Bots are getting good at mimicking engagement
joindatacops.com
joindatacops.com
The bots are getting creepily good at mimicking engagement. I wrote up my findings, including some of the bizarre patterns I saw and the off-the-record conversations I had with ad tech insiders. It seems like a massive, open secret that nobody wants to talk about because the whole system is propped up by it.
I'm curious if other developers, founders, or marketers here have seen similar discrepancies in their own data.
We at engineering decided to filter out bots. Figures fell dramatically by more than 50%.
In less that a day business mandated us to remove the filter.
Bots are real people after all
There are defensible business reasons for this, in having a contract already in place at the old CPM, so being unable to double the CPM and half the views mid-contract... but still pretty much fraud.
How is that not substantive?
Some countries have property taxes where you declare the value and the government retains the right to purchase the property for that value for example.
My first thought was to make the advertising cost driven by revenue on the site. But that just reverses the incentive.
Altering data would mess with everything. Why is unverified traffic increasing? What's wrong with new marketing efforts? Marketing just requires fixed definitions. e.g. if you have 97% bots but it remains constant that's okay. I know I am spending $x to get $y conversions. I can plan over time, increase or decrease and I can plan. I won't be willing to pay as much as with 0% bots (will pay far far less) but I can set my strategy on this.
It's not that it's x% bots that is the problem. Growth team doesn't adjust strategy on percentage-bot. Growth team adjusts strategy based on return on ad spend. If 0% bots but no return, way worse than 5x ROAS with 99% bots.
In others you'd want, say, auditing or independent third-party verification.
In this case, perhaps an audit involving the deliberate injection of a mix of legitimate and bot traffic to see how much of the bot traffic was accurately detected by the ad platform. Rates on total traffic could be adjusted accordingly.
This of course leads to more complications, including detection of trial interactions, see e.g., the 2010 VW diesel emissions scandal: <https://en.wikipedia.org/wiki/Volkswagen_emissions_scandal>, or current AIs which can successfully identify when they're being tested.
That's a subset of multi-party decisionmaking situations, though it's a useful one to keep in mind.
Basic ad ops has ad buyers buy ads from different vendors, track which converts (attribution, which has flaws, but generally is decent signal), and allocate spend via return on ad spend. So it hurts the vendor at least as much as the buyer by inflating the cost per action / damaging roas.
If YouTube served gigabytes of the video file for 40 munutes and human watched it for that time, but they didn't send a request to `youtube.com/api/stats/atr/` and periodically to `/stats/qoe?`, did the video actually get viewed?
I think a reasonable person would say that the person viewed that video. Only a programmer would suggest that wasn't a view because they didn't properly engage the statistics and tracking endpoints.
But so much of the industry is built on deeply invasive tracking and third-party ad networks that this is a normal thing.
If that causes a problem with Youtubers making a paycheck from external sponsors, Youtube really really does not care, because sponsorships are money that youtube doesn't get!
Youtube is downright hostile to creators who don't make the "right" content, which means new videos at a perfect schedule that all have the exact same "taste" such that viewers can mindlessly consume them in between the actual ad breaks youtube cares about. The more and faster people accept this, the more likely we get improvements.
It took over a year for youtube to agree to let people swear in videos without punishment, including videos that are set as "These are not intended for children", after they unilaterally introduced this swear downranking several years ago.
Youtube cares about Mr Beast and that's about it. If you do not run your channel like Mr Beast, youtube hopes you die painfully and think about your mistake. Youtube actively drives creators making actual art, science, knowledge, and other types of videos to burnout, because Youtube considers creators to be a renewable resource worth exploiting, because there will always be 15 year olds who want to become influencers.
It is not "deeply invasive tracking" or "programmer thinking", it's entirely business. Google's business is ad views, not video views. They want to measure what they care about
And just the fact that if users have a privacy extension blocking the view tracker, is it not just respecting their wishes to not be tracked?
Did something similar at a small company I was working at. The VP of marketing sat me down and told me to do the same thing.
After the meeting, I was told by another dev that the VP was tying a monetary value to specific clicks and if I was filtering out the bots, it would make his data look bad and reduce the amount of potential revenue for the company he was touting.
I think you can see how the bots were actually helping him promote how awesome a job he was doing with our web properties to the owners.
Interestingly, I think money is increasingly its own falsehood now. A lot of rich people are finding that they pay a lot to get what's basically a scam, like Sam Altman's swimming pool that leaked and ruined the rest of the house [1]. There's a reason that Billionaire's Bunker [2] is entering the cultural zeitgeist despite fairly terrible plotting, dialogue, and acting.
[1] https://fortune.com/2024/07/17/sam-altman-infinity-pool-mans...
I worked for one of the mag7 doing customer support bot tech. Clients internal metrics around containment consistently showed better results than ours - even though you'd normally expect them to be putting pressure on their vendor. because it was a kpi for their internal team to look good to their bosses
The correct thing to do, probably, is to just provide the new data to the customer without changing what they were already looking at. So a new widget appears on their dashboard, "52% bot traffic", they click on that, and they see their familiar line chart of "impressions over time" broken down as a stacked line chart, bottom is "human impressions over time," top is "bot impressions over time," and the percentage that they were looking at is reported either above or beneath the graph for the same time intervals. Thus calling attention to the bottom graph, "human impressions over time," and they can ask your sales people "how do I get THAT number on my normal dashboard?" and they can hem and haw about how you have to upgrade to the Extended Analytics Experience tier for us to display that info and other business nonsense...
Point is, you stimulate curiosity with loud interference rather than quietly interfering with the status quo.
It’s not as honest, but more palatable unfortunately
When a company puts up a billboard or an ad on the bus, they don't care if the ad is seen by dashcams and dogs. All that matters is impact on the bottom line.
Remember what matters is how the ad affects the bottom line. Everything else is just a proxy - you need to check to see if your proxies are good enough.
But within their advertising market, you compete for placement with other advertisers. If everyone is getting lots of fraud traffic, presumably they adjust their bids for it, if you're getting outbid consistently, it's reasonable to expect that the other advertisers are either getting a better ROI or they have a lower ROI target than you do.
About a million years ago, I was on a team that had a significant ad program, and it was primarily data driven, we'd come up with keywords to advertise on, measure the results and adjust our bids. With a little bit of manual work to filter out inappropriate keywords or to allow a lower ROI on "important" keywords. Of course, our revenue was largely also from advertising, so it was a bit circular.
Effective advertising depends on iterative testing, which is very hard if the signal to noise ratio is way off.
If you display ads on your website, then sure, it doesn't matter if a bot or real person viewed/clicked on your ad if you get paid regardless - to the point where there's a big industry of putting ads on sites then having bots engage with the site and its ads.
Billboards are a bit different because that's paid per week / month, but internet ads are paid per impression or click.
they also don't get charged retroactively for each dashcam and dog that looks at the billboard
You aren’t paying for conversion rate, you are paying for a link being put on a website when a query is made. You can’t control whether a bot follows that link. You can’t control how sophisticated that bot is. You can’t expect an advertiser to filter out every type of illegitimate traffic (although it sounds like they probably have the capability to filter out more but don’t have any incentive to do so).
I have seen recommendations from across the Internet to not bother with Google ads and other similar paid ad services. It’s basically like paying for a cold lead, you’re attracting one of the least interested types of customers.
The recommendation I’ve always seen is that it’s better to build legitimate interest in your product by producing content. Or perhaps move to an advertising platform where there’s more of a guarantee of reaching human users.
But still, I’ve heard that trying to spend customer acquisition dollars on one-time purchases is a losing battle.
If Tesla was able to start a massive car company without buying ads you can go without AdWords, too.
Google Ads used to be very effective. You are catching the punter as they are actually looking for a solution to their problem. However Google have inflated bids and increased the complexity to the point where very few people can make a decent return now.
>The recommendation I’ve always seen is that it’s better to build legitimate interest in your product by producing content.
That is becoming rapidly less true as AIs steal all the traffic.
>If Tesla was able to start a massive car company without buying ads you can go without AdWords, too.
We can't all be Teslas.
But it's incredibly unlikely that the amount of fraud is stable, any attention from a small player can generate enough volume to eclipse all of the background. Thus no, you can't just account for it on your calculation.
When a company puts up a billboard, it doesn't get taken down because a flock of birds passed by before anyone could look at it.
Billboards sell for flat rates, online advertising is sold on the basis of impressions.
Contrast the billboard with an online ad: If you pay for 10,000 impressions, and 9,000 of those impressions end up being automated traffic, then you’ve effectively lost 90% of your advertising budget. Like you mentioned, ad buyers can and do account for this, so it’s not clear that it’s as big of a problem as this thread is making it out to be. The theory I see thrown around is that Meta and Alphabet are tacitly allowing this bot traffic to continue (or even that they are responsible for it themselves) because it allows them to sell another round of ads, but if this is what is going on it isn’t clear why firms would keep buying ads that aren’t converting.
Uh, yes.
If you get 47 sales on $10k in ad spend (pay per click) and $9900 of that $10k was fraudulent then you got 47 sales on $100 of ad spend. Imagine if you could stop those fraudulent clicks.
Your calculations make as much sense as pouring out 90% of a can of beer, claiming you have just made vodka, and simultaneously trying to pay only for the 10% of beer that is left.
You do understand that when I stop/reduce the fraudulent clicks that leaves extra ad budget available to be spent by clicks from real people who will actually buy things.
I have 0 incentive to not work to stop/reduce the fraudulent clicks. My ad spend stays the same but my sales go up.
Fake clicks give the illusion that ads are working and instead you have to optimize your funnel or whatever else.
I am fine with spending $10,000 on ads (or whatever amount).
The issue is that when I know $5,000 of it was spent on clicks that had 0% chance to convert. For every fraudulent click I can block/prevent that is one more click that can be made by a real person who may actually make a purchase.
The online ad market is extremely inefficient because it has no idea how much ad spend is even reaching people.
You can’t stop fraudulent clicks just like you can’t stop your SuperBowl ad from playing while your viewers are in the bathroom. How much of ESPN’s viewership happens at bars where nobody is watching?
At some point it’s not reasonable to expect ad networks to be able to stop sophisticated bots or exclude them from your billed impressions.
They should definitely try to minimize it if they want to maintain the value of their impressions but I think there is a good argument that OP just isn’t the right customer for this type of advertisement.
If you’re trying to sell a t-shirt you don’t hire a salesperson to cold call people, maybe OP shouldn’t be using web ads in the first place. If fraud was cut down by half would their situation really be that much better?
It does, because it changes the strategy.
If you think the ads are working and have 10k potential customers then you start thinking about how to increase your conversion rate thinking you could get a chunk of those 10k, you might think distribution is solved.
But if it turns out only 2.5k are real humans then your conversion rate might not even be an issue and it’s just the marketing strategy that needs tweaking.
The whole point is that they are giving you fraudulent traffic which you use as real data to figure out the next steps. If you don’t know it’s fraudulent or how much of the clicks are fraudulent then you are taking decisions under the wrong assumptions.
> You can’t stop fraudulent clicks just like you can’t stop your SuperBowl ad from playing while your viewers are in the bathroom
That’s not even a good analogy, we are taking clicks, not impressions.
Yes, you literally can stop fraudulent clicks. I'm not sure where you get the idea that it's not possible?
This is nothing like the Super Bowl analogy.
> During my investigation, a source from the e-commerce data industry provided a crucial piece of the puzzle. He explained that his former company was responsible for scraping 70 million retailer web pages every single day. This is a legitimate and massive source of automated traffic.
> Why do they do this? For vital business intelligence. Major retailers like Amazon do not always notify vendors when they run out of stock. So, brands pay for data scraping services to monitor their own products. These "good bots" check inventory levels, see who is winning the "buy box," ensure product descriptions are correct, and track search result rankings. They even scrape from different locations and mobile device profiles to analyze what banner ads are being shown to different audiences.
I think a lot of the players involved in this would say those are bad bots. Having your competitors scrape your site for data would probably be something most website owners wouldn’t like, but getting data about THEIR competitors would be something they WOULD like.
All bot traffic is good from SOMEONE’S perspective, otherwise it wouldn’t be happening. Someone had to program the bot, someone has to be running it. They obviously think they are good bots.
The people running AI scraping bots think they are good bots. Many content creators think those are bad bots. Price comparison sites scraping retailers think their bots are good. The sites being scraped often think they are bad bots.
I just don’t think we can clearly separate good and bad bot traffic without specifying whose perspective we are talking about.
This case seems "good" by most perspectives.
"Bad bots" would be ones that primarily support grift. They siphon money from advertisers, or get other people to waste advertising $, or game product listing metrics so that the bots' owners products get better placement, or so that other manufacturers' products get worse placement. They're not creating any new value; they're removing value from the economy by causing bad allocation of resources.
For example, bots that scrape content for AI training; is that good or bad for a healthy economy? AI can be productive, but is it ‘stealing’ other’s productivity, which could hurt in the long term if it causes decrease in future human production because of diminished rewards?
It's not even a question of whether, in that particular case, AI will displace some forms of human labor, even if that causes a macroeconomic crisis where society needs to figure out a better way to operate in light of new AI-driven economic realities.
It's a simple question of whether internet bots are causing economic inefficiencies and mis-allocation of resources in the classical sense.
The bad bots, discussed in the article, are. AI scraping bots are not. All it takes to see this are all the mentions of fraud in this HN thread. Bad bots are, at their core, fundamentally fraudulent in how they interact with sites. That fraud is compounded into how companies realizing the extent of that fraud react: by hiding or lying to investors and boards, because the fraudulent bot activity has significantly altered market perception. Nothing about pure scrapers is fraudulent, regardless of whether their ultimate goals are good or bad for the economy or humanity, if that could even be known.
Bot traffic has been an issue for years. It has given rise to a new cottage industry of ad fraud detection services, none of which I’ve found particularly valuable. It always comes down to “so what do we do about it?” and no one seems to know how to get bots to stop viewing or clicking on paid media placements. Consumers use Google search and are on FB, IG, TikTok, LinkedIn, etc., and there aren’t really competing ad networks with “fewer bots”. So they keep buying fake traffic knowing that a significant chunk of it is invalid.
I don’t see anything changing unless big tech companies making billions in ad revenue have a large enough incentive to do so. At the moment they have plenty of incentives to keep things as they are.
“Half of the money I spend on advertising is wasted; the trouble is I don’t know which half.” - John Wanamaker
[1] https://jimalytics.com/industry/youre-buying-bots-an-inconve...
But maybe not legislation, because it might upset a ton of people. Any time a social media site goes through and seems to get rid of bots, people complain about how many "followers" they've lost. The fake user/clicker situation is quite pervasive, and a lot of parties, not just the advertising networks, benefit from the inflated numbers.
But it can be so insidious..."wow, that video got 1M YouTube views? Must be very popular!" Orrrr a lot of those views were from bots? Who knows?
So maybe a better approach than legislation is to talk about how the ad/bot fraud can both help us and hurt us and not demonize one side but see how we all may be implicated in it somehow. Maybe that will help us to be more aware of the problem and not fight against people, but try to work together to solve it.
I was never really a punk-rebel kid, but a certain part of me rooted in the optimism of the early-internet kinda wants to see ad-models crash and burn.
Even advertising "working normally" always had a psychic odor of exploitation and deceit. Ex: "You absolutely need this product or else your peers will hate you."
You could even install them on some digital glasses for a kind of They Live sunglasses situation [0], but please never wear such things while driving...
With that said, what do you think the error is here? Could your script have had some false positives? Enough to move the dial?
Repeated purchase attempts, particularly on the cheapest product in the catalogue is a bad sign.
Strange that the item cost $47 and the number of sales was 47.
The thing was that site kept reflowing the layout over and over again and I think the point was that they were hoping you were going to click on something you saw on the sidebar and then it would reflow and an ad would be there right at the second when you clicked and then... Ka-Ching!
[1] Funny reversal that the enterprise product is mass market and the consumer product is overpriced if not gold-plated.
Only considering old school "Ads by google" banners as ads and then patting yourself on the back for never clicking them is pretty faint praise for your ad evasion skills.
I don't click display ads or links to Amazon or YouTube listings, etc. If I want to buy something I try to go directly to the manufacture's site and search for it there. If I have to go to a third party sales platform I'll go there and search for the thing I want.
I rarely click on YouTube recommedations, because they are more and more just AI slop. I subscribe to people whose content is interesting, and that's the vast majority of what I watch.
In Google Ads you can just turn off the option to run your ads on non-Google sites; I think it's called their Display Network. Just run your campaign only on Google's search pages.
I'm surprised the article doesn't mention this rather common solution.
The numbers that Google Ads shows as invalid clicks is dramatically different from what our platform detects.
In our cases, most fraudulent clicks come from mobile devices and mobile networks. Also, bots don't load page resources in full, making them easier to detect.
If your organization is struggling with fraud click detection, don't hesitate to reach out by email. We are ready to help.
This is not likely to distort the basic numbers in your story, but it makes the premise questionable
If your script can correctly segregate bot traffic from human traffic, and website operators can conclude that bot traffic does not make purchases, then what -- you still don't want to blackhole that traffic right ?
I'm none of those things, 10x or otherwise, and would be ashamed if I were. I do run a personal website, though, and most of the traffic is bots that ignore robots.txt. I wish those developers, founders, and marketers would stop doing that to me.
How does the advertising money get to them to make it worth their while to run bots at such a scale?
I mean sure, I suspect that people in the advertisers are doing it, but surely thats massively risky. there must be a grey market for this kind of transaction?
I agree, I have a hard time understanding the motivation for this behavior. It's obviously happening, I just don't know how someone benefits from a bot pretending to browse the internet.
I was thinking about ad traffic from one of the meta/twitter/tiktok products and couldn't think what the motivation would be
there is also a massive industry of fake accounts and fake engagement for social media and SEO (google). bots are designed to create plausibly real engagement, which is used to trick ranking algorithms into boosting content. these bots have to be real enough to bypass platform detection. clicking through on ads is a way of incentivizing platforms not to shut them down and possibly improving the ranking results, working with the theory that platforms give stronger weight to engagement signals from clients that generate more revenue.
Too often I have been working on ecommerce gigs where the traffic comes from internal 'SEO' tools. Normally the person in charge is marketing, not technical, so it has always been difficult to get past identifying the problem to fixing it.
Often ecommerce companies are very siloed, so one person in one part of the 'team' is scraping the site with their special tools, only for another person to be doing another scrape with their special tools. You can have the guy doing the newsletter doing his thing, the guy doing organic search doing their thing, the guy doing paid ads doing their thing and someone in sales doing their own thing.
The sad thing is that they are typically just a few SQL joins away from exactly the data they want in a format they can digest. However, due to silo-ing, it can be hard to have that conversation.
On top of that, you do get new bots that need to be dealt with. The Huawei bot will scrape everything yet the store might not be delivering to China. So there are legit bots not doing ad fraud that need to be dealt with.
What I also find interesting is that nobody is interested in the server logs. They come for free, and, although not having CDN cache hits, they still record checkout transactions and any pages that don't have a freshly cached page.
Ad fraud also goes on in companies. I worked for a very successful company once and we only measured sales and what was out of stock. We didn't need open rates for emails or click through rates, our main problem was selling too much, which was a nice problem to have. Note that if you sell too much then you aren't going to get it all out the door in a timely fashion, or you run out of big lorries to put the orders in.
Since then I have not worked on a site that is as successful. Instead we have people getting praise for 'false metrics'. Anything an SEO person creates or a marketeer measures will always have some nonsense aspect to it. Or the accounting is mixed with brick and mortar sales even though free shipping and discounts have been given on each sale, with adwords used to get people through the 'door'.
Since sales manager has to report to someone on the board, if the sales numbers aren't good, the nonsense stats can be used to obfuscate the facts. The board only ever care about profit, so I don't like the way this goes down with false metrics of nonsense.
Another fundamental problem in ecommerce is a lack of basic salesmanship. If you work the shop floor doing specialist sales where you have to listen to the customer's needs, then you gain experience in the art of sales. You haven't got to be good at it, in fact it can be better to know your limitations, for me that is big ticket items where I don't have the product knowledge, however, I could always hand those sales over to a much more capable colleague.
You don't win every sale, but, in retail, you can have some really good streaks where no customer leaves empty handed. Your conversion rate is going to be more like 90% in face to face sales if you have the right product at the right price, with customers that don't buy coming back the following day or week to splash the cash.
In High Street retail there is no way you give customers 15%+ off just for stepping through the door. Yet this is table stakes in ecommerce, particularly for small to medium size shops. This instantly devalues the product.
Often there will be chatbots for whatever reason, and I am sure the likes of Dell can get that right, but your typical small ecommerce site will fluff this up too, so any customer daring to use the chatbot will not get instant help from a sales person.
So what to do?
It depends on your product, however, the goal is to get customers for life, not to churn through them. To achieve this it comes down to product, price, availability, shipping times, customer service and incentives for the customer to advertise for you, with reviews, word of mouth and all that hard stuff that needs real human skill. Sometimes it will only be a one-off sale, for example, if someone is buying a mattress. But, even then, the customer service basics matter.
What is also silly is how, with everything you buy, you will get adverts and incentives to buy what you have just purchased. I don't know why this is not considered career limiting, but nobody seems to have fixed this.
Next time, I will find out what data everyone needs and have my own script to collect just that data, then have a live backup that can be used for all internal purposes such as report generation. There will be no mystery script containers downloading 150 scripts with cookies, outside of developer supervision. My hunch is that the only numbers that really matter are sales.
As for why the sales aren't happening, so long as functionality is as it should be on all devices, then you have to dig deeper, to the knowledge level that only someone that has worked the showroom floor understands. Really, a website should be the sales person's knowledge condensed into HTML. Yet nobody asks the guy serving customers what they upsell or what product they recommend if a customer baulks at a given product. Instead we have mystery-meat AI scripts that manage these things.
Thanks for the heads up on ad fraud, normally I am a long way from that due to internal scraping efforts being far more damaging to website performance, and, as a developer, page load times matter to me far more than how much was paid for the traffic.
The word "revenue" above doesn't sound right. Did you mean "ad impressions", "clicks", "visits", or something else?
If the bot traffic is giving you 40% more revenue, well that's excellent news, right? If it's revenue, then keep doing it.
I've nicked that sentence to use as the title above, since it's more neutral and representative of what the article is about.
Why not use Microsoft Clarity?
> The bots are getting creepily good at mimicking engagement.
You would think Google Analytics would help a lot with this, but they seem to not care.
Answer: pride, curiosity, inspiration, creativity, a sense of ownership, integrity, control, vision
I can't believe these concepts are so foreign to so many.
Like, asking if there's a real technical reason.
But, go off I guess.
Not just not care, but don't they have incentive to report higher traffic?
- they don't allow double tracking, so you have to trust their numbers
- if you look at IP from their "clicks", you see often a FB/Google datacenter IP range
- and for most of the traffic they might send you, they did just clever algorithm and heavy profiling to stole your organic traffic. So they get this "amazing" performance by claiming people that would have bought on your site anyway
I have seen and been working in companies trying do to the impact metrics well, but these are outliers
- websites showing ads are annoying their user and get no benefit of it
- stores/brands/people that want to advert pays a bug chunk of money for nothing - only the middle men are getting benefits
Sure, a lot of your visitors were handed a flyer by google before they showed up in your store, but were any of them new customers?
You can either buy the top spot on queries that are looking for you or google will sell it to your competitor.
I used to advertise in Adwords on the name of my products, but I no longer bother.
The reality is that a shockingly high percentage of people just don't get upset about that. Pay attention to anyone using a web browser without an ad blocker. They are so deep in the learned helplessness hole they don't even complain as their web page takes ages to load and reflow while they want to read, and they just silently click skip on all three youtube ads.
The ads work great on them. Think of how many people bought HeadOn.
They don't just take credit for people that were coming to your site anyway, they actively steer them away with competing ads if you aren't paying enough.
Facebook Ads and SA360 both allow 3rd party impression and click tracking. Not to mention the myriad of 3rd party analytics tools you can use to track the website.
> if you look at IP from their "clicks", you see often a FB/Google datacenter IP range
I've never heard this before and seems hilariously simplistic. These megacorps don't have VPNs?
> they did just clever algorithm and heavy profiling to stole your organic traffic
Not sure if you had a stroke while writing this, but this makes no sense.
As someone who has also been in the industry for 10+ years, none of what you said passes the smell test. Maybe after 15 years you still never understood the industry or technology?
>> they did just clever algorithm and heavy profiling to stole your organic traffic
>Not sure if you had a stroke while writing this, but this makes no sense.
I have only ever been a buyer; so my technical understanding of the back end implementation is incomplete.
However, I think the point OP is trying to make is that the back end of the ad targeting infrastructure ends up attributing "spend" to folks who would have otherwise been organic traffic and found your site anyway. ie placing your ad in effective organic pathways and/or in front of well targeted users.
This of course makes sense. A well targeted ad is going to be presented to lots of people who would have otherwise been organic traffic.
This is just a problem with measuring ROI on ad's in general though. I think what has changed is improved attribution of digital ad's has confused people. They see 10 clicks, $5 spend, 1 sale @ $10 and think a 200% ROAS.
In the old days (and still sometimes for non digital) ad effectiveness was measured as a lift over baseline. Different media had different decay rates.
Depending on your digital property, a similar model may need to be applied to your digital campaigns. As I understand though, this is harder to do in modern times with digital ad's.
This is insanely important to do now, and the people who know how to successfully do it are old and retiring.
But nobody really wants to know the number of advertising dollars spent to get customers who were already going to spend on your product to ... spend on your product.
> Not sure if you had a stroke while writing this, but this makes no sense.
Well, that what is happening in practice. It is an attribution game, ad companies try to have the best conversion numbers more than increasing the overall sales of a store.
I have been working in a drive to (physical) store company, we were measuring performances with an expose group + a control group and we were showing 3% visit in store increase during a campaign, while our competitors were claiming things around 10%. We knew that their tracking technology was just about "we shown an ad, they visited the store, it is because of us"
If you want to second guess the psychology of whether or not the ad "drove" the user to make a purchase or not, that is up for discussion. Or whether Google/META do shady shit in their bidding environment. Absolutely.
But this idea that you can't track everything is bullshit.
It just doesn’t ring true. That aspect of the story isn’t novel at all, and someone in that line of work should surely have known all this, right?
Now the section on categorising different bot patterns, that’s more interesting, and I haven’t seen so much said about it.
They had plenty of other problems, including an unworkable business plan, but maybe they would have had more time to pivot before selling out for pennies if they'd not been chasing their tail so much.
I can also say that with every passing year I see that when the Bible talks about people who love lies it's a lot less metaphorical than I thought it was when I was younger. The stories in this very thread about managers demanding that the original, bot-inflated numbers be put back up on their internal metrics rather than have the correct and accurate one is just an example, and one particularly vivid and easy to see.
Though there is also a business opportunity there... I am also reminded of one of my favorite Dilbert cartoons where the boss is viewing a defective spreadsheet that has a higher profit for the year the farther down he scrolls into it. Why not just multiply your click stats by e^(number of months since start / 20) or so? I mean, if the click counts don't have to match any particular reality, go for the gusto.
And what is advertising sold on but statistics, hmmm.
I'm pretty sure that whole market is a game of horse shoes and hand grenades.
For most consumers, it’s entertainment, but for industrial use it’s great for this kind of fraud. And very difficult to detect at scale, since the only cost effective tools for this kind of analysis are also ML/AI, and hence can be fooled more predictably/trained against.
And here's me thinking it was producing a tonne more data scraping.
Looks like a very fresh business [1]; it only became worth writing about once they were ready to pitch their product.
This is a thing that the R&D folks at my company had lots of evidence for, but since we wanted to partner with Meta and Google to be the measurement company of record for FB/Insta and YouTube CPG advertising, we weren't allowed to discuss.
Nothing wrong with using ChatGPT for help of course, I just found it interesting and kind of ironic given the contents of the article. It would be even more interesting if this was the actual writing style of the author, since this must be what ChatGPT was fine tuned to adhere to. Is it the predominant style of communication in adtech?
Tried to keep going, but "a strange, unsettling feeling crept in," yes indeed.
Also please note OP is marketting copy for an analytics product that presumably says it will filter out bot traffic better?
"It's not X, it's Y" is for when you know your reader will think X, and you want them to understand the difference between the X they are expecting and the Y that is actually the case.
Here are the two "not just" passages in the article:
> Driven by this discrepancy, I built a simple tracking script. It was not a sophisticated piece of software, just a tool designed to observe how "users" actually interacted with the website. I was not just counting clicks; I was watching behavior.
> The bots operating today are disturbingly good. They are not just hitting your site and leaving; they are programmed to mimic engagement, making your marketing ROI calculations dangerously inaccurate. I began to categorize them.
Both of these passages are using the "not just X, but Y" trope in a purposeful, meaningful way, anticipating what a typical audience member might assume and signaling for them to adjust their expectation. Not just trying to maximize the usage of the trope with random non-sequitur values of X and Y, as GPT does.
(Oh dear, I just said it too.)
It would be interesting to assemble a corpus of meaningful human text with superficial use of AI tropes, put it side-by-side with true AI slop, and see who can tell the difference. I suspect that, in the coming years, this will be a test posed in higher education and job interviews.
> I was not just counting clicks; I was watching behavior.
This is completely superfluous. The rest of the paragraph is fine, but the only real ChatGPTism is this one.
> They are not just hitting your site and leaving; they are programmed to mimic engagement
In this case it’s a valid sentence. What makes it sound overly AI are the semicolon and the timing, which is exactly the usual “it’s not just {1-5 words}; it’s {1-5 words}”.
Also, perhaps I’m not a typical audience member (I’m not), but I was not assuming any of that. I would try to explain how I parse this kind of language, but the gist is that a 10 sentence word with unambiguous meaning doesn’t really leave space for interpretation. Your attempt to “anticipate what the audience is thinking” is just a writing style, nothing deep about it.
My point mostly being that, as I mentioned in the parent, this is clearly a text written in the ChatGPT style, with simple yet over-expressive language. I personally heavily dislike the style, but I’m a non native speaker used to either scientific or fiction prose. I don’t encounter it too often, and to me it sounds like a parody of Silicon Valley vernacular. Scam alerts immediately start firing in my brain. There’s no “good way” of doing it nowadays.
It's very disappointing that it now seems to be commonly believed that semicolons and em dashes are a sign of AI. Semicolons have been used as a useful tool for a long time; a comma here would simply be wrong.
Those heavily exposed to media before ChatGPT feel there's a shift on the way sentences and narratives are presented. We usually don't use that many semicolons because we structure our sentences not to use them, same for dashes or other weird constructs used by ChatGPT right now. It's hard to explain with words but there are a lot of tell signs when you start looking for them. And not even talking about the explicit ones like bullet points with "bold title:", one that really pulls the curtain open for me is the phrases that end with "one, two, and three of something". When you see it once it's hard to unsee it used everywhere, even on sites that are supposed to be user content only like here or reddit.
Facebook? They have the motivation. They are getting paid per click. But I don't think they would dare. They would get sued into oblivion if they got caught.
A competitor? They could be trying to burn up my marketing budget. But it seems it would require a bit more technical sophistication than most companies have. And are they really going to pay some shady outfit to do it for them? It seems unlikely?
An ad agency? It will quickly become obvious that you are getting crap return on your spend, so that seems unlikely as well.
Someone else? What is their motivation?
These are bot networks run by big firms just building reputation overtime, to be switched on during some huge - usually political - campaign.
Suddenly, in some point in the future, all these "legitimate" looking "people" start inflating discourse against some political figure in some country and completely flip the democratic game.
It happened way too many times in the past decade.
I think the history of corporate bad behavior has demonstrated that any time you say "Surely, company X wouldn't dare to do [some profit-making thing]" you're probably wrong. This applies to all companies, not just Facebook. Revenue = R, Cost of the activity = A, Probability of being caught = P, Likely fine = F. If R > A + P x F, you do the thing.
If stock regulators weren't so worthless they'd make them disclose fraudulent traffic percentages.
While this might sound extreme and paranoid, this is exactly the kind of thing corporations do to transfer risk and create plausible deniability.
Typically, it's people who game the algorithm (for ad dollars), and use these bots to click on other things (including ads) to make it more difficult for FB / Google / etc to detect all the traffic going to their pages and content is their own fake clicks, trying to train the algorithm to send other people to their stuff.
Fraudsters might not be able to as easily game Facebook since they don't pay per 1000 views or whatever - but they can still drive tons of traffic to websites filled with ads, scams, etc, and they can also be making money to get fake engagement to wannabe influencers, politicians, propaganda, etc
Facebook is WAAAY too big of a business to take on this liability themselves, and try to keep it secret.
They'd face securities fraud from their investors, likely huge government fines, and lawsuits from all of their advertisers.
Idly sitting by and letting the numbers pump a little higher is less of a problem.
That said, 73% doesn't come as a surprise. If anything I expect it to be higher.
I guess this quote sums up the situation
> When I tried to bring this up with a few major ad platforms, the conversation always followed a predictable script. The sales reps were incredibly friendly until I mentioned click fraud or bot traffic. Then, the tone shifted instantly to corporate-speak: "Our AI detection is industry leading" and "We take ad fraud very seriously." It was a polite but firm wall, a clear signal to stop asking questions.
> One rep I had known for years finally admitted the truth off the record. "Dude, we know," he said. "Everyone knows. But if we filtered it all out properly, our revenue would drop 40% overnight, and investors would have a meltdown."
Lawsuit time?
If the SEC weren't weak maybe they could go after them for fraudulent numbers.
Wayyyy back in 2000 I ran ad ops for Lycos which included a ton of other sites that they had acquired. We did an audit that uncovered the fact that 25-75% of traffic/pageviews/visitors/ad impressions were due to bots. We did our best at the moment to block some more of them, but it was losing game then, as now.
Advertising, especially online advertising, is a largely a waste of money. Overall, it's obsolete and while it may generate what seems like economic activity, it's a net loss as a use of our time and money.
Advertisement has always been largely a waste of money, the problem is always trying to figure out what part is a waste and what part is effective. Internet advertising promised to be more specifically targeted and attributable than previous advertising models, but I don't think that really turned out well.
As an advertiser, you really have to work hard to know how things are working. Ask customers to tell you how they found you, but know that customers don't always know and don't always share what they know. Adjust advertising campaigns (including turning them off) and see how things change, but changes take a while to filter though. And you've got to do all the other things too --- a lot of new or smaller companies make the mistake of spending a lot on advertisements to drive traffic to them when they can't service the traffic: coupon campaigns that get too many people in, or sending you to a page that's unclear or difficult to use so people bail out, etc.
The article claims 50k visitors, 47 sales, $4000 in ad spend; it's not really clear if that's overall visitor numbers or just from the traffic attributed to advertising. Measuring success by visitor numbers or sales counts was never the right way to measure success; the numeric measure of success for a business should be an accounting measure of net income [1], which is only somewhat related to visitor and sales count. If this business generates $500 of net income for every sale (excluding advertising costs), then $4000 in ad spend to generate 47 sales seems reasonable; if they generate $1 of net income for every sale, then $4000 in ad spend is highly problematic.
[1] GAAP or adjusted as appropriate to the business
After we implemented advanced bot traffic detection and filtering, their reported traffic plummeted by 71%. [...]
But then the sales report came in. Their actual sales went up by 34%.
Their real conversion rate optimization (CRO) efforts had been working all along, but the results were buried under an avalanche of fake clicks. They were not bad at marketing; they were just spending thousands of dollars advertising to robots programmed never to buy anything. Their marketing ROI went from "terrible" to "excellent" overnight.
Can you just tell e.g. Google Ads that you don't want to pay for certain clicks?
Did they modify their targeting to try to avoid bots?
You wouldn't necessarily want to track conversions for retargeting, since depending on your product or service, a second buy might be unlikely. But someone who checks out multiple product pages or articles on your site might be interested and buy in the near future. That of course are also actions bots could easily do.
>Can you just tell e.g. Google Ads that you don't want to pay for certain clicks?
No
If a bot network hits all the conversing events then Google will tailor the traffic to look more like the bot network.
If you filter the bot traffic out then Google can tailor the traffic to look like real converting users instead.
How would you do that on Google or a third-party site?
Bot prevention measures can be good, but the more hoops you make your users jump though (CAPTCHA etc), the more legitimate users will drop off. Those have significant impacts on conversion rates.
I would think fixing this should involve the analytics and attribution side rather than adding friction to your e commerce flow.
Especially as bot tech continues to get better and more indistinguishable from real traffic.
The second is best "blocked" by doing nothing at all and instead silently dropping them from analytics.
I guess there is no incentive in most markets. Facebook, etc make only a token effort to reject non-troublesome bot traffic.
[0] bots and other automated traffic which cannot generate revenue or human ad views
One rep I had known for years finally admitted the truth off the record. "Dude, we know," he said. "Everyone knows. But if we filtered it all out properly, our revenue would drop 40% overnight, and investors would have a meltdown."Yes, if you only count purchase/sale conversions
Maybe no, if you also count clickthrough and view conversions, perhaps even lead conversions sometimes (because fake sign ups are possible).
But you’re right. Purchase conversions are one incentive
I'd love to have an agent that goes online whenever I'm running low on toilet paper or something, browses all the stores and clicks all the ads, and automatically orders the best deal it can find.
Why do they do this? For vital business intelligence. Major retailers like Amazon do not always notify vendors when they run out of stock. So, brands pay for data scraping services to monitor their own products. These "good bots" check inventory levels, see who is winning the "buy box," ensure product descriptions are correct, and track search result rankings. They even scrape from different locations and mobile device profiles to analyze what banner ads are being shown to different audiences.
_---------------_
Guilty as charged. You quickly learn to bypass bot detection measures and create a fully automated system to gather all this information just because amazon doesn't provide it in an accessible manner causing harm to businesses who need this intel and their own internet infra.
https://www.wired.com/story/bots-online-advertising/
And before that, in the 2010s, an employee of ebay discovered the perverse incentives of online advertising. Freakonomics did a podcast on the subject. It is a two parter, but really worth your time if you want to get some of the behind the scenes thoughts and machinations of the advertising industry.
https://freakonomics.com/podcast/does-advertising-actually-w...
I'm working right now on an inventory management system for a clinic which would really benefit from pulling the prices and availability from a very specialised online shop. I wish I could just get a large, fully cached status of all items in a json/CSV/whatever format. But they're extremely not interested, so I'm scraping the html from 50 separate categories instead. They'll get a few daily bot hits and neither of us will be happy about it.
If people are scraping data that you're not selling, they're not going to stop - just make it trivially accessible instead in a way that doesn't waste resources and destroy metrics.
Sure you might be willing to build the bot to scrape it... but some other competitors won't go to this effort so it still means a bit of information asymmetry and stops some of your competitors poaching customers / employing various marketing tactics to exploit short term shortages or pricing charges etc.
> 'Why hand your competitors data on a silver plate'?
To lessen the issue from the article and free up server resources for actual customers.
A simple and reliable API? That's different.
Here are answers to the main questions I'm seeing:
Regarding the writing style: Look, I used AI to edit the article for clarity and professionalism only for the blog. This is a business website, not a personal blog. If you want the raw, unedited version, check this Reddit link:
https://www.reddit.com/r/DataCops/s/vieMBQO8Lg
I find it ironic that some people are more focused on criticizing the writing style than addressing the actual problem I'm documenting.
About sharing the script: Sophisticated bot detection can't rely on Microsoft Clarity or traditional analytics alone. It requires multiple data points:
Modern bots use specific behavioral patterns at scale with customized browser configurations (not just basic headless browsers). You need to understand frameworks like Puppeteer, Playwright, and Selenium very well to detect them.
You need access to live global datacenter IP databases.
Even when click farms use real phones with real browsers, IP intelligence helps distinguish human from automated behavior. Each device needs a unique IP, when it runs its target activity to avoid fraud detection.
Detection involves analyzing dozens of factors simultaneously.
The data in this article is real. I stand by the findings.
https://successfulsoftware.net/2025/08/11/what-i-learned-spe...
It seems that if you're a social media platform, the most profitable option is to ignore bots while swearing up and down that your users are legitimate and you do everything you can to fight bots.
That or dark patters that get people to click on ads in the middle of content, where the user instantly backs out once they realize.
I'm sure that is at least part of it. And if they can be creative about how they get the page to draw, they can get you to click on something unintentionally. The shitty website of my local paper is good at this.
If you are running ads you should be running statistics on them. You should know that you spent $X on ads, and got $Y on sales. This despite how messy the data is (Car manufactures want you to buy a new car every 3 years, so most ads are buy in the future not today - which makes getting answers hard). What matters isn't engagement, clicks, or any other data analysis can give you - they are (or trying to be) a proxy for what matters: how did advertisements affect sales. There are other ways to get that data, those ways have long been known. You shouldn't ignore them - if only to verify that that analysis is still working.
Mainly, it avoids the main point- 73% of your traffic is "faked" enough to look real.
Who are the players in that scenario that stand to benefit from your traffic being fake?
You pay for Google (search ads) and Facebook ads but the traffic is faked by them (unlikely)
You pay other publishing networks (maybe adsense?) and the website owners profit from sending fake traffic (maybe true? if the article were really trying to make a case for this, just name them?)
Or, you work inside a company and just want to make your department look good?
I'm not sure I know what the point of this article is besides a click bait title.
Just tell me exactly what the mechanism is for this fake traffic- don't hint at some kind of conspiracy.
Despite the “LinkedIn influencer” writing style of the article, the results don’t seem that shocking or unexpected.
> just name them Does Adsense even gives you information where exactly your ad is getting published?
Unlikely?
I'd say it's unlikely they are generating fake clicks on search ads.
Mainly because they just don't need to in order to make money. They can still just charge you for impressions on searches that aren't the most well matched. No need to create fake traffic like this blog post conspiracy says.
Eh, you know what, let's just not think too much about it.
Fraudulently creating traffic and charging you for it is a whole other level of scam, though.
A lot of our customers use post purchase surveys and on-site surveys to help with this sort of thing. For example a really common use-case is an attribution survey which appears after a sale is made. The survey will ask something like "how did you hear about us?" which helps determine what actually drove the sale so they can get some clear insights outside of Google and Meta. It's not perfectly reliable but it's an additional data point that helps with the mess out there...
Because I could probably write that, and I'm currently (mostly) on the bench
EDIT: It seems that people actually reject this filtering for business-incentive reasons. Well here's what I propose: In order to get out of this "debt", we "pay it off" at some rate. So for example, every week we subtract an additional 1% of the traffic detected as "fake", in order to hopefully get back to some real numbers at some point
If you feel trapped now, just wait until you're in prison!
If conversion rate drops 90%, the value of a click should also drop 90% as companies adjust the bids for ads.
I'm sure this isn't always true in practice and it takes time for bids to adjust. I'm just wondering how much impact this actually has.
And on the other side, if ad companies were better at filtering bots, their clicks become more valuable and companies should be willing to pay more.
I sincerely wish we could get past this phase of advertising... imagine any other product where you're trying to extract money from 1% by annoying 99% of your audience. We don't have to tolerate this, and many aren't anymore (thanks to ad blockers)
I still have to hear a compelling argument about why I should use computers “by hand” and ignore these powerful tools. Price checking, comparison shopping, buy when released for sale… All of these things point me to using bots.
This feels a lot less like “fraud” and a lot more like “the world has moved on”. Maybe it’s time to route traffic that looks like bots to a bot-optimized shopping experience.
That feels a lot more like 'fraud' than 'the world evolving'
"First-party Web Analytics That Powers CRM: Stop losing leads to blockers, bots, and fake clicks. DataCops gives your CRM the real customer journey."
I wouldn't understand what would be motivation of anyone else to create these bots to click on ad's on facebook. It must be the advertising's company.
The overall point stands, yet the specific "73% fake traffic" figure should be taken with caution until independent data backs it up.
Is he paying for inbound traffic? To whom? Clearly they're not delivering.
I mean, it's burning ad dollars and causing advertisers to rethink their strategy. Who knows, maybe that will eventually lead to the realization that web pages that are 20% content and 80% ads are just luring bots and not customers.
On the other hand, the money being burnt is going to Google, Meta, etc... and helping fund massive surveillance infrastructure. To be honest, I'd prefer it if it all just went to shareholders. Heh, maybe that'll be the sign that we've hit peak surveillance infrastructure: Google and Meta dividend payments go up :-)
But I have trouble sympathizing with someone who writes this:
> Mouse Movements: Did the cursor move in natural, human-like arcs, or did it snap between points?
> Scrolling Patterns: Was the scrolling speed variable, with pauses and upward scrolls, or was it a perfectly smooth, mechanical glide?
> Time Between Interactions: How long did a "user" wait between clicking a link, hovering over an image, or adding an item to the cart?
I read that as: "We're tracking every movement, every hesitation... so that we can feed it to our models and determine how best to keep you addicted".
I knew it was happening, and I know I'm editorializing there... but they are getting closer and closer to just coming out and saying it.
edit: Added newlines in quoted part.
You know, the kind that existed before Google created their thing.
I know of fingerprint.js which is semi open source but I have never tried them out. Any feedback?
Advertising: it's lies all the way down.
Here is recent example:
https://diysolarforum.com/threads/ai-bots-are-here-on-diy-so...
The irony of complaining about bot traffic with bot copy is too funny.
> It was not a sophisticated piece of software, just a tool designed to observe how "users" actually interacted with the website. I was not just counting clicks; I was watching behavior
The dead Internet theory is a conspiracy theory which asserts that since around 2016 the Internet has consisted mainly of bot activity and automatically generated content manipulated by algorithmic curation, as part of a coordinated and intentional effort to control the population and minimize organic human activity.
WTF? I’m pretty sure that’s either impossible or you have access to pretty scary malware.
> This was not the obvious spam that gets filtered out. This was sophisticated bot traffic designed to fool standard analytics platforms.
> Humans are messy; these bots were clinically precise.
AI slop writing is so tiresome. At least it is somewhat noticeable.
PS: For what it's worth, I still think the article was moderately interesting, even if lazily and poorly written by delegating the task to an AI.