Hacker News new | past | comments | ask | show | jobs | submit | page 2 login
Google and Facebook ad traffic is 90% useless (youexec.com)
540 points by somid3 on Jan 15, 2017 | hide | past | web | favorite | 370 comments

This is due to the ad fraud bubble. We have covered it extensively on Software Engineering Daily: https://softwareengineeringdaily.com/?s=ad+fraud

Thank you! This makes perfect sense: "A huge percentage of online advertisements are never seen by humans. They are viewed by bots–automated scripts that are opening web pages in a browser and pretending to be a human. Advertising scammers set up web pages, embed advertisements on those pages, and then pay for bot traffic to come and view those advertisements."


The details and numbers which are revealed in this podcast are really eye opening. highly recommended.

we have those, the amount of invalid clicks in adwords does not mach the ones i see in my logs, there 20-40% difference where adwords shows less. i cannot fix that from my site. i block ip ranges when i find them but they can just move.

So the bots are driving billions in trackable revenue for companies then? Branding advertisers aside (which search isn't always great for), direct response advertisers don't often care about some of these metrics between spend and sale other than directionally.

Sure you might have attribution questions, but a bot likely won't spend like a person would.

Ah, thanks for your email!

I recently started writing a book on programming and set-up a landing page for it, to do a pre-launch campaign and I set up some ads on Facebook.

Fortunately, I work in the Analytics industry, so I set-up scroll tracking on my website using Google Tag Manager and started noticing the same patterns as the author as well. Just so to add on to the author, I even excluded the top fraud-prone countries (Egypt, Bangladesh, Pakistan, India, etc.)

In effect, the only pre-orders I've gotten are from Amazon's own internal organic advertising combined with my personal Social media posts. I tried increasing the budget, but that didn't work out either.

Mind you, I track almost every important action on my site (click a button, fill in a form, etc.). The behaviour of these "bots" are very strange and the only thing I've gained in this gamble is an ultra high bounce rate (~90%, < 2 seconds session duration) like the author. Again, I feel it's too early for me to conclude that these networks are useless as I keep reading posts from other people saying Facebook works extraordinarily well for them.

Wouldn't 90%+ of your target market ad block your ad?

We've been doong FB for almost a year now and have learned alot in this space, (with still alot to go).

Here are some things that might help you.

1. The type of ad you run and the copy you use has a big impact on the user's behaviour. Not all ads are created equal. Good copy that really taps i to why someone should be clicking through and what value they can expect to attain the other end makes a big difference. But this isnt an exact science. Its more of an art.

2. If you're targeting is off, you will also get undesired behaviour. People might click through because they like the ad, but if they havent been targeted correctly, it will cause issues. You may need to get more laser focused in terms of targeting. Age range, custom audience (through email upload), lookalike audiences, interests etc.

3. People via email, trust you to some extent and have a relationship with you. They're a warm audience. This is the real reason the organic approach is working for you. You have a relationship with those people. Its not the same with paid ads. You need to build that trust and warm them up. You might take a layered/re-targeting approach by offering them some value and then re-targeting the people that interact with your content with a second and third round of ads. You might also grab their email and then target them both via email and fb.

FB can work, but it take alot of iterations and time to get it right.

Alternatively you could just run an FB ad that does something to collect email addresses, then run an email nurturing campaign, warm up the audience and then target via email.

FB can work, but it takes time.

This is true. What would be amazing is if there was a way to audit their advertising algorithms to know it is fair. I tried their new Contact Form tool, but the cost per lead is about $17/lead.

At the end of the day, if someone is advertising a product that is incredibly good, they should be able to use CPM instead of CPC. The point I am trying to make is that I believe the CPM model is dependent on making CPC & other more costly services look like a better alternative.

Publishers want to get paid for the opportunity (CPM) but Advertisers want to pay a portion of their revenue (PPA): PPL, PPV, PPC and PPM are all proxies for that customer acquisition, so if you want to buy effectively, you need to start by figuring out what your costs per acquisition are. If you can't even hypothetically afford anything close to $17/lead, you might not be able to purchase customers yet, but if you can, then figuring out your lead-to-sales ratio is a lot easier than working out everything between the impression and the sales.

You might consider looking for an agency that specialises in PPL/PPA; basically doing revshare, and letting them build appropriate site lists, and user demographics, and copy testing and so on to help figure out what makes your product sell -- they will put skin in the game because they can leverage their expertise.

This. Not every business has the economics to make advertising work at certain scales.

For many, $17/lead is fantastic because they have an AOV or LTV that supports it. Not every business can do that, so other forms of marketing may make more sense at that stage.

In the contract economy, imagine that all of us, with our small contracting companies and associated specialities, are day labourers standing at the dockside waiting to be hired.

The dock manager (Google) opens a gate to the strong looking fellows allowing them into the dock company's holding area. If they are not in this area (front page of appropriate SERPs) then they aren't considered by the task master who later comes to the holding area, scans the candidates that the dock manager has selected, and takes his pick to get the days work done - from which point it is finally up to the individual and their abilities to prove their ongoing worth.

Imagine if you needed to be an SEO, adwords and general internet marketing guru just to get the dock manager to look at you so you could get paid....

EDIT: Clarity

In the contract economy every one of the dock workers has an agent whose job it is to make sure they have the support they need to provide ongoing worth.

Right now most agents won't take someone who can't immediately bill invoices in the high thousands of dollars. In a contract economy the agent services market would have a low end. Tools will emerge to help agents take on large volumes of low dollar value clients.

Double digit percentages of the entire unrepresented labor market are up for grabs. The current setup is unstable.

Unions are different in that they represent a class not an individual and they don't take a percentage of earnings.

OP is simply doing it wrong. Advertising works, and the title's claim only shows how inexperienced op is. I work at a digital ad agency. Ads work if you do it correctly, but it's incredibly easy to waste money. People like op are why google and Facebook make buying ads overly complicated - if it were easy you wouldn't have to spend so much money.

If op is reading this, I'd be happy to hop on a phone call and help you out.

@Soared, aside from the self-marketing of your firm how about you put your money where your mouth is. Let's use the same landing page - I'll make a campaign, you make a campaign, we each spend $500 and we track conversions in a specific period of time.

The one who converts more for less will owe the other $1000 - we can make the accounts public so that (1) the world can see your "expertise" and (2) how "confused" I am. Game?

I would love to see this.

Every article about rampant ad fraud inevitably contains a comment thread where ad consultants appear to make claims that other people "just can't do it right". Let's see it done right, then.

I agree, @Soared - where you at?

The landing pages (most likely multiple, targeted and tailored differently for different incoming traffic) are a big part of doing things right - for many people in this thread, a big part of what they are doing wrong is where they send the traffic. Many successful ad campaigns would have negative ROI if they simply sent people to the product home page.

Once upon a time there was an affiliate link that paid me $5 per referred sale. I sent them Adwords traffic at 30c/click and a 10% conversion rate ($3/sale = $2 profit for me) until they eventually realised they could just do Adwords themselves.

It doesn't matter if you're only converting 1% of advertising traffic if that 1% is profitable.

This article seems to demonstrate that bots are clicking links though, any comments on that? I mean, if it's true, and those are costing you money, you better be able to design some pretty effective campaigns with that extra overhead.

> People like op are why google and Facebook make buying ads overly complicated

Buying ads is literally one of the easiest things I've ever done on Facebook.

Buying ads is easy. Buying effective ads is not.

Agreed. I'm just objecting to the notion that Facebook makes buying ads hard to protect their advertisers.

Sure, can I can show you my data and make you believe otherwise, I really want to get to the bottom of this since others also believe it to be an issue.

can you email me at somid3 at gmail?

90% useless is still better than most of the alternatives.

Seriously though, it's hard to track success.

If I'm showing an ad wanting people to rush out and buy sneakers... it's rare that it will hit the right person at the right spot in their sneaker purchase decision journey.

But what I want is to bombard them with information so that when they are ready to buy sneakers, they think positively of the sneakers I am selling.

I don't care if they don't follow a perfect path on my site... if they click around and click off... they most likely aren't going to click and buy on the spot. I'm OK with this.

How do I know if my advertising is working? Well... tracking users is one way, but just asking users is something a lot of stores don't bother with. A simple "how did you hear about us?" or hit them up with a customer satisfaction survey after the purchase with a few questions -- incentivized with a coupon. You'll get some great metrics that will help you understand if you marketing budget is being well-spent.

Agreed. You definitely have to do retargeting campaigns too so you can stay in front of them and help drive them to convert when they are in the consideration phase.

However, if you are bidding on keywords showing purchase intent (usually on Google Adwords), the user's should be more engaged. If the Google PPC traffic isn't engaging you may need to optimize your landing page and do some A/B testing.

Facebook is quite different because the user isn't expressing intent at the time they see your ad (usually interest/demographic based targeting) and are having to jump off FB to check your site out. So your landing page has to be really compelling and quicker to digest for that audience. You have to try and capture an email and drop a FB retargeting pixel for sure.

The strategies should be different based on those channels because of the context and frame of mind the user is in.

Well it still tells something that 90% of majority of Googles revenue is based on fabricated BS.

Correct. This is the main point of the claim, and that their CPM is a fabricated stream of visitors with a specific mix of "active" vs. "non-active" users - which they control.

Which you tell Google/Facebook how much you want to pay for it. The better quality they deliver the more you (and your competitors) will be willing to pay.

True. Early on when I used the Google Keyword tool it told me it could drive 400 clicks with a max bid of $0.50 per click. In reality I can get 10 clicks for $1.25 per click -- and that is when the conversion rate is an order of magnitude worst.

For me that means I am paying 100X more than what I had expected.

Keyword tool covers the total market, not what you can expect to get. Quality score increases your cpc, and you'll never ever get 100% of clicks.

I had set my market at United States only. You're correct, about quality score.

The case I am making is I believe Google & Facebook mix "active" and "non-active" users to create an artificial stream of users such that paying by impression is always worst than CPC.

Sorry man, but you're not experienced enough to make these claims. Search, display, and social are entirely different channels and have completely different strategies. Display is for awareness, you shouldn't be trying to get clicks.

You get charged for misclicks so maybe that's what you are seeing. But cpm bidding generally isn't for driving traffic anyways. Cpc is usually more expensive because it is way more targeted and is intent based.

Correct, this is not a statistical trial or anything official. In fact, my data is not significant. After years of using Google and Facebook ads that I have never seen CPM out beat CPC.

Because cpm doesn't get measured with conversions. Cpm is almost strictly awareness which is very difficult to measure.

> But what I want is to bombard them with information so that when they are ready to buy sneakers, they think positively of the sneakers I am selling.

This is horrifying. It's so weird to me that people think this is normal, ethical behavior that it's okay to admit to.

This whole thread is horrifying to me. And I'm saying that as someone who has been online for 35 years, and who (I cringe to admit) worked on the first web-based advertising in 1993. It's like watching a child grow up to be a cruel dictator.

Care to elaborate?

On being horrified, or on early web ads?

For the latter: I worked on Global Network Navigator [1] from 1993-1995. GNN was the first commercial website, published by O'Reilly & Associates. I was the technical director of GNN. We experimented with various types of financial support, including what today would call sponsored content. My first gig at GNN in October 1993 was coding the HTML for the the initial sponsored content article (for a Bay area law firm). Around the same time, Wired (as HotWired) invented banner ads -- I lay no claim to that! ;)

[1] https://en.wikipedia.org/wiki/Global_Network_Navigator

For the former: I think systems like FullStory are incredibly intrusive. Why would I want my personal scrolling & clicking actions captured, played back, and analyzed? How have we come to accept this as normal? I realize most people don't think about this, but that's because they don't realize the level of information that is accrued. I know when I've tried to explain how much information is gathered, my non-techie friends are horrified themselves. Sure, it might be anonymized (or might not); regardless, it's just fucking creepy.

Reading this thread made me double-check my ad-blocker/anti-tracker settings, clear all my cookies, and think again about just disabling Javascript entirely.

This article is 99% useless. Adwords and Facebook are very complex and take experience, time, and $ to get campaigns dialed in right. The targeting and audience choices have become increasingly complex but you can create cost efficient ad placements if you know what your doing and I think in this it's not the case. Plenty of businesses rely upon these platforms to drive traffic, leads, and sales, so your argument for leaving offers no compelling reason why to leave these platforms.

I'm not suggesting we leave these platforms, but rather that we evaluate their true offering, and perhaps audit the algorithms to make sure its fair for everyone. So far, in this long thread it seems others have noticed the same.

Maybe it depends on the product.

I haven't use google ads in years but I did study my logs and watched them in real time and came to the conclusion they were either hitting me with bots or foreigners getting paid to click their ads. It was blatant theft in my opinion.


A fun evening for me is getting a 6six pack and creating a website like "Bob's BBQ", then running google ads for it. The access_log is hilarious. People from Pakistan, people not requesting css (wtf?), on and on.

I have definitely seen the same thing.

I would like to add that having done both, ads in foreign languages are clicked on by bots many times less than ads in English.

Google Adwords seems to be almost a scam. If you don't disable them, most of your clicks will come from Nigeria, Afghanistan, and other random countries, with Google doing nothing about it.

I've had success with Facebook in Polish and Italian, ads in English seemed to be mostly bots.

Pretty sad for people who spend thousands without knowing.

I guess content marketing is the way to go, most people know this. Write good stuff that is useful to your audience, and share it on Reddit. I've done it in the past, with great results.

You're so right. FB & Google doesn't work at all. Stop advertising on those sites, you're making my CPC go up.

Well said, except for the second sentence :)

Agree with this post. Approximately 90% bots in our FB ad network acquisitions. It's a total scam.

There are some niches that are overflowing with stupid advertisers that are bidding higher than the amount the referral is actually worth. In those cases, there really isn't anything you can do. You can wait it out, but there are rubes already lined up to replace the ones that go bankrupt and leave.

That's why you'll find people swearing both cases..."adsense is worthless" vs "adsense is a goldmine". It really depends on the niche.

Was it a Search or a Display Campaign for Google? I could understand that adsense webmasters used clickfarms. However finding out that google is using clickfarms for Search campaigns would be uncovering the greatest scam of the Internet.

I was optimizing for best behavior - so, in the case of Google - Google Search Network only, no network partners. I also moved Mobile devices to -100% so I don't bid in that space. So, desktop & tablet only.

Automating clicks on competitors ads might be beneficial for companies who are competing on the same keywords.

This could be of course just ordinary user behavior. When searching something, I'm opening tens of different pages. Probably I scroll quite quickly through most of them. Just made some quick tests and I think on mobile phone I also act as described, so just flick to the end of the page glancing stuff on my way.

Interestingly the behavior is similar to what I have observed with Mouseflow recordings for organic Google+ and Facebook posts. The page is opened, the mouse never moves and the 'user' scrolls a couple of times. I wonder if these records could just be the regular reviews that Google and Facebook do; their visits do not show up on Google Analytics but Mouseflow is able to take recordings.

Isn't that how legitimate mobile or even laptop trackpad users look ? If I don't intend to click anywhere, the "mouse" cursor never moves - the page loads, I scroll, and go back by keyboard.

With a mouse, you get slight movement even while scrolling or pressing buttons, but with a trackpad or on a mobile device you don't, there's "mouse movement" even if you explicitly do some aiming before a click.

Well, I have 100s of such visitors, I doubt they have 100s of auditors to review my page after 10 seconds each.

A bit off-topic, but how did you create those videos? It seems like a useful tool to test/optimize design, find UI bugs, etc.

EDIT: They used FullStory (https://fullstory.com). I missed that when I read it the first time.

There are quite a few of these screen-capture tools around (Inspectlet, Mouseflow, Hotjar), most of them for UX optimization.

If you're looking for a more developer-focused product, you can also check out LogRocket (https://logrocket.com).

LogRocket looks interesting and may be very helpful for debugging problems in production or optimizing performance. But FullStory or something like it fits my current needs better. I just want a video of real users interacting with the site to see where they get confused. And I don't want to spend a lot of time integrating code to get it to work. That is what FullStory is promising but I haven't tried it yet.

How does this reflect on a sites privacy policy? Are there any open source versions?

You just don't understand how to use the platforms. I'm generating millions of dollars in revenue for startups from both these channels each month.

These posts pop-up every now and then, what OP doesn't consider is they just aren't a good marketer.

People being paid by large advertising companies like google an facebook should declare this in their responses.

I'm making the assumption that the 90% is your bounce rate for traffic from Facebook and Google Advertising.

I'm also assuming that you probably have left all channels (mobile, tablet, desktop, various display channels, etc) on, and not using remarketing/retargeting.

If all that is true, then what you're seeing is normal. Heck it can even be higher than that in a lot of cases. Ideally, you need to be working on reducing that useless traffic over time.

But this easier said than done especially if you have a a small budget that will be completely gone within a few weeks.

PPC is such a hard thing to do because you need consistent funds for at least 1-4 months to truly test what works and what doesn't. If you don't have that, then there is no point in doing it.

My word of advice is either stop doing PPC and focus on less costly things, or if you're going to do it, then start by using only one ad network (doesn't matter; just pick one) and working within a single channel of that network.

Once you feel comfortable within that one channel, then start adding another channel and another channel over time within that ad network.

Hey there, can you tell us a little more about how you were targeting your users and with what kind of content? There could be some bots in there but it also looks like folks land, do a quick check of the content on the site and bounce because it's not what they are looking for. I do this sometimes, especially if I accidentally click on top level ad content on accident. I know you said you did some super targeted ads - what does that mean? It kind of looks like the people converting on your ads are not the ones you're looking for. I'd look at the audience first and then ad content.

It's great to budget your spend where it makes the most sense and your communities and earned ads/content should do great for you. Just watch your spend there and lways budget some for experimentation and try NEW things. Just because you have written off a gold mine doesn't mean there isn't gold there. Maybe you were mining it incorrectly. However, If you feel there is gold all over the ground I'd pick it up before I started digging for it.

As someone who spends a lot of money on ads for aBigCo. globally, I can tell you GDN ads and Twitter prospecting ads are quite terrible.

Facebook has a tremendous amount of data about users, but for a B2B campaign, its quite terrible and worthless. Be sure you completely disable all mobile app "networks" placementd (esp facebook - audience network ) --unless you're doing managed placements -- I've seen tens of thousands burned with these networks, which show ads in worthless apps, designed to generate fraudulent clicks.

That said, retargeting across these networks (esp Facebook) can drive responses, although I can't say we've seen anything turn into a big opportunity later in the funnel (big/complex sales cycles)...retargeting requires traffic, so I'm constantly trying to pull in a variety of 3rd party/behavioural data to improve front end traffic, to get the right people to pages.

The ad model is unsustainable, I want BigCo to switch (they have the resources), but it's not easy for a global company.

I know what you mean, my 9-5 is also at a BigCo, similar story.

Well since the ad pricing is based on bids, it doesn't matter if 90% is useless as long as some flush-with-VC-cash competitor isn't artificially bidding up keywords on that 90% useless traffic.

It's really easy to get ads wrong. The wrong placement, objective, bid amount, bid strategy, targeting, creative, incorrect tracking, poor landing page, etc can all waste your money.

But just from looking at the home page of the site you were advertising I'm not surprised your ads had poor performance. At a glance, I can't tell what exactly i'm 'Starting today' and I doubt your ad creative would be much better, so most people who click are probably curious and have low intent. If you want high CTRs its easy to 'optimize for clicks' and use a catchy ad copy or image, but then obviously you'll have low conversion. Not to mention 'optimizing for clicks' on display networks obviously runs into issues with bots.

Readers that come from your newsletter are already converted to you. Readers that come from an advertisement do not know who you are and are evaluating your website to see if they want to convert to you. It is reasonable to expect that only 10% of these people want to convert to you. The other 90% after scrolling down your website are not interested in converting to you. Your website did not contain the hook to make them convert. You should improve your website.

TL;DR The adverts performed 100% in bringing visitors to your website. That is their job. Your website is not converting them to real users 90% of the time. Your website failed.

So many people are saying "you're doing it wrong" but that doesn't explain what seems like artificial users.

exactly, Veritasium covers this in meticulous detail in his two videos on facebook fraud:



Yup, pretty much. Probably 99% is closer to the truth. If you want to not give your ad budget to crooks you need to buy 100% cost-per-action where you only pay if the click converts. The other end of that spectrum, impressions, is really like throwing bills out your 10th story window into the crowds below.

"Click fraud" is rampant and Yuuuuge. Because it is both easy and it doesn't get the international police on your back like stealing CC's or user accounts do. So it is a fun game for smart people who make anywhere from a few dollars to a few thousand dollars a day just for writing clever code.

I'd love to hear more. I'll email you tmrw. Interesting background.

Why are the two gifs so vastly different? The times are fairly close to each other just 10 seconds difference but the Facebook/Google looks to be altered to give the impression of them doing nothing.

The replay shows them both at 1x but they're not the same at all, the second gif is like 10x the speed of the first.

That behavior looks exactly correct for mobile traffic.

Flagged the story, looks like title click bait to probably hit front page; There seems to be manipulation on the gifs to drive their point.

The second gif animation is actually slower - you are correct. Both gifs were speed up 3X. The Google/Facebook one is a very very slow scroll.

The point I am making is that 80-90% of the traffic is a simple user interaction.

Sorry, but the scroll is so slow that I decided to speed up both animations 3X - that is stated in the article.

See the bottom-right of the GIF "Skip inactivity", which explains the difference you're seeing.

Correct, thank you.

I basically used Camtasia and changed the clip to 3X the speed, I didn't do it via the Fullstory UI

Could it be a third party click fraud? For example, if you buy ads in Adwords on thrid-party sites, their owners' income depends on how many clicks are made so they might be motivated to add bots to real user traffic.

And even if you buy ads on Google's own sites (like Youtube), the channel owners could hire bots to maximize their income.

Regarding Facebook, I don't know if there is any third party, maybe this is just people who clicked on ads accidentaly?

Adwords used to work really well for me, over the years the return declined and the system got more complex until it wasn't worth the effort anymore.

There is a good post from Andy Brice on this: https://successfulsoftware.net/2013/05/26/the-declining-prof...

90% seems accurate:

- Most times I click in ads, I do it only to help the content creators, I have no interest at all in the ad I'm clicking.

- The remaining clicks I often do it by mistake.

I used to click only on ads I didn't like, to charge the creators money, until I realized the networks were optimizing for showing me only ads I hated.

People coming from scroll-through-attend-nirvana kind of platforms like fb, twitter etc are expected to behave that way.

Shouldn't you build a rather very specific landing pages for these click-throughs and then compare their behavior? I personally feel that redirection to very good article on your site would rather have them more interested than sending them to a landing page.

I wonder if it's a touch-screen vs mouse/keyboard difference.

It depends on what you want the new users to do. Ads mostly bring visitors who don't know about your company. If you are in a business where branding or trust is important, it's more difficult. If you are in a commodity business where prices matter, I think it's easier.

That's normal they have a different behavior. Comparing Google Ads and Facebook Ads is hard as there is a user behavior difference between them.

- Visitors from Google Ads often have a purchase intent - Visitors from Facebook Ads are currently entertaining thanks to Facebook photos & videos. (you need to create a need, a urgency to sign up / purchase)

-> my startup has experienced that. Google ads bring qualified users to Moneytis.com (comparison of money transfer options), whereas Facebook ads bring curious people to sign up on neomy.io (fun alerts about currency exchange rates)

Does the auction-style pricing already take this into account?

I don't believe it does. It seems from my data that CPM will always be more costly than CPM. I am suggesting this is artificial.

You set your bid though. What you mean is that the going rate for CPM is higher?

In-house email will always outperform paid media, so the fact that you record a higher degree of site interaction from people who arrive via email should be of no surprise.

With regard to the advertising on Facebook and Google, what audience targeting was tested and applied? What ad formats were used? How many ad concepts were tested? What testing was done on the landing pages to improve performance? What was the "ROI" of the advertising investment, and what was the +/- on these results relative to the goals ascribed to the investment at the outset?

If you have no answers to these questions, or cannot back up your response with data from the type of testing I refer to here (which is part and parcel to running online advertising), then your article is completely pointless.

I recently ran a campaign to validate an idea. Ultimately I dropped the idea, but out of 300 impressions, i got 10 clicks.. and 6 of those clicks left their email address. Not bad.

I then clicked the "optimize my campaign" button. Suddenly my clicks went up, and my impressions went up... but no one left their email. I think the trick to advertising on Google is to target niches, and scale up "manually". I also think there's something to targeting the long tail keywords rather than the fat popular keywords. "Sofa" might get 50k clicks, but the people searching for "Black modern leather sofa" will probably be a better match.

There are many traffic exchange sites that use automated software to click or view ads. Most of them run on server ips or they're using a proxy. http://getIPIntel.net is a free and useful service to combat bad / low quality traffic. Though to save you money, your advertiser needs to not bill you for automated traffic. It's unlikely that'll happen because the more views and clicks, the more money they'll get. You can still use GetIPIntel's api to help you pick which advertising company gives you legitimate traffic.

Isn't this just Sturgeon's law in action?


Google worked better for me since people search with intent whereas on facebook you rely on serendipity. They are there for something else and you hope they will have interest in your service.

All in all are you sure you use the right keywords in google so you get the people with right intent?

Facebook's strength is audience targeting. Are you sure the audience you target is the right one on Facebook?

Start by making a landing page dedicated to your google campaign traffic. Pay attention to the results and try to make changes to improve conversion rate.

Some of those will be changes to the LP, some of it will be narrowing the targeting of your ads to exclude unprofitable segments.

Is easy really, to get something from ads. You just need to stop half-assing it.

One of the visitors to this thread shared this link with me via email: https://softwareengineeringdaily.com/?s=ad+fraud - might help others researching the topic.

"Facebook fraud" is a video anyone who considers paid ads MUST watch: https://www.youtube.com/watch?v=oVfHeWTKjag

Is it common that people move the mouse cursor around while reading a page? What on earth for? Unless it's mindless fidgeting with the physical mouse, wouldn't it slow their reading down?

Most people do not take their hands away from the mouse, and will scroll with the mouse wheel. People are unable to do those things without moving the mouse in some way.

@OP: What's the product? I get that its a newsletter but I can't figure out the content from the copy. Why dont you test putting sample content on the landing page? :)

14 Year PPC guy here. I have built businesses using PPC traffic.

With Google - Turn off search partners and content network and try running the numbers again.

> This visitor is very different - it feels like it has ADD, or its a paid slave boy somewhere, or a bot that has clumsy intelligence.

I'm happy to see you associate someone with ADD with slaves or a non-intelligent bot.

"Each week receive hand-picked insights to help you develop your career & get promoted faster" by learning how to randomly insult people and equate them with slavery.

"BE GREAT AT WHAT YOU DO", like insulting people you've never met.

Do your testimonials come from people who learned how to insult people?

does anyone know if the OP is talking about search ads, or display ads? I stay away from display ads because it's chock full of fraudulent activity, but even search ads I see about 90% lower conversions than organic search. so wondering if this is the same problem....

i dont know if anyone's behavior matches mine, but when i click a fb add, i usually just vertically scroll to the bottom (quickly) looking for anything compelling me to stop.

...Google and Facebook ad traffic useless to 90%

So a 10% conversion rate? Seems normal.

10% of the entering traffic, acts normal. The other 90% is garbage. So if your initial organic conversion rate is X%, this article suggests your Google Ad, or Facebook ad conversion rate will be 10% * X%. That said, its not an extensive study or anything scientific.

This is on you. Build a better audience.

"but yeah, that 10% though"

So Adblock is 90% effective

No, of the entering traffic, after they click on the ad, its as if 90% of them are not your typical organic visitor.

I wonder how snapchat ads are converting

Its pretty covert because so many people are dismissing it as something their daughters use to do frivolous things

But I bet they didn't even know about the election ads, or the short form stuff thats decent at getting the point across quickly, or the location filter ads

@somid3 I would like to invite you to our analytics community: http://www.innerjoins.org

and to @vivekd's request for data on this topic: https://my.caura.co/investigating-fraudulent-clicks-in-googl...

- plenty of data here. Happy to provide a SQL interface for it if you can prove that you've got the chops

love the analysis

I have found that all display ad clicks and all mobile ad clicks are worthless for our business with adwords. You have to blow a thousand bucks to learn that lesson. Once you eliminate that traffic, its not bad. Makes me wonder if google is one ftc lawsuit away from halving in value (tinfoil hat).

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