Pinned Questions
      Back to home
      1. Sales Enablement Centre
      2. Conversation Analytics - Sales FAQs
      3. Pinned Questions

      Qualifying Conversation Analytics Requests

      In order to trial conversation analytics, we need to make sure we can access the underlying conversations, in a suitable format.

      Please find out the following answers, and add them to your HubSpot notes before arranging an in-depth with our subject matter experts. 

       

      For text analytics 

      Very little additional qualification is needed in order to demo - the only requirement is that they use a ticket management system/software that we integrate with, and they are comfortable turning on "full-fetch" integrations so that we're able to access conversation text, in addition to the conversations meta-data.

       

      For speech analytics 

      More qualification is needed as the audio needs to be of a certain quality in order to be accurately transcribed. 

      The questions are as follows: 

      1. Who is your current call recording provider?
        1. Please note, that outsourcers may use two systems, so please ensure you collect details of every system within their operation. 
      2. Can you bulk-export calls from your existing call recording platform?
        1. Many platforms will allow you to play and export calls one at a time,
          but this is not sufficient. You need the ability to select a range of calls
          using filters and export in bulk. Please check with your call-recording
          provider as without the appropriate access rights (which some vendors
          scandalously charge for) you will not be able to bulk-export your calls
      3. Are your call recordings stored in split-channel (stereo) audio?
        1. In this case, agent and customer voices are heard separately on the
          left and right channels?
      4. What languages are in scope?
        1. Please check here for the latest languages supported.
      5. What is the compression format of your call recordings? 
        1. We need to know the file format, codec, bit rate and sample rate. You can obtain this information by loading a call recording into a suitable free application such as Audacity.
          The quality of the audio in these calls is an important factor in the success of transcription and analysis. The information here is a guide to the best formats for good quality transcriptions. We understand that many clients do not record with these settings to save disk capacity. We will need to validate that your audio is of sufficiently high quality before commencing the beta trial. The ideal format is uncompressed PCM WAV files at 64 kb/s or higher obtained using audio transmitted over high bandwidth G711 Codec telephony networks. Split channel audio is also
          ideal with one speaker on each channel. Most telephony will have some compression, using the G729 Codec standard normally, so naturally, the audio that
          is recorded (from the inbound caller at least) will be compressed. If compressed files are provided using Codecs such as .MP3 or . WMA are provided, then the ideal set-up would be:
          • Sample rate 8KHz or more (as standard on the G729 Codec)
          • Bit rate 32 kb/s or more
          • Bits per sample (bit depth) 32 but 24 ok if other requirements are met.

            Typical formats we see that are NOT good enough for speech analytics are:
          • MP3 with a bit rate of 16 kb/s or less
          • WMA with a bit rate of 24 kb/s or less