Contact Us
General Enquiries Tel: 020 8326 8326
Need Support? Tel: 020 8326 8300
red box logo
NICE fully colourised logo
NICE in contact fully colourised logo
Verint fully colourised logo
Microsoft fully colourised logo

Speech Analytics: Deployment options, timescales and proving it works

8 min read
Author Business Systems UK
Date Feb 10, 2015
Category The Inner Circle Guide

ContactBabel recently published the ‘The Inner Circle Guide to Customer Contact Analytics’ which was sponsored by Business Systems. For those with a hectic schedule, we created a short series of blogs covering some of the main points highlighting the latest insights on speech analytics covered in the guide. Please find the fifth in the series below.
The majority of speech analytics implementations tend to be deployed on-premise (CPE) as opposed to in the cloud and according to a recent ContactBabel survey only 8% of customer contact analytics implementations in the US were cloud-based, compared to 22% in the UK.

While most speech analytics solution providers state that the functionality available in the cloud-based solution is similar to that in a CPE environment, there are some issues to consider. Cloud based deployments are likely to offer a reduced implementation cost, be quicker to deploy and do not require additional hardware. Having said that some vendors offer SaaS-based pricing options even for CPE implementations and the transmission of call recordings into the cloud is likely to require a lot of bandwidth.


For those organisations that don’t have the necessary internal analyst or IT resources to implement their own speech analytics technology, access to the outputs of this technology can still be acquired on a managed service basis. The provider will typically have their own dedicated analyst or team that will run analysis of your organisations call recordings and report back on the insights and actions required. The service will also typically check the effectiveness of the analysis, making changes to topics, reports and provide general solution tuning.

This approach works well where a proof of concept is required before taking the plunge with a full investment or where a one-off project is needed to get to the root of a particular issue or where you want to push the onus for results, delivery and commitment of the project back on to the supplier.


Typically lasting between two to eight weeks a proof of concept is where a vendor will come in (or take your recordings offsite) and look at a specific issue or KPI to target, analyse and act on the results. This typically should provide proof of the solutions ability to deliver the ROI and engage the business more closely with the solution and its potential outputs.
Typically at low or no-cost these engagements focus on cost reduction or revenue enhancement for example looking at calls with longest handle time, identifying correlations in this and silence on calls and looking at the costs associated with reducing those calls.


Having spent quite a lot of time and money in getting your call recording system up and running over the years, your choice of speech analytics solution should not necessarily be dictated by your incumbent call recording system it does not necessarily need to be from the same manufacturer.

There are data extraction tools which will export the audio data from the live production environment without endangering it but this comes at an additional cost per seat and should be considered in any study of total cost of ownership. In some cases this can be a struggle and cause some delays, but a solution is usually found in the end.

Another option may be to have the incoming analytics provider implement their own recording solution, which is used only for analysis of calls. Running two recording environments side-by-side negates the need to touch the recording production environment at all.


Speech Analytics roll outs vary, for phonetics-based solutions allow a 1 or 2 week technical implementation followed by a 4-6 week period of fine-tuning the base model.

For more complex, larger implementation that may adopt a hybrid solution (both phonetic and speech-to-text elements) it may take longer with 2 weeks to set up servers and a further 6-12 weeks to carry out the implementation, category and topic creation as well as reporting and dashboards.
Analytics is effectively an ongoing, learning process and where professional services are being provided as part of the package with ongoing training and consultancy, the project is never really over, it simply evolves.

Keep an eye out for the next in our speech analytics blog series – where we’ll be looking how to get the most out of speech analytics post implementation – not one to be missed!

Download the full ‘Inner Circle Guide to Contact Centre Analytics’ here >