Hacker News Reader

Comment by chrismorgan | Hacker News Reader

I’m puzzled by this: I thought it was well-understood, at least in the industry, that traffic numbers were at least mostly nonsense, and that ad click metrics especially were suuuper shady, typically more than half fraud; yet OP, in the business of “accurate ad spend analytics”, only just discovered this!?

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.

View on HN

The last startup I worked at invested a ton of money- much of their marketing team's time plus a good chunk of development time for a few months- trying to "fix the funnel" in terms of conversions from these so-called leads.

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.

Some of us knew it, but it's getting worse. And I've worked adjacent to or on marketing teams for the the last decade... and I can't explain why but my theory is they all trust the numbers implicitly, esp. the ad spend. Having to justify your department's continued existence via KPIs that need to exponentially increase year over year means bloated numbers are a good thing. They can just blame sales for never closing deals.

There's the old story about Charles Babbage and his not understanding how you can feed garbage to a computer and people will believe the output, but in a sense it goes deeper than that; put a number in front of someone and I think a good 75% of the population at least will simply believe it. Not only believe it, but if someone comes along 15 seconds later with a different number, they'll argue with the new guy, even though in a parallel universe where the second guy got there first they'd have argued equally vociferously with the first guy! When objectively they have no real basis for either number, or any other.

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.

The religion of Science is being taught all over the world and the only thing people can remember from that teaching is that Numbers are good, are true, are the only thing we should take into account and are computed by kind smart rational workers.

It's always important to remember: lies, damned lies, and statistics.

And what is advertising sold on but statistics, hmmm.

This is well understood for well over a decade at this point -- the blogpost is a marketing effort by datacops.

Then it is securities fraud. Advertisers' shareholder money is being wasted in a process that is "well understood" to be lies, funneling money to tech industry.

That may be true but good luck ever bringing a case and validating any of it. I am sure those companies that are suspicious are also aware that they are looking at a money pit of fees without any chance of a reliable outcome.

I'm pretty sure that whole market is a game of horse shoes and hand grenades.

50k to 47 conversions seems like a noticeable increase in fakeness to me at least. that's going from more than half fraud to essentially only fraud with a rounding error amount of real users

This is the ‘gold’ that agentic/generative AI really produces.

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.

> This is the ‘gold’ that agentic/generative AI really produces.

And here's me thinking it was producing a tonne more data scraping.

That’s the coal for the steam engine

> Yet OP, in the business of "accurate ad spend analytics", only just discovered this!?

Looks like a very fresh business [1]; it only became worth writing about once they were ready to pitch their product.

[1] https://whois.domaintools.com/joindatacops.com

OP is selling their own web analytics product, so take the article with a grain of salt.

I have no idea who OP is, but as someone who worked in the ad measurement pipeline, it tracks.

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.

It's more a matter of "how much" personally, I knew it was not reliable but I thought it would be around 5-10% and recently more around 20%. I was far from thinking close to 75%!!

Reply on news.ycombinator.com