Chronotope’s avatarChronotope’s Twitter Archive—№ 74,979

                      1. So I've talked about my quest to increase performance on one of my experiment sites.
                    1. …in reply to @Chronotope
                      I had been doing my own caching and playing around with various techniques (as is the point of such a site) when I decided to try CloudFlare
                  1. …in reply to @Chronotope
                    (CloudFlare is a very popular external caching service that my host for that site DreamHost has tools to hook in easily & freely)
                1. …in reply to @Chronotope
                  Google Analytics automatically filters bots out of its reporting. You never see them. CloudFlare does not. So, opportunity to count my bots!
              1. …in reply to @Chronotope
                Keep in mind, this site has almost no canonical content (I use it to archive news stories) and explicitly tells Google not to index it...
            1. …in reply to @Chronotope
              Google Analytics reads yesterday's unique users as 35. CloudFlare: 1,156. So that means about 3% of users hitting my server are humans.
          1. …in reply to @Chronotope
            And keep in mind *it isn't an indexed site*
        1. …in reply to @Chronotope
          Now, CloudFlare isn't catching most of the bots that Google Analytics is catching and dismissing. So Google has some good bot fingerprinting
      1. …in reply to @Chronotope
        Why don't we collaborate on making the web a safer place where servers cost lest and ads are served more accurately?
    1. …in reply to @Chronotope
      If Google cared about veracity of its ad serving stats or health of the open web it would release its methodology for fingerprinting bots.
  1. …in reply to @Chronotope
    Anyway. because 90% of my traffic comes from Facebook and most of the links are shared to Facebook I'm going to posit a theory:
    1. …in reply to @Chronotope
      The decline in referrals from search and rise in Facebook referrals to top slot is accompanied with bots moving crawling operations to FB.
      1. …in reply to @Chronotope
        I'm not saying that it is the causation of the switch, more likely the bots are merely following trends because they want to look human.
        1. …in reply to @Chronotope
          But here's the difference: Facebook operates a closed garden system. They have total control & could eliminate bots on their platform.
          1. …in reply to @Chronotope
            This isn't the only evidence that they have not. Which leads to the obvious question: how much of Facebook's ad inventory is shown to bots?
            1. …in reply to @Chronotope
              Facebook doesn't allow third party tools to verify ad impressions so we will likely never know.
              1. …in reply to @Chronotope
                A fundamental assumption by advertisers that causes Facebook to be seen as trustworthy is the assumption that their traffic is 100% human.
                1. …in reply to @Chronotope
                  If it isn't, does Facebook deserve that trust? Or a slice of the ad marketplace that large? I have an obvious bias, but Qs worth considering
                  1. …in reply to @Chronotope
                    Fun additional reading on this topic: In 2012 a startup claimed that 80% of its Facebook ad clicks came from bots: techcrunch.com/2012/07/30/startup-claims-80-of-its-facebook-ad-clicks-are-coming-from-bots/
                    OpenGraph image for techcrunch.com/2012/07/30/startup-claims-80-of-its-facebook-ad-clicks-are-coming-from-bots/
                    1. …in reply to @Chronotope
                      Distil (runs a product to block bots, so grain of salt) reported in 2015 that 22.8% of all web traffic was bad bots. emarketer.com/Article/China-Bad-Bots-Plague-Mobile-Web-Traffic/1012553
                      OpenGraph image for emarketer.com/Article/China-Bad-Bots-Plague-Mobile-Web-Traffic/1012553
                      1. …in reply to @Chronotope
                        In Q2 2013 Solve Media reported that only 30% of desktop, 57% of mobile was proven as human traffic forbes.com/sites/alexkonrad/2013/09/10/as-websites-see-more-zombie-traffic-the-bots-now-come-from-southeast-asia/#1b5e8e873f4f
                        1. …in reply to @Chronotope
                          And, as noted here, a WSJ article found 46% of display ads were shown to bots: adcontrarian.blogspot.co.il/2013/06/the-75-billion-ad-swindle.html
                          1. …in reply to @Chronotope
                            Worth noting that one of the common ad fraud tactics (pixel-size-compression) is even better supported by new IAB standards for ratio ads.
                            1. …in reply to @Chronotope
                              Bot fraud ad arbitrage may drive up to a 100% profit, so there is plenty of incentive. digiday.com/publishers/confessions-bot-traffic-buyer/
                              1. …in reply to @Chronotope
                                A Moz author crunchs numbers: while more ads could have been seen by humans we can only be absolutely sure of 8%. moz.com/blog/online-advertising-fraud
                                OpenGraph image for moz.com/blog/online-advertising-fraud
                                1. …in reply to @Chronotope
                                  W/this context it is no wonder AdContrarian says ad fraud could be the largest criminal activity in under a decade adcontrarian.blogspot.com/2016/10/agencies-profiting-from-online-ad-fraud.html


Search tweets' text