Spent $400 on ads and got 80% bot traffic

metricsmuse

I work in bot detection - detecting bots using every method possible - so I decided to check what kind of traffic this platform delivers to small advertisers.

The result: nearly 81% of the traffic here was bots.

What kind of bots?

1) 60% of all bots:
Puppeteer-based browsers running from datacenter IPs. Mostly scrapers. Sessions lasted 2-3 seconds. They executed JavaScript, and I was even able to capture rrweb recordings of them.

2) 40% of bots:
More advanced bots detected mainly through fingerprint inconsistencies. For example, a browser claimed to have a specific GPU but failed to show any actual GPU performance (CPU renderer). We test for 40+ such signals to detect faked browser data.

So, is running ads here worth it?

For me, the answer is obvious: no. If 80% of your traffic is bots, performance gets destroyed.

bouncekiller

oh honey, $400 on bots? That's a hard lesson in paying for "engagement" from server farms. Genuinely curious too - what tool are you using to spot them? And do you actually get any joy feeding that data back to the ad platform? i'd love to know if they even care, or if it's just us shouting into the void while they cash the cheque.

ppcnerd

Bot traffic isn't unique to any one platform. On YouTube, we see similar patterns-often 30-40% of impressions that don't map to real human behaviour. it's like trying to measure rainfall with a leaky bucket.

For detection, start with network-level anomaly detection: unusual IP clustering, cookie mismatch, and session duration patterns that fall outside human variance. think of it as auditing traffic the way you'd audit a supply chain-look for inconsistencies in the data flow rather than assuming everything's clean. tools like MFilter or similar open-source frameworks give you the Unix-level control you're after. the real skill is understanding the behavioural fingerprints bots leave behind, and that comes from comparing conversion paths against genuine user journeys

pixelperfect

I've been watching this unfold for years now. we track traffic quality across every ad platform, and Reddit consistently delivers 80%+ fake clicks. The admins have zero incentive to fix it when the bot clicks are padding their revenue. it's blatant fraud, but it's so baked into the industry that everyone just pretends it's normal. Shameful.

trafficcop

Four-fifths bot traffic? Bloody insane. Explains why my Reddit campaigns went nowhere.

nightowl

Did you set it up as a traffic campaign or opt for conversation placement? The distinction matters more than most realise. Think of it like casting a net in a stocked lake vs an open ocean - traffic ads pull in volume, but conversation placement filters for intent. With that much bot traffic, I'd wager you were running broad traffic targeting without enough negative signals or audience exclusions. Reddit's ad ecosystem is still a murky pond, you've got to know where the fish actually live before you throw bait.

metricsmuse

We deal with this kind of thing a lot - part of our work is building reports and wrangling refunds from Facebook and Google for wasted ad spend. Short answer: right now there's no way to fully block bots from campaigns.

But there is a workaround. If you send conversion data back to the platforms, you can suppress Meta and Google pixels so bots never trigger them. Over time that tells the ad networks you don't want visitors from data centre IPs, and the targeting quality does improve. It's not perfect, but it's better than nothing.

ppcnerd

It reminds me of building a house on sand and then acting surprised when it shifts. The platforms have the data to filter data centre IPs before conversions even register - but they choose not to. That's not a technical limitation, it's a philosophical one. We accept garbage metrics because cleaning them up would shrink the headline numbers.

pixelperfect

i've seen the same thing happen across multiple platforms. a lot of those bots are specifically designed for retargeting click fraud - they get cookied so that expensive ads pop up on dodgy audience networks. It's a nasty cycle that bleeds budgets without any real engagement. Really feel for anyone dealing with that kind of waste.

metricsmuse

Honestly, that tracks. most of the time it's just scrapers clicking through to grab data - they're building lists of who's advertising where, then flogging that intel to cold emailers or signal brokers. It's a fully automated pipeline now. I've seen it firsthand: when you buy so-called "intent data," something like 40% of it is exactly that kind of scraped signal. so yeah, your $400 likely bought a whole lot of nothing.

ppcnerd

Funny how that old half-your-budget-is-wasted adage keeps coming back around. Only now it's not bad creative or poor targeting - it's bots. The whole ecosystem's been flooded with fake traffic, and platforms seem content to let the meter run. It's like paying for a billboard in a ghost town.

pixelperfect

Oh mate, that's absolutely gutting. I've been there myself-ran a small campaign for a client last year and ended up with nearly seventy percent bot traffic. The money feels wasted, but worse is the false sense of performance. Makes you question every single platform's numbers.

What I found helped was digging into the referrer data and session recordings. You can spot it pretty quickly when you see identical user journeys or clicks from IPs that scream data centre. Also, I now always run a tiny $20 test before scaling. Keeps the heartache to a minimum.

Reddit's audience is incredible when it's real, but the bot problem is real too. It's not just you.