Recherche avancée

Sur d’autres sites (10404)

• Is Google Analytics Accurate ? 6 Important Caveats

  8 novembre 2022, par Erin

  It’s no secret that accurate website analytics is crucial for growing your online business — and Google Analytics is often the go-to source for insights. 

  But is Google Analytics data accurate ? Can you fully trust the provided numbers ? Here’s a detailed explainer.

  How Accurate is Google Analytics ? A Data-Backed Answer 

  When properly configured, Google Analytics (Universal Analytics and Google Analytics 4) is moderately accurate for global traffic collection. That said : Google Analytics doesn’t accurately report European traffic. 

  According to GDPR provisions, sites using GA products must display a cookie consent banner. This consent is required to collect third-party cookies — a tracking mechanism for identifying users across web properties.

  Google Analytics (GA) cannot process data about the user’s visit if they rejected cookies. In such cases, your analytics reports will be incomplete.

  Cookie rejection refers to visitors declining or blocking cookies from ever being collected by a specific website (or within their browser). It immediately affects the accuracy of all metrics in Google Analytics.

  Google Analytics is not accurate in locations where cookie consent to tracking is legally required. Most consumers don’t like disruptive cookie banners or harbour concerns about their privacy — and chose to reject tracking. 

  This leaves businesses with incomplete data, which, in turn, results in : 

  • Lower traffic counts as you’re not collecting 100% of the visitor data. 
  • Loss of website optimisation capabilities. You can’t make data-backed decisions due to inconsistent reporting

  For the above reasons, many companies now consider cookieless website tracking apps that don’t require consent screen displays. 

  Why is Google Analytics Not Accurate ? 6 Causes and Solutions 

  A high rejection rate of cookie banners is the main reason for inaccurate Google Analytics reporting. In addition, your account settings can also hinder Google Analytics’ accuracy.

  If your analytics data looks wonky, check for these six Google Analytics accuracy problems. 

  You Need to Secure Consent to Cookies Collection 

  To be GDPR-compliant, you must display a cookie consent screen to all European users. Likewise, other jurisdictions and industries require similar measures for user data collection. 

  This is a nuisance for many businesses since cookie rejection undermines their remarketing capabilities. Hence, some try to maximise cookie acceptance rates with dark patterns. For example : hide the option to decline tracking or make the texts too small. 

  

  Sadly, not everyone’s treating users with respect. A joint study by German and American researchers found that only 11% of US websites (from a sample of 5,000+) use GDPR-compliant cookie banners.

  As a result, many users aren’t aware of the background data collection to which they have (or have not) given consent. Another analysis of 200,000 cookies discovered that 70% of third-party marketing cookies transfer user data outside of the EU — a practice in breach of GDPR.

  Naturally, data regulators and activities are after this issue. In April 2022, Google was pressured to introduce a ‘reject all’ cookies button to all of its products (a €150 million compliance fine likely helped with that). Whereas, noyb has lodged over 220 complaints against individual websites with deceptive cookie consent banners.

  The takeaway ? Messing up with the cookie consent mechanism can get you in legal trouble. Don’t use sneaky banners as there are better ways to collect website traffic statistics. 

  Solution : Try Matomo GDPR-Friendly Analytics 

  Fill in the gaps in your traffic analytics with Matomo – a fully GDPR-compliant product that doesn’t rely on third-party cookies for tracking web visitors. Because of how it is designed, the French data protection authority (CNIL) confirmed that Matomo can be used to collect data without tracking consent.

  With Matomo, you can track website users without asking for cookie consent. And when you do, we supply you with a compact, compliant, non-disruptive cookie banner design. 

  Your Google Tag Isn’t Embedded Correctly 

  Google Tag (gtag.js) is a web tracking script that sends data to your Google Analytics, Google Ads and Google Marketing Platform.

  A corrupted gtag.js installation can create two accuracy issues : 

  • Duplicate page tracking 
  • Missing script installation 

  Is there a way to tell if you’re affected ?

  Yes. You may have duplicate scripts installed if you have a very low bounce rate on most website pages (below 15% – 20%). The above can happen if you’re using a WordPress GA plugin and additionally embed gtag.js straight in your website code. 

  A tell-tale sign of a missing script on some pages is low/no traffic stats. Google alerts you about this with a banner : 

  

  Solution : Use Available Troubleshooting Tools 

  Use Google Analytics Debugger extension to analyse pages with low bounce rates. Use the search bar to locate duplicate code-tracking elements. 

  Alternatively, you can use Google Tag Assistant for diagnosing snippet install and troubleshooting issues on individual pages. 

  If the above didn’t work, re-install your analytics script. 

  Machine Learning and Blended Data Are Applied

  Google Analytics 4 (GA4) relies a lot on machine learning and algorithmic predictions.

  By applying Google’s advanced machine learning models, the new Analytics can automatically alert you to significant trends in your data. [...] For example, it calculates churn probability so you can more efficiently invest in retaining customers.

  Vidhya Srinivasan, VP of Measurement, Analytics and Buying Platforms at Google

  On the surface, the above sounds exciting. In practice, Google’s application of predictive algorithms means you’re not seeing actual data. 

  To offer a variation of cookieless tracking, Google algorithms close the gaps in reporting by creating models (i.e., data-backed predictions) instead of reporting on actual user behaviours. Therefore, your GA4 numbers may not be accurate.

  For bigger web properties (think websites with 1+ million users), Google also relies on data sampling — a practice of extrapolating data analytics, based on a data subset, rather than the entire dataset. Once again, this can lead to inconsistencies in reporting with some numbers (e.g., average conversion rates) being inflated or downplayed. 

  Solution : Try an Alternative Website Analytics App 

  Unlike GA4, Matomo reports consist of 100% unsampled data. All the aggregated reporting you see is based on real user data (not guesstimation). 

  Moreover, you can migrate from Universal Analytics (UA) to Matomo without losing access to your historical records. GA4 doesn’t yet have any backward compatibility.

  Spam and Bot Traffic Isn’t Filtered Out 

  Surprise ! 42% of all Internet traffic is generated by bots, of which 27.7% are bad ones.

  Good bots (aka crawlers) do essential web “housekeeping” tasks like indexing web pages. Bad bots distribute malware, spam contact forms, hack user accounts and do other nasty stuff. 

  A lot of such spam bots are designed specifically for web analytics apps. The goal ? Flood your dashboard with bogus data in hopes of getting some return action from your side. 

  Types of Google Analytics Spam :

  • Referral spam. Spambots hijack the referrer, displayed in your GA referral traffic report to indicate a page visit from some random website (which didn’t actually occur). 
  • Event spam. Bots generate fake events with free language entries enticing you to visit their website. 
  • Ghost traffic spam. Malicious parties can also inject fake pageviews, containing URLs that they want you to click. 

  Obviously, such spammy entities distort the real website analytics numbers. 

  Solution : Set Up Bot/Spam Filters 

  Google Analytics 4 has automatic filtering of bot traffic enabled for all tracked Web and App properties. 

  But if you’re using Universal Analytics, you’ll have to manually configure spam filtering. First, create a new view and then set up a custom filter. Program it to exclude :

  • Filter Field : Request URI
  • Filter Pattern : Bot traffic URL

  Once you’ve configured everything, validate the results using Verify this filter feature. Then repeat the process for other fishy URLs, hostnames and IP addresses. 

  You Don’t Filter Internal Traffic 

  Your team(s) spend a lot of time on your website — and their sporadic behaviours can impair your traffic counts and other website metrics.

  To keep your data “employee-free”, exclude traffic from : 

  • Your corporate IPs addresses 
  • Known personal IPs of employees (for remote workers) 

  If you also have a separate stage version of your website, you should also filter out all traffic coming from it. Your developers, contractors and marketing people spend a lot of time fiddling with your website. This can cause a big discrepancy in average time on page and engagement rates. 

  Solution : Set Internal Traffic Filters 

  Google provides instructions for excluding internal traffic from your reports using IPv4/IPv6 address filters. 

  

  Session Timeouts After 30 Minutes 

  After 30 minutes of inactivity, Google Analytics tracking sessions start over. Inactivity means no recorded interaction hits during this time. 

  Session timeouts can be a problem for some websites as users often pin a tab to check it back later. Because of this, you can count the same user twice or more — and this leads to skewed reporting. 

  Solution : Programme Custom Timeout Sessions

  You can codify custom cookie timeout sessions with the following code snippets : 

  • _setSessionCookieTimeout. Set a custom new session cookie timeout in milliseconds.
  • _setVisitorCookieTimeout. Sets a custom Google Analytics visitor cookie expiration time frame in milliseconds.

  Final Thoughts 

  Thanks to its scale and longevity, Google Analytics has some strong sides, but its data accuracy isn’t 100% perfect.

  The inability to capture analytics data from users who don’t consent to cookie tracking and data sampling applied to bigger web properties may be a deal-breaker for your business. 

  If that’s the case, try Matomo — a GDPR-compliant, accurate web analytics solution. Start your 21-day free trial now. No credit card required.

  21 day free trial. No credit card required.

  Your email address
  

  Your website address
  

  Your Analytics subdomain will be .matomo.cloud (change)

  Choose your analytics subdomain
  

  .matomo.cloud

  I hereby accept the terms and conditions and the data processing agreement (DPA)

  
  Your information will be used to create an account on our cloud service. For more information please consult our privacy policy.
  
  

  » Start improving your websites now

  

• ffmpeg doesn't work when the script is launched by cron - No protocol specified, Cannot open display :0.0

  14 janvier 2024, par a k

  ffmpeg in my script doesn't work when the script is launched by cron of the root

  


  Error : No protocol specified, Cannot open display :0.0

  


  OS : Ubuntu 20.04

  


  #!/bin/bash
log=/var/log/log2/log2.txt

echo ______________ $(date) >> "$log"
echo "$""DISPLAY" "=      "  "$DISPLAY" >> "$log" ;
echo whoami '         ' $(whoami) >> "$log" 
echo pwd '            ' $(pwd) >> "$log" 
echo "$""USER" "=" '        ' "$USER" >> "$log"
echo PATH '           ' "$PATH" >> "$log"
echo which ffmpeg '   ' $(which ffmpeg)>> "$log" 
echo whereis ffmpeg ' ' $(whereis ffmpeg) >> "$log" 
echo "\nls -l /bin/* | grep ffmpeg" '        ' >> "$log"
ls -l /usr/bin/* | grep ffmpeg >> "$log"
echo "ls -l /var/log | grep log2" ' '>> "$log"
ls -l /var/log | grep log2 >> "$log"

ffmpeg -y -f x11grab -s 1366x768 -i :0.0 -r 25 /var/log/log2/test.mp4 -loglevel error 2>>"$log" &

echo "pid ffmpeg ""$""!"" = " "$!" >> "$log" 
sleep 5
kill "$!"
echo exit >> "$log"
exit


  


  When Cron (Cron of the root) launches the script, ffmpeg shows the error "No protocol specified, Cannot open display :0.0"

  


  /var/log/log2/log2.txt :

  


  ______________ ven. 05 mai 2023 04:10:01 CEST
$DISPLAY =       
whoami           root
pwd              /root
$USER =          
PATH             /usr/bin:/bin
which ffmpeg     /usr/bin/ffmpeg
whereis ffmpeg   ffmpeg: /usr/bin/ffmpeg /usr/share/ffmpeg /usr/share/man/man1/ffmpeg.1.gz
ls -l /usr/bin/* | grep ffmpeg  
-rwxr-xr-x 1 root   root      284976 mai   18  2022 /usr/bin/ffmpeg
ls -l /var/log | grep log2  
drwxrwxrwx  2 root              root               4096 mai    5 04:09 log2
pid ffmpeg $! =  74590
No protocol specified
[x11grab @ 0x56244aa06740] Cannot open display :0.0, error 1.
:0.0: Input/output error
exit


  


  When I launch the script manually as a not-root user (with sudo), everything work correctly :

  


  ______________ ven. 05 mai 2023 04:10:47 CEST
$DISPLAY =       :0
whoami           root
pwd              /home/an
$USER =          root
PATH             /usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/snap/bin
which ffmpeg     /usr/bin/ffmpeg
whereis ffmpeg   ffmpeg: /usr/bin/ffmpeg /usr/share/ffmpeg /usr/share/man/man1/ffmpeg.1.gz
ls -l /usr/bin/* | grep ffmpeg  
-rwxr-xr-x 1 root   root      284976 mai   18  2022 /usr/bin/ffmpeg
ls -l /var/log | grep log2  
drwxrwxrwx  2 root              root               4096 mai    5 04:09 log2
pid ffmpeg $! =  74618
exit


  


  When I launch the script manually as the root, everything works correctly :

  


  ______________ ven. 05 mai 2023 04:11:27 CEST
$DISPLAY =       :0
whoami           root
pwd              /root
$USER =          root
PATH             /usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/snap/bin
which ffmpeg     /usr/bin/ffmpeg
whereis ffmpeg   ffmpeg: /usr/bin/ffmpeg /usr/share/ffmpeg /usr/share/man/man1/ffmpeg.1.gz
ls -l /usr/bin/* | grep ffmpeg  
-rwxr-xr-x 1 root   root      284976 mai   18  2022 /usr/bin/ffmpeg
ls -l /var/log | grep log2  
drwxrwxrwx  2 root              root               4096 mai    5 04:11 log2
pid ffmpeg $! =  74683
exit


  


  When crontab of an ordinary user launches the script, it works correctly :

  


  $DISPLAY =       
whoami           an
pwd              /home/an
$USER =          
PATH             /usr/bin:/bin
which ffmpeg     /usr/bin/ffmpeg
whereis ffmpeg   ffmpeg: /usr/bin/ffmpeg /usr/share/ffmpeg /usr/share/man/man1/ffmpeg.1.gz
ls -l /usr/bin/* | grep ffmpeg  
-rwxr-xr-x 1 root   root      284976 mai   18  2022 /usr/bin/ffmpeg
ls -l /var/log | grep log2  
drwxrwxrwx  2 root              root               4096 mai    5 05:05 log2
pid ffmpeg $! =  77601
exit


  


  I have tried to put into the script :

  


  export DISPLAY=":0"

  


  export DISPLAY=":0.0"

  


  export DISPLAY=":1"

  


• Could not open file v_YoYo_g07_c04.jpeg av_interleaved_write_frame() : Input/output error

  26 octobre 2017, par Joseph

  the purpose of the following code is to extract frames from a list of videos.
  To do so l defined in python 3.5 extract_frames function which calls extract_frame function. In this latter a ffmpeg instruction is given to extract the frames

  def extract_frames(vidlist,vidDir,outputDir):
  
      f = open(vidlist, 'r')
  
      vids = f.readlines()
  
      f.close()
  
      vids = [video.rstrip() for video in vids]
  
      for vid in vids:
  
          videoName = os.path.join(vidDir,vid.split('.')[0]+".avi")
  
          frameName = os.path.join(outputDir, vid.split('.')[0]+".jpeg")
  
          extract_frame(videoName,frameName)
  
  extract_frames(trainlist01, ucf101_path,training_output)

  such that :

  ucf101_path = "/local/common-data/UCF/UCF-101" # Path to the root dataset
  
  trainlist01="/local/common-data/UCF/ucfTrainTestlist/trainlist01.txt"
  
  training_output = '/local/common-data/UCF/tmp_frames/train/'

  where trainlist01.txt contains a set of lines as follow (first column the path to the video and the second is the class of the video) :

  ApplyEyeMakeup/v_ApplyEyeMakeup_g08_c01.avi 1 .....

  And videoname and framename correspond for instance to :

  videoname="/local/common-data/UCF/UCF-101/YoYo/v_YoYo_g07_c04.avi"
  
  framename="/local/common-data/UCF/tmp_frames/test/YoYo/v_YoYo_g07_c04.jpeg"

  The error is returned by the following function

  import subprocess
  
  
  
      def extract_frame(videoName,frameName):
  
  
  
          if not os.path.exists(videoName):
  
              print('%s does not exist!' % videoName)
  
              return False
  
          # call ffmpeg and grab its stderr output
  
          p = subprocess.call('ffmpeg -i %s -r 1 -s qvga -t 1 -f image2 %s' % (videoName,frameName), shell=True)
  
          return p

  The error is :
  [image2 @ 0x422ff80] Could not open file : /local/common-data/UCF/tmp_frames/train/YoYo/v_YoYo_g25_c05.jpeg
  av_interleaved_write_frame() : Input/output error

  and the output files training_output where l supposed to get my frames is empty.

  files and folders ad subfolders have all right .

  What’s wrong with my code ?

  EDIT 1

  After creating a subfolder YoYo/ in /local/common-data/UCF/tmp_frames/train/ and run again the code l got this new error :

  Could not get frame filename number 2 from pattern '/local/common-data/UCF/tmp_frames/train/YoYo/v_YoYo_g25_c05‌​.jpeg' (either set updatefirst or use a pattern like %03d within the filename pattern) av_interleaved_write_frame(): Invalid argument