5 ms·

Well I tried running some adwords on the 'display network'. I've tried at various times over the years with this and always abandoned it as a festering pit of

by brc 16y ago

Well I tried running some adwords on the 'display network'. I've tried at various times over the years with this and always abandoned it as a festering pit of spam and fraud. I decided to try it out again, in case things had improved somewhat. And before I knew it, I had a couple of conversions. There was also a lot of 50% CTR rates for obvious spam farm sites. So I excluded all of those and upped the budget. And I got a few more conversions. Some really odd ones too - from myspace.com and other sites that I would think my customers have no business being on (or being on and in a purchasing mood, to be more accurate). But then I compared actual sales with analytics and other information. While my adwords account was reporting 'conversions' my sales were actually not showing these conversions. Now I might have the conversion code wrong (doubtful, it's just a copy/paste) but something tells me varios spammers have worked out ways to make some of your ads show conversions to stop being excluded from the list. Whatever the reason, it's switched off again. Traffic and sales show no ill effects. So I'm back to my original hypotheses : the display network is a festering pit of scammers and money wasting.

patio11 16y ago
I was of a similar mind several years ago. Can I suggest that you try Conversion Optimizer? I have my occasional issues with Google, but that product is everything they do right in a little packet of joy with a bow on top. It virtually completely solved my issues with the Content Network. I have end-to-end tracking on my conversions these days and I can verifiably trace ~25% of sales to trials started immediately after a visit from an ad on the Content Network. The network certainly is a hive of scum and villainy, but they give options for not doing much business with that portion of it. P.S. I cannot speak authoritatively for your conversion code, but I'd suspect you're seeing the effects of residual conversions where someone comes in from myspace.com, then later searches via brand name, and you're crediting the brand search but not myspace.
- ericd 16y ago
  Would you mind sharing some of those options you set to avoid the bad parts? Or does the conversion optimizer handle most of that after a while? If you ever wrote a guide to how to get started with google advertising and get it to work for you, I'd gladly pay cash money for it. There's just way too much noise out there on the subject.
- brc 16y ago
  The conversions code is the standard adwords conversion javascript, added to the final 'thanks for your order' page. I would agree with you about the myspace thing- but I ended up with 5 or 10 of these bogus conversions, all on pretty low total clicks (I'm talking 10's of clicks resulting in a conversion). My best referral partners don't go anywhere near those sorts of numbers. In other words, one case makes a fluke (and myspace is highly visited, so I can agree with your theory) but after a couple of rare cases, you start to invoke the law of unlikely statistics in your head and say 'wait a minute...' Here are some stats: domain___cl_impr__CTR__ACPC_Cost_Cv_Co/Cv_CvRate xqno.com 13 1,138 1.14% $0.64 $8.33 2 $4.16 15.38% There's more like this. Some have 1 click, 3 impressions and a 100% conversion rate. No way is that going to happen more than once in a very long time, yet my adwords account has about 5 of these. And none of the conversions shows up in analytics. Anything that was giving me numbers like that- you'd say 'yeah, go for it'. But I could find no actual conversions in analytics for that domain. Looking at that domain we see a Url shortener with one google ad at the bottom. Just knowing my other referral rates and conversion rates for highly-relevant sites, makes me believe firmly that these stats are just plain wrong.
  - Natsu 16y ago
    Assuming your site is small enough that they don't optimize to beat you specifically, the best way to defeat any kind of automatic spamming is to make your site unique in some trivial way. Even if it would take them almost no effort for them to change their tools to spam you, they won't bother if you're too small for them to notice.
    - brc 16y ago
      Not sure what you're saying here. What I'm witnessing is click-fraud. It would seem like nobody is stupid enough to run 1000 clicks on my ad on their site, but they seem to nickel and dime each advertiser - a click here, a click there, nothing that couldn't be disproven. There seems to be a lot of work needed to wade through the display network. I don't know how the false conversions happen. It would seem to me they have found a way.
      - 16y ago

by brc 16y ago

patio11 16y ago

I was of a similar mind several years ago. Can I suggest that you try Conversion Optimizer? I have my occasional issues with Google, but that product is everything they do right in a little packet of joy with a bow on top. It virtually completely solved my issues with the Content Network. I have end-to-end tracking on my conversions these days and I can verifiably trace ~25% of sales to trials started immediately after a visit from an ad on the Content Network. The network certainly is a hive of scum and villainy, but they give options for not doing much business with that portion of it. P.S. I cannot speak authoritatively for your conversion code, but I'd suspect you're seeing the effects of residual conversions where someone comes in from myspace.com, then later searches via brand name, and you're crediting the brand search but not myspace.

ericd 16y ago
Would you mind sharing some of those options you set to avoid the bad parts? Or does the conversion optimizer handle most of that after a while? If you ever wrote a guide to how to get started with google advertising and get it to work for you, I'd gladly pay cash money for it. There's just way too much noise out there on the subject.
brc 16y ago
The conversions code is the standard adwords conversion javascript, added to the final 'thanks for your order' page. I would agree with you about the myspace thing- but I ended up with 5 or 10 of these bogus conversions, all on pretty low total clicks (I'm talking 10's of clicks resulting in a conversion). My best referral partners don't go anywhere near those sorts of numbers. In other words, one case makes a fluke (and myspace is highly visited, so I can agree with your theory) but after a couple of rare cases, you start to invoke the law of unlikely statistics in your head and say 'wait a minute...' Here are some stats: domain___cl_impr__CTR__ACPC_Cost_Cv_Co/Cv_CvRate xqno.com 13 1,138 1.14% $0.64 $8.33 2 $4.16 15.38% There's more like this. Some have 1 click, 3 impressions and a 100% conversion rate. No way is that going to happen more than once in a very long time, yet my adwords account has about 5 of these. And none of the conversions shows up in analytics. Anything that was giving me numbers like that- you'd say 'yeah, go for it'. But I could find no actual conversions in analytics for that domain. Looking at that domain we see a Url shortener with one google ad at the bottom. Just knowing my other referral rates and conversion rates for highly-relevant sites, makes me believe firmly that these stats are just plain wrong.
- Natsu 16y ago
  Assuming your site is small enough that they don't optimize to beat you specifically, the best way to defeat any kind of automatic spamming is to make your site unique in some trivial way. Even if it would take them almost no effort for them to change their tools to spam you, they won't bother if you're too small for them to notice.
  - brc 16y ago
    Not sure what you're saying here. What I'm witnessing is click-fraud. It would seem like nobody is stupid enough to run 1000 clicks on my ad on their site, but they seem to nickel and dime each advertiser - a click here, a click there, nothing that couldn't be disproven. There seems to be a lot of work needed to wade through the display network. I don't know how the false conversions happen. It would seem to me they have found a way.
    - 16y ago