Today we will look at speech analytics as a tool to improve the efficiency of call centres and sales teams.
Speech analytics to assess calls and managers' performance
Telephone conversations, as a way to communicate with the consumer, are used in a huge number of areas. We can call the online shop to clarify terms of delivery or place an order. We contact the technical support of online services by phone. Ordering a pizza. Sign up for a test drive of your vehicle.
There is a need for supervision of call centre operators and sales managers. Moreover, when communicating with customers, they can point out any shortcomings in a product or service, or make suggestions which were not considered before, but could help make the product better. But a call centre operator who receives hundreds of calls a day is unlikely to pay attention to this.
To monitor the performance of managers, and to analyze conversations with customers, companies are either hiring dedicated specialists on staff, or seeking the services of a speech analytics service.
Speech analytics is a tool for evaluating call recordings by tagging (sorting) them by content.
Speech Analytics call center dashboard example
Hiring a person or team to listen to telephone conversations
This method was used in the early days of call centre services. Specially trained people would listen to recordings and identify calls in which:
- operators deviated from prescribed scripts;
- there were excessively long pauses in conversations;
- conflict situations broke out;
- clients expressed their wishes for the improvement of a product or service;
- the call centre script needed improvement.
And attention was drawn to other nuances, problems and suggestions that might arise as a result of communicating over the phone. And while the face-to-face approach is effective, it has two significant disadvantages.
Firstly, the specialists have to listen to a huge number of phone calls and sort them manually, which without additional supervision, which is often lacking, inevitably leads to errors.
Secondly, in most cases they are employees of the company that sells the product, which means additional costs for salaries and workplace equipment.
Using speech analytics services
With advances in technology, the need to hire individual employees is no longer necessary, and today almost all companies are using speech analytics services for contact centres. They provide higher accuracy in call tagging, enable the generation of fast reports, and are much cheaper for companies.
Voice and Speech Analytics call center dashboard example
Automated speech analytics services can often be found. Such programmes use dozens of dictionaries compiled in advance by linguists to monitor call quality, with which the machine identifies problematic conversations, analyses script compliance, monitors politeness of communication and categorises calls accordingly.
Artificial intelligence is also added to automation, where neural networks monitor not only the presence of specific phrases, but also analyse the speed of the conversation, the presence of pauses and interruptions and deal with transcribing conversations - translating audio into text and dividing the dialogue into customer and manager's speech.
But even modern technology is not perfect. This is due to the fact that in a conversation the client can use abbreviations, slang, other words and phrases, by which only a person can understand what the client really meant. Yes, they can be added to the dictionary. But to do this, you need to first hear them and identify them as necessary. Adding and reconfiguring the system itself also takes time. Equally important is the emotional colouring of the conversation, which as yet the artificial intelligence is unable to interpret correctly.
And so some services are once again returning to engaging specialists to listen to calls, but already with the use of online analytics and additional control.
Such specialists are called taggers (from the word tag). Their work is structured as follows:
- Together with the customer the service team determines from 2 to 15 tags, which will be used to mark the calls;
- taggers listen to the calls and assign tags to them, for example: product: grill, told about the promotion, offered delivery, manager did not introduce himself;
- further this data is processed by the online analytics system and generated into reports for further work with them;
- These manually assigned tags are also used to train neural networks for automated call classification.
Unlike in-house wiretapping, the work of the taggers is optimised:
- Special interfaces have been created for the taggers to listen to calls and tag them;
- There are acceleration and pause cutting options, which reduces call processing time;
- Automated unloading of calls from PBX, CRM and call-tracking systems;
- Employees work on a part-time basis, minimising the potential for errors that can arise with increased fatigue from full-time professionals;
- Taggers are recruited from the regions, which reduces the cost of handling calls without losing quality;
- The work of the taggers is additionally monitored.
Together we can achieve 99% accuracy in call handling with much lower financial costs.
Possible Speech Analytics call center infrastructure setting
What do you use call analysis for?
As we stated above, voice analytics programs provide call quality control in call centres and sales departments. But voice analytics software does not stop here. With their help you can:
- Analyse the effectiveness of advertising channels, optimise advertising costs and reduce the cost of targeted leads. This is especially relevant when working with contractors because it often happens that unscrupulous advertising agencies engage in lead scam in highly competitive areas. Calls are received, but are not converted into deals. Classification of leads into targeted, non-targeted and SPAM calls together with automatic uploading of this data into web analytics and contextual advertising systems allows you to obtain more accurate data on the effectiveness of advertising channels and optimize them, getting more targeted leads and sales within the same budget;
- Improve the performance of each call centre manager and sales team individually. Managers make mistakes. Often trivial, easily corrected by control and substantive feedback. For example: not greeting, interrupting, not addressing by name, not communicating benefits, not communicating discounts/special offers/bonuses, not offering related products and additional services. With tags managers see what kind of mistakes made by a particular manager and can give him feedback not in the format of "Ivan, you bad manager, not enough sell, sell more," but in the format of "Ivan, when selling printers, you offer to buy an extra set of cartridges to only 30% of printer buyers. And the average figure for the department is 75%. Do it more often, it will increase your revenue and company sales."
- Increase sales by redirecting calls to more effective managers. Managers are not the same and sell different products with different efficiencies. And call distribution in 9 out of 10 companies is built without regard to effectiveness (evenly or on the principle of who picks up the phone first). With the speech analytics service you can count the number of calls each manager receives for each product group, compare it to sales data, and calculate the conversion from call to transaction for each manager and each product group. And then reconfigure the distribution of calls so that they are directed to those who are more likely to turn a call into a sale. A simple and seemingly logical scheme, but one that many companies still do not apply;
- Monitor the achievement of KPIs;
- Identify new customer needs;
- Find deficiencies in products and services.
An example of using speech analytics
Let's look at how speech analytics works using the call centre of a chain of clinics as an example. The service has processed more than 10,000 calls. According to the results, the company found out that
- only 75% of operators adhere to the script when communicating with clients;
- only 55% of the calls are targeted;
- conversion from targeted calls to transactions was 63%, i.e. 3,465 out of 10,000 people became clients of the clinic;
- 42% of the calls without an appointment have no objection work.
From the data obtained, we can draw conclusions:
- It is necessary to improve the quality of the call centre, namely: increase control over compliance with scripts, refine scripts to work with objections, refine scripts to increase the conversion of targeted calls.
- It is necessary to revise the settings of advertising channels to improve the quality of traffic and optimise marketing costs.
This is a small part of the data obtained and the real company's conclusions based on it. In addition, using speech analytics software helped the call centre identify the most effective agents and cut those who did not meet the requirements from the team.
So we see that the use of speech analytics is possible in all areas involving phone calls. And with the help of the results it provides, it is possible to increase the efficiency of departments, increase conversion rates, optimize marketing costs, identify the most effective employees, and make other decisions that ultimately lead to an increase in company profits.