as an account totally".

For example, Adwords defaults into advertising on every device which now includes TVs. This channel would never work for our business (an app that integrates with GitHub), and I'm sure this would work badly for many other businesses. We only found it by digging into the campaign stats once it had launched.

It also strikes me how long the approval takes for new ads - 48 hours. How many advertisers drop off in that time? That feels like a real value add thing to focus on.

I wonder how google handles vpn users, particularly in countries where vpns are fairly ubiquitous like China or even Iran. If someone in Turkey is clearly identified as a Turkish user despite having a vpn in the United States, I shouldn’t be charged for them right?

You might be technically right but Google is incentivized to charge you. Also, it's probably really hard (technically) to resolve your real identity.

It is super simple for Google to resolve real identify of most of users and fraudsters. At least there absolutely no justification for Google not to detect fraud in cases where OP could do it using his own statistics.

Your browser by default leak your OS version (Chinese using older versions very often), locale (of course there is tons of users with Thai or Turkish locale outside of their home country, right?), fonts, etc, etc. Basically everything is can be used and actually actively used for fingerprinting. In cases of some proxies WebRTC can just leak your real IP. I wont even start on the fact that Google controls web browser that is used by 50% internet users and it's sent a lot of data to the mothership.

And yeah if you try to mess with Google itself by let's say mass registering accounts you'll understand how good they are at protecting their own infrastructure and money. At the same time why should they care about someone else lost money? So they don try to fix it.

So glad I'm not involved in PPC anymore. We used to catch junk clicks all the time, from many different countries. One company I worked with was local and wanted very targeted local clicks. Nothing I could do could stop clicks from other states and countries, which were useless to them.

Don't get me started about Twitter advertsing. Want some followers for your business? Twitter will gladly send you tons of obviously fake profiles as soon as you get too many impressions without a click. Like profiles without real pictures and timelines consisting only of retweets.

You have to tolerate an accepted amount of cruft with any PPC campaign.

The cruft is getting worse as Google takes away control from managers and turns it over to AI. It's like how exact match no longer means exact match.


And in addition they no longer show average position metric (1) which increases barrier to entry, they now only show impression share.

(1) https://www.practicalecommerce.com/say-goodbye-to-average-po...

The wording makes targetting useless.

But other questions came up.

Why do so many quit google after less than five years?

How does a company who's product is a push notification when someone clicks get funding?

> Why do so many quit google after less than five years?

Usually initial stock grants are for four years. If the price goes up enough and then relatively flattens out, it can happen that one’s total compensation begins decreasing after the fourth year. That, and 4-5 years is enough time to build a sufficiently large safety net that one can attempt to strike out on one’s own.

Did he get funding?

Or he's just unemployed working on an app?

Being a SWE on Ads gave him an inclination to use Google products?!

Honestly I think this highlights how divorced eng is from real world usage, otherwise he'd know this shit happens. every.damn.day.

ad tech is a house of cards and makes me reconsider working at Google every day.

You work on ads at Google and feel that ad tech is a house of cards?

I actually disagree, primarily because of the auction. It's an incredibly clever way to ensure Google's incentives are aligned with the advertisers'. Assuming rational bidders, Google can only charge more per click if the clicks deliver more value to the advertisers. If Google just sends a bunch of garbage clicks to the advertisers, they will just lower their bids.

OP, how did you track location? If it's IP address then it could be someone in the US using a VPN. If I'm on T-mobile my IP says I'm in Seattle so DFP will show me ads for Seattle but I'm in SF.

So HN wants Google not to track prople but still exactly and completely accurately know the location of each and every one of them for Ad targeting?

That sounds incompatible.

What about just not serving the ad to an IP in Turkey without even looking at the person using it?

How reliable is IP geolocating nowadays?

It's pretty reliable outside some known bad blocks.

Gotta wonder if there’s a point where people just stop paying for this nonsense.

This is completely deceptive, especially showing them as from the US!!! I imagine 90% of AdWords users are not aware of this. Thank you for writing this!

I'm surprised that you're surprised.

You worked for Google. You should have known this before today.

Why exactly should he have known this, especially as an SWE? Google is a gigantic 110k+ people company by now. It's totally possibly to spend years head deep in code specializing on some obscure parts of codebase with near zero understanding how everything else works outside of your narrow area of focus.

He has been working for two years in the AdWords team. Sure it is possible to work somewhere for two years and don't get even an basic understanding of the product, but it is unsurprising if somebody would call that surprising.

Google is a scam company. Stop using it.

Would you please stop posting flamebait and/or unsubstantive comments to HN? We've already had to ask you about this, and we ban accounts that keep doing it.


Is it possible that they've detected someone using a proxy and traced them back to the US?

Not even remotely.

Just watched this vidoe re: other adword shenigans where Tulsi Gabbard discusses her suit against Google with Joe Rogan. They suspended her ads and adwords account during the first democratic debate & never provided a reason.


Applications are open for YC Summer 2021

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact