Recherche avancée

Sur d’autres sites (6826)

• Cohort Analysis 101 : How-To, Examples & Top Tools

  13 novembre 2023, par Erin — Analytics Tips

  Imagine that a farmer is trying to figure out why certain hens are laying large brown eggs and others are laying average-sized white eggs.

  The farmer decides to group the hens into cohorts based on what kind of eggs they lay to make it easier to detect patterns in their day-to-day lives. After careful observation and analysis, she discovered that the hens laying big brown eggs ate more than the roost’s other hens.

  With this cohort analysis, the farmer deduced that a hen’s body weight directly corresponds to egg size. She can now develop a strategy to increase the body weight of her hens to sell more large brown eggs, which are very popular at the weekly farmers’ market.

  Cohort analysis has a myriad of applications in the world of web analytics. Like our farmer, you can use it to better understand user behaviour and reap the benefits of your efforts. This article will discuss the best practices for conducting an effective cohort analysis and compare the top cohort analysis tools for 2024. 

  What is cohort analysis ?

  By definition, cohort analysis refers to a technique where users are grouped based on shared characteristics or behaviours and then examined over a specified period.

  Think of it as a marketing superpower, enabling you to comprehend user behaviours, craft personalised campaigns and allocate resources wisely, ultimately resulting in improved performance and better ROI.

  Why does cohort analysis matter ?

  In web analytics, a cohort is a group of users who share a certain behaviour or characteristic. The goal of cohort analysis is to uncover patterns and compare the performance and behaviour of different cohorts over time.

  An example of a cohort is a group of users who made their first purchase during the holidays. By analysing this cohort, you could learn more about their behaviour and buying patterns. You may discover that this cohort is more likely to buy specific product categories as holiday gifts — you can then tailor future holiday marketing campaigns to include these categories. 

  Types of cohort analysis

  There are a few different types of notable cohorts : 

  1. Time-based cohorts are groups of users categorised by a specific time. The example of the farmer we went over at the beginning of this section is a great example of a time-based cohort.
  2. Acquisition cohorts are users acquired during a specific time frame, event or marketing channel. Analysing these cohorts can help you determine the value of different acquisition methods. 
  3. Behavioural cohorts consist of users who show similar patterns of behaviour. Examples include frequent purchases with your mobile app or digital content engagement. 
  4. Demographic cohorts share common demographic characteristics like age, gender, education level and income. 
  5. Churn cohorts are buyers who have cancelled a subscription/stopped using your service within a specific time frame. Analysing churn cohorts can help you understand why customers leave.
  6. Geographic cohorts are pretty self-explanatory — you can use them to tailor your marketing efforts to specific regions. 
  7. Customer journey cohorts are based on the buyer lifecycle — from acquisition to adoption to retention. 
  8. Product usage cohorts are buyers who use your product/service specifically (think basic users, power users or occasional users). 

  Best practices for conducting a cohort analysis 

  So, you’ve decided you want to understand your user base better but don’t know how to go about it. Perhaps you want to reduce churn and create a more engaging user experience. In this section, we’ll walk you through the dos and don’ts of conducting an effective cohort analysis. Remember that you should tailor your cohort analysis strategy for organisation-specific goals.

  

  1. Preparing for cohort analysis : 

  • • First, define specific goals you want your cohort analysis to achieve. Examples include improving conversion rates or reducing churn.
    • Choosing the right time frame will help you compare short-term vs. long-term data trends. 

  2. Creating effective cohorts : 

  • • Define your segmentation criteria — anything from demographics to location, purchase history or user engagement level. Narrowing in on your specific segments will make your cohort analysis more precise. 
    • It’s important to find a balance between cohort size and similarity. If your cohort is too small and diverse, you won’t be able to find specific behavioural patterns.

  3. Performing cohort analysis :

  • • • Study retention rates across cohorts to identify patterns in user behaviour and engagement over time. Pay special attention to cohorts with high retention or churn rates. 
      • Analysing cohorts can reveal interesting behavioural insights — how do specific cohorts interact with your website ? Do they have certain preferences ? Why ? 

  4. Visualising and interpreting data :

  • • Visualising your findings can be a great way to reveal patterns. Line charts can help you spot trends, while bar charts can help you compare cohorts.
    • Guide your analytics team on how to interpret patterns in cohort data. Watch for sudden drops or spikes and what they could mean. 

  5. Continue improving :

  • • User behaviour is constantly evolving, so be adaptable. Continuous tracking of user behaviour will help keep your strategies up to date. 
    • Encourage iterative analysis optimisation based on your findings. 

  

  The top cohort analysis tools for 2024

  In this section, we’ll go over the best cohort analysis tools for 2024, including their key features, cohort analysis dashboards, cost and pros and cons.

  1. Matomo

  

  Matomo is an open-source, GDPR-compliant web analytics solution that offers cohort analysis as a standard feature in Matomo Cloud and is available as a plugin for Matomo On-Premise. Pairing traditional web analytics with cohort analysis will help you gain even deeper insights into understanding user behaviour over time. 

  You can use the data you get from web analytics to identify patterns in user behaviour and target your marketing strategies to specific cohorts. 

  Key features

  • Matomo offers a cohorts table that lets you compare cohorts side-by-side, and it comes with a time series.
    • All core session and conversion metrics are also available in the Cohorts report.
  • Create custom segments based on demographics, geography, referral sources, acquisition date, device types or user behaviour. 
  • Matomo provides retention analysis so you can track how many users from a specific cohort return to your website and when. 
  • Flexibly analyse your cohorts with custom reports. Customise your reports by combining metrics and dimensions specific to different cohorts. 
  • Create cohorts based on events or interactions with your website. 
  • Intuitive, colour-coded data visualisation, so you can easily spot patterns.

  Pros

  • No setup is needed if you use the JavaScript tracker
  • You can fetch cohort without any limit
  • 100% accurate data, no AI or Machine Learning data filling, and without the use of data sampling

  Cons

  • Matomo On-Premise (self-hosted) is free, but advanced features come with additional charges
  • Servers and technical know-how are required for Matomo On-Premise. Alternatively, for those not ready for self-hosting, Matomo Cloud presents a more accessible option and starts at $19 per month.

  Price : 

  • Matomo Cloud : 21-day free trial, then starts at $19 per month (includes Cohorts).
  • Matomo On-Premise : Free to self-host ; Cohorts plugin : 30-day free trial, then $99 per year.

  2. Mixpanel

  

  Mixpanel is a product analytics tool designed to help teams better understand user behaviour. It is especially well-suited for analysing user behaviour on iOS and Android apps. It offers various cohort analytics features that can be used to identify patterns and engage your users. 

  Key features

  • Create cohorts based on criteria such as sign-up date, first purchase date, referral source, geographic location, device type or another custom event/property. 
  • Compare how different cohorts engage with your app with Mixpanel’s comparative analysis features.
  • Create interactive dashboards, charts and graphs to visualise data.
  • Mixpanel provides retention analysis tools to see how often users return to your product over time. 
  • Send targeted messages and notifications to specific cohorts to encourage user engagement, announce new features, etc. 
  • Track and analyse user behaviours within cohorts — understand how different types of users engage with your product.

  Pros

  • Easily export cohort analysis data for further analysis
  • Combined with Mixpanel reports, cohorts can be a powerful tool for improving your product

  Cons

  • With the free Mixpanel plan, you can’t save cohorts for future use
  • Enterprise-level pricing is expensive
  • Time-consuming cohort creation process

  Price : Free basic version. The growth version starts at £16/month.

  3. Amplitude

  

  Amplitude is another product analytics solution that can help businesses track user interactions across digital platforms. Amplitude offers a standard toolkit for in-depth cohort analysis.

  Key features

  • Create cohorts based on criteria such as sign-up date, first purchase date, referral source, geographic location, device type or another custom event/property. 
  • Conduct behavioural, time-based and retention analyses.
  • Create custom reports with custom data.
  • Segment cohorts further based on additional criteria and compare multiple cohorts side-by-side.

  Pros

  • Highly customisable and flexible
  • Quick and simple setup

  Cons

  • Steep learning curve — requires significant training 
  • Slow loading speed
  • High price point compared to other tools

  Price : Free basic version. Plus version starts at £40/month (billed annually).

  4. Kissmetrics

  

  Kissmetrics is a customer engagement automation platform that offers powerful analytics features. Kissmetrics provides behavioural analytics, segmentation and email campaign automation. 

  Key features

  • Create cohorts based on demographics, user behaviour, referral sources, events and specific time frames.
  • The user path tool provides path visualisation so you can identify common paths users take and spot abandonment points. 
  • Create and optimise conversion funnels.
  • Customise events, user properties, funnels, segments, cohorts and more.

  Pros

  • Powerful data visualisation options
  • Highly customisable

  Cons

  • Difficult to install
  • Not well-suited for small businesses
  • Limited integration with other tools

  Price : Starting at £21/month for 10k events (billed monthly).

  Improve your cohort analysis with Matomo

  When choosing a cohort analysis tool, consider factors such as the tool’s ease of integration with your existing systems, data accuracy, the flexibility it offers in defining cohorts, the comprehensiveness of reporting features, and its scalability to accommodate the growth of your data and analysis needs over time. Moreover, it’s essential to confirm GDPR compliance to uphold rigorous privacy standards. 

  If you’re ready to understand your user’s behaviour, take Matomo for a test drive. Paired with web analytics, this powerful combination can advance your marketing efforts. Start your 21-day free trial today — no credit card required.

• Increased File Size When Converting MP4 to WebM using FFmpeg

  23 décembre 2024, par kimgijeong

  I am using FFmpeg to convert MP4 to WebM with the following command :

  


  ffmpeg -y -hide_banner -nostats \
-f mov,mp4,m4a,3gp,3g2,mj2 -i "http://127.0.0.1:80/lotteon-low-bitrate.mp4" \
-threads auto -f webm -acodec libopus -b:a 96.059k -vcodec libsvtav1 -preset 11 -pix_fmt yuv420p \
-vf "scale='min(-1, iw)':'min(-1,ih)':force_original_aspect_ratio=decrease,crop=trunc(iw/2)*2:trunc(ih/2)*2" \
"/usr/local/m2/m2temp/xcdrtmp/2052_1.webm"


  


  However, the output webm file size is larger than the source MP4 file. For example :

  


  


  • Source MP4 : 4.6 MB (bit rate : 994,053 bps)

    


  


  • Output WebM : 16 MB (bit rate : 3,902,037 bps)

    


  


  


  I know SVT-AV1 encoder defaults to CRF mode. Due to not specifying the bitrate explicitly, the SVT-AV1 encoder automatically sets the bit_rate. It appears that the encoder is setting it to a much higher value (3,323,104 bps), causing the increase in file size compared to the source MP4 (994,053 bps). Here are the methods i tried to reduce the WebM file size compared to the source MP4 :

  


  

  1. -b:v 994k

  


  


  I tried to match the target bitrate with the source MP4's bitrate to reduce the output size, but encountered the following error :

  


  Svt[error]: Instance 1: Force key frames is not supported for VBR mode Last message r
epeated 2 times [libsvtav1 @ 0x239dd100] Error setting encoder parameters: bad parameter (0x80001005)


  


  Looking at the official documentation, this mode change (from CRF to VBR when setting a target bitrate) appears to be expected behavior. However, the error is puzzling since I haven't set any force keyframe parameters in the FFmpeg command.

  


  

  2. svtav1-params "mbr=994k"

  


  


  The second method i tried was using the svtav1-params "mbr=994k" option to set the maxrate while maintaining CRF mode This method showed some improvement, but the output file size was still larger than the source MP4.

  


  

  • Output WebM : 5MB (bit rate : 1,209,877 bps)

  


  


  The more critical reason why we can't adopt the second method (using svtav1-params "mbr=994k") is that even for the same MP4 source file, the output file size varies slightly with each encoding.

  


  

  3. -b:v 994k -svtav1-params “rc=2:pred-struct=1”(CBR, low delay)

  


  


  The final method I tried was setting the target bitrate while using CBR (Constant Bit Rate) and low-delay mode The default value for pred-structure is 2(random access), but I had to use low-delay mode due to the following error :

  


  Svt[error]: CBR Rate control is currently not supported for SVT_AV1_PRED_RANDOM_ACCESS, use VBR mode


  


  This way was the only approach among those i tried that successfully reduced the output size.

  


  

  • Output WebM : 4.3MB (bit rate : 1,032,373 bps)

  


  


  In summary, I have three questions about this MP4 to WebM conversion issue :

  


  


  1. When setting the target bitrate with -b:v 994k, the switch to VBR mode is expected behavior. However, why does the force keyframe error occur when we haven't explicitly set any force keyframe parameters ?

     


  


  2. Why does the output WebM file size fluctuate when setting maxrate either through -maxrate or svtav1-params "mbr=994k", even when using the same MP4 source file ?

     


  


  3. Besides using -b:v 994k -svtav1-params "rc=2:pred-struct=1" (CBR with low delay), are there any other methods that can guarantee a WebM output size smaller than the source MP4 when converting from MP4 to WebM ?

     


  


  


  I am using a recent version of the SVT-AV1 codec :

  


  Svt[info]: SVT [version]:       SVT-AV1 Encoder Lib 58146ca
Svt[info]: SVT [build]  :       GCC 11.5.0 20240719 (Red Hat 11.5.0-2)   64 bit
Svt[info]: LIB Build date: Oct 28 2024 07:40:59
ffmpeg video svt-av1


  


• Ffmpeg RTSP stream with size=N/A bitrate=N/A [closed]

  6 octobre 2024, par Comycos

  When I try to stream a video it seems to work but bitrate and size are N/A frame=  277 fps= 26 q=-1.0 size=N/A time=00:00:11.07 bitrate=N/A speed=1.05x

  


  I run the following command to stream a video to a RTSP server :
ffmpeg -re -stream_loop -1 -i idurre-2024-09-17-12h33m19s.mp4 -c copy -f rtsp -rtsp_transport tcp rtsp://localhost:8554/test

  


  The output ffmpeg is like this :

  


  ffmpeg version 7.1 Copyright (c) 2000-2024 the FFmpeg developers
  built with Apple clang version 15.0.0 (clang-1500.1.0.2.5)
  configuration: --prefix=/opt/homebrew/Cellar/ffmpeg/7.1 --enable-shared --enable-pthreads --enable-version3 --cc=clang --host-cflags= --host-ldflags='-Wl,-ld_classic' --enable-ffplay --enable-gnutls --enable-gpl --enable-libaom --enable-libaribb24 --enable-libbluray --enable-libdav1d --enable-libharfbuzz --enable-libjxl --enable-libmp3lame --enable-libopus --enable-librav1e --enable-librist --enable-librubberband --enable-libsnappy --enable-libsrt --enable-libssh --enable-libsvtav1 --enable-libtesseract --enable-libtheora --enable-libvidstab --enable-libvmaf --enable-libvorbis --enable-libvpx --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxml2 --enable-libxvid --enable-lzma --enable-libfontconfig --enable-libfreetype --enable-frei0r --enable-libass --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg --enable-libspeex --enable-libsoxr --enable-libzmq --enable-libzimg --disable-libjack --disable-indev=jack --enable-videotoolbox --enable-audiotoolbox --enable-neon
  libavutil      59. 39.100 / 59. 39.100
  libavcodec     61. 19.100 / 61. 19.100
  libavformat    61.  7.100 / 61.  7.100
  libavdevice    61.  3.100 / 61.  3.100
  libavfilter    10.  4.100 / 10.  4.100
  libswscale      8.  3.100 /  8.  3.100
  libswresample   5.  3.100 /  5.  3.100
  libpostproc    58.  3.100 / 58.  3.100
[mov,mp4,m4a,3gp,3g2,mj2 @ 0x127704880] st: 0 edit list: 1 Missing key frame while searching for timestamp: 0
[mov,mp4,m4a,3gp,3g2,mj2 @ 0x127704880] st: 0 edit list 1 Cannot find an index entry before timestamp: 0.
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from 'idurre-2024-09-17-12h33m19s.mp4':
  Metadata:
    major_brand     : isom
    minor_version   : 0
    compatible_brands: mp41avc1
    creation_time   : 2024-09-17T10:34:16.000000Z
    encoder         : vlc 3.0.21 stream output
    encoder-eng     : vlc 3.0.21 stream output
  Duration: 00:00:55.68, start: 0.000000, bitrate: 16205 kb/s
  Stream #0:0[0x1](eng): Video: h264 (High) (avc1 / 0x31637661), yuvj420p(pc, bt709, progressive), 1920x1080, 16203 kb/s, 25 fps, 25 tbr, 1000k tbn (default)
      Metadata:
        creation_time   : 2024-09-17T10:34:16.000000Z
        handler_name    : VideoHandler
        vendor_id       : [0][0][0][0]
Stream mapping:
  Stream #0:0 -> #0:0 (copy)
Output #0, rtsp, to 'rtsp://localhost:8554/test':
  Metadata:
    major_brand     : isom
    minor_version   : 0
    compatible_brands: mp41avc1
    encoder         : Lavf61.7.100
  Stream #0:0(eng): Video: h264 (High) (avc1 / 0x31637661), yuvj420p(pc, bt709, progressive), 1920x1080, q=2-31, 16203 kb/s, 25 fps, 25 tbr, 90k tbn (default)
      Metadata:
        creation_time   : 2024-09-17T10:34:16.000000Z
        handler_name    : VideoHandler
        vendor_id       : [0][0][0][0]
Press [q] to stop, [?] for help
frame=  516 fps= 26 q=-1.0 size=N/A time=00:00:20.63 bitrate=N/A speed=1.03x    

[q] command received. Exiting.

[out#0/rtsp @ 0x6000008e0240] video:40672KiB audio:0KiB subtitle:0KiB other streams:0KiB global headers:0KiB muxing overhead: unknown
frame=  528 fps= 26 q=-1.0 Lsize=N/A time=00:00:21.11 bitrate=N/A speed=1.02x    


  


  The server I'm streaming to is a MediaMTX which is supposed to receive the stream. It seems to be connected as it logs it :

  


  mediamtx-rtsp  | 2024/10/06 10:35:13 INF MediaMTX v1.9.0
mediamtx-rtsp  | 2024/10/06 10:35:13 INF configuration loaded from /mediamtx.yml
mediamtx-rtsp  | 2024/10/06 10:35:13 INF [RTSP] listener opened on :8554 (TCP), :8000 (UDP/RTP), :8001 (UDP/RTCP)
mediamtx-rtsp  | 2024/10/06 10:35:13 INF [SRT] listener opened on :8890 (UDP)
mediamtx-rtsp  | 2024/10/06 10:35:16 INF [RTSP] [conn 192.168.65.1:17069] opened
mediamtx-rtsp  | 2024/10/06 10:35:16 INF [RTSP] [session b3bb2623] created by 192.168.65.1:17069
mediamtx-rtsp  | 2024/10/06 10:35:16 INF [RTSP] [session b3bb2623] is publishing to path 'test', 1 track (H264)


  


  But when I try to read it with ffplay nothing is showing...