What are live analytics for CPaaS?
You might have seen Live Analytics being offered by many CPaaS providers, but its meaning remains fuzzy. In this post, we try to look deeper into live analytics and how various CPaaS providers interpret it, what they provide to users and what are their capabilities.
Live Analytics implies the ability to receive information and data points about the platform, services, users, session quality etc., in real-time.
Put simply, you can think of it as an intelligent traffic camera mounted at an intersection, which maps the current state of signal lights, the number of cars on each signal, pedestrians crossing the road etc. Anyone seeing this camera’s feed knows the instant status of the traffic signal.
In CPaaS, it translates to seeing data points on a continuous timeline including parameters such as:
- Active calls/sessions and detailed information on each
- Number of users in each session with details of their geographical location, devices, bandwidth, browser used etc.
- Quality parameters including audio rates, drops, video freeze rate, video data rates etc.
- Instant data indicating usage by different projects etc.
What’s the utility of live analytics data?
In a nutshell, we can break its utility in the following areas:
- Developer experience
- Customer experience
- Pattern identification
- Fraud detection
When developing a product, especially one that works with complex cloud communication services, developers need real time information. It is important to know what is happening in your business in real time. As a developer, live analytics helps you achieve many objectives: you can monitor coding outcomes, accelerate bug identification and rectification, provide instant feedback among many other actions. Also, the ability to monitor the functioning of applications on distributed systems and geographies in real time facilitates more optimal code development.
For businesses with large number of geographically distributed users, monitoring the communications solutions’ performance is critical to ensure users get an optimal experience. The ability to respond to any issues due to platform failure, application failure or maybe even human failure, gets enhanced if you have access to live information and data. IT Teams understand the value of being able to view data points in real time- it helps them keep a finger on the pulse of their application and empowers them to maintain readiness levels required to resolve any issues if they arise.
Pattern Identification involves mapping continuous data streams of various metrics and parameters to develop patterns on user behavior, device distribution, geographical availability of bandwidth, operating systems preferences and more. Patterns across location, demographics, and subscription channels (SMS, USSD, Web, Android App) help quantify marketing campaign performance, provide critical insights to business leaders, product managers, sales teams etc. It helps them identify emerging markets and trends and specific opportunities for optimization in specific geographic locations or for device types etc. The advantage of live analytics in this scenario compared to post-facto reports is that it allows seeing dynamic movements of data streams that are difficult to model from static data without any significant statistical and modelling expertise, not available in every organization.
Online fraud- whether it is stolen cards, illegal video or voice transmissions, identity theft etc. is becoming a growing problem for businesses globally. The ability to see events in real time, including payments, sudden and unexpected usages in high-risk geographies etc. is key to quick response and rectification, avoiding business losses.
Live Analytics Across CPaaS Providers
By now we know that what is Live Analytics and how it can be very useful for developers and businesses alike, let’s see how various CPaaS providers fare when it comes to giving access to such highly useful data.
Twilio has a comprehensive dashboard for analytics, but it doesn’t provide real time information. Also, the time frame for these dashboards’ availability is limited to 7 days only. There is a good degree of filtering and customization of the dashboard formats.
One can select time periods and see detailed information in visual formats including:
Graphs are available for data pertaining to rooms, participants, usage etc.
These dashboards show details across rooms and participants for distribution of operating systems, devices, browsers region etc., as well as participant level errors and issues like Participant Packet Lost and Participant Round Trip Time.
Quality Metric Graphs
Quality metrics are provided on a per room, per participant basis and are displayed in 10 second intervals. On the send (publishing) side, metrics for bitrate, packet loss, and round-trip time (rat) are provided on a per track basis to assess the quality of the participants outgoing audio or video. On the receive (subscribing) side, the total inbound bitrate for the connection is provided to help you diagnose when a participant has low downlink bandwidth.
Agora offers a comprehensive range of both live analytics and past session level data up to 30-90 days based on the pricing plan you have taken.
Live Session Statistics
You can view the quality of calls of all active sessions through a single dashboard to view essential such as concurrent users and channels, video and audio fluency, streaming quality, freeze ratio, usage and network latency, and many more. Also, the data automatically refreshes every 20 seconds.
You can access to visual representation of a variety of session and user-level statistics to help monitor enterprise scale calls. Intuitive diagrams that help you grasp data in multiple dimensions, such as line charts, heat maps, bubble charts, and ranking lists. The heat map charts offer insights on the Average login time, Audio & Video Freeze Rate, and Network delay rate display data for the last 30 minutes.
* Only video freezing longer than 600 milliseconds is considered. For Web Video SDK, only video freezing longer than 500 milliseconds is considered. Also, only end-to-end network delays more than 400 milliseconds are considered.
Users can also access to various metric such as Geography, Network type, Operating system, SDK version, Device type, and Channel size. It is possible to view these statistics in bar chart form.
Usage & Call Quality Insights
The Quality Overview page gives you Usage Metric like Peak Channel Traffic, maximum number of in-call users across all channels, total number of video and audio minutes. Metric Analysis for quality helps you locate dates and hours to give you an idea about abnormalities faced in different date ranges. You can set data granularity by Day/Hour in the filter.
* The success rate is considered if a user joins the channel within 5 seconds.
Vonage allows you to view historical session level data such as duration and quality of each session and stream.
Advanced Insights for In-Stream Data
Live analytics data is available through Advanced Insights API which allows you to view data at the Session and Stream levels. It gives 30-second snapshots of audio & video latency, bitrate, packet loss ratio, and codecs.
However, data latency is of 5 minutes which isn’t optimum.
Source: Vonage API Developer (See Advanced Insights Section) Insights Dashboard for Historical Data.
You can see what your users are experiencing, around the world and across devices. Basically, it provides the following three information usage (participant minutes, subscriber minutes etc.), video quality, and errors, such as failure rate to connecting sessions.
Insights data is retained for 60 days, and it is aggregated on the daily basis. However, it is available at the latency of 36 – 48 hours.
You can view the histogram of video bitrate and latency for streams within the project that gives you a good idea of the session quality.
EnableX is one of the newer CPaaS providers, with a full stack platform and some cutting- edge live analytics. The width and depth of the real time data that EnableX provides is unmatched among all providers.
While their Live Analytics Dashboards are very comprehensive, we will talk about some key dashboards that are made available, and can give highly valuable insights.
Live Conference Data
This includes a wide number of parameters that includes distributions of active rooms/conferences at the moment, current users, events, media zone utilisation etc. Each conference can be drilled down to provide various other parameters specific to the selected conference including users’ distribution by geography, device type, operating system, browser, errors etc.
Call Quality Dashboards
These live dashboards show visualizations by categorizing each connected user into buckets of call quality, by intelligently assessing their bandwidth usage across all videos, audio or data streams, error occurrence rates, failure/dropout rates etc.
Service wise Statistics
A unique advantage of EnableX is it provides statistics across multiple parameters specific to the video service, audio service, data streams etc., including but not limited to RTT, Send/Received Bitrates, Packet Losses, Errors etc.
Accounts/Participant Level Dashboards
The granularity of the live analytics goes to a user account and even a participant level, providing a comprehensive picture of the real time situation that exists in a video meeting, audio call or others collaborative modes. All the live analytics dashboards in EnableX are not only real time but can also be viewed across various timelines for deeper analysis as required. Only Live Analytics is not enough. Availability of logs/reports is equally important from a customer point of view.
Conclusion- Although the Live Analytics provided by the above CPaaS frontrunners seem similar, functionality-wise they are very different. However, partnering with EnableX can help you gather real-time analytics and feed your respective business requirement in the most seamless manner, ensuring an end-to-end smooth experience.
Disclaimer: The information on this website is for general informational purposes only. All brands mentioned are trademarks of their respective owners, and no representation is implied. All data provided are represented to be accurate at the time of their publishing.