MyQM – SOME PRACTICAL SITUATIONS
Here you’ll find various use cases of our solutions in several situations.
Case 1
FIND KEY WORDS WITH THE “MOST USED WORDS” GRAPH AND THE DIFFERENCE FILTER
Bank
An IQM analyst working for the operation of a bank was analyzing the “most used words” graph and realized that the word “problem” was frequently used by the Bank’s customers. This obviously attracted the attention of operations and the request to analyze this specific word “problem” in more detail. In a detailed analysis using the “Filter – 1 difference” tab, the IQM analyst discovered and realized that the word “problem” was recently often used in combination with the word “ATM”. Indeed, after an internal audit, the Bank confirmed an ATM problem. After further investigation, a software problem was involved which prompted customers to call the call center due to the problem with ATMs. The bank immediately took action to resolve the problem.
An IQM voice analysis user wanted to find and analyze calls where the customer called for cancellation issues. So she adjusted the dates for the correct period. Then she wrote the word “cancel *” (“*” is the wildcard to replace any sequence of characters) in the quick search bar and chose the client channel to detect words or portions of sentences such as canceled, cancellation request etc… which could have been said by the customers. Once the calls were selected, she created a MyQM campaign and injected the selected calls to audit the reason for these cancellation issues.
Case 2
WORD DETECTION USING GENERIC CHARACTERS AND THE CLIENT CHANNEL – MYQM CAMPAIGN.
Tour operator
Case 3
USE OF QUERIES AND DIFFERENT VIEWS – CAMPAIGN OF NEW CUSTOMER PRODUCTS
Tour operator
An IQM voice analysis analyst wanted to check the forbidden words spoken by agents during calls over the past few weeks. To access this information, he opened the “Search calls” menu and chose the “forbidden words” query he had created last month. Next, he considered the “daily view” instead of the “general view” and inspected the calls on the daily list. What caught his attention was the fact that the first two days of last week were the days with the highest number of calls containing prohibited words. The analyst then focused on these 2 specific days using the date period selection and used the “agent view” this time. He could then control who used the most forbidden words.
Thanks to IQM, he noticed that a group of agents was at the top of the list of “forbidden words” users. When further analysis was done, he discovered that these agents were dealing with a new product campaign. When customers called and obtained information through these specific agents, words like “I don’t know”, “let me find out” etc … were very often used. To follow up on this discovery, an immediate training program was scheduled to quickly remedy this problem.
Site in case 3 described above, the client created a MyQM audit campaign with the calls selected from the list of “forbidden words”. After the audit process, the coaches realized that clients were difficult to manage and that the agent was under high pressure with these types of calls. Agents were then asked for feedback using MyQM’s feedback functions.
Case 4
USE OF QUERIES AND A QM CAMPAIGN (FEEDBACK) CAMPAIGN OF NEW CUSTOMER PRODUCTS
Tour operator
Case 5
USE OF FILTER-1, FILTER-2 AND DIFFERENCE OF FILTERS, TRENDS ANALYSIS
Credit Company
An IQM analyst wanted to analyze the fundamental reasons for canceling a credit card for a credit company. She used a query with the words “cancel my card” and tracked the changes in the trend analysis module. Following this, she realized that the product cancellation trend had increased dramatically since last week compared to the monthly average. So she tried to understand the main root cause using the IQM statistics comparison module.
She therefore chooses last week as the date range and the “cancel card” request as the main request in filter-1 and the same request but with 2 previous weeks as the date range for filter-2. She then compared the statistics with “client only” as the voice channel to focus on the client only. She obtained the following findings:
- The word “advertising” has increased significantly in the “Filter 1 difference” column compared to the previous week’s cancellation calls.
- When she clicked on the word “advertising”, she discovered that phrases like “shopping”, “bought”, “no bonus” and “no loading” appear most often and frequently used with the word “advertising”.
By clicking on one of these highlighted words on the text analysis screen, the IQM analyst selected and listened to a few calls. What she discovered was that customers were calling the contact center because of a credit card advertisement that promised an additional bonus on purchases made at certain malls. However, the promised bonuses were not granted after shopping. Due to customer frustration, cancellation calls have increased dramatically. Based on this IQM study, the credit company took immediate action to resolve the problem.
In a contact center where certain life insurance products are sold by the telemarketing service, agents are rated in relation to their sales performance. To improve the sales performance of all agents, the supervisors decided to use speech analysis to see the difference between the best performing agents and the rest of the call center.
Using the IQM statistics comparison module, the comparison of acoustic data and word comparison is performed for both sets of agents during the same period. What was discovered in the “agents only” channel was very interesting:
- Successful agents used more quantitative explanations compared to the rest of the call center, as highlighted in the filter-1 difference report.
- The performance agents’ silence rate was higher than the others.
- The performance agents’ interruption rate was lower than the rest.
- The monotony and speed of the performing agents was lower than the others.
Based on this information, the supervisors created a MyQM campaign and started listening to and auditing some of the calls again and discovering the following:
- Successful agents are more likely to use exact numbers to indicate customer earnings / savings based on customer specific profiles.
- By speaking slowly but more effectively, successful agents paused just after presenting their offers or revenue opportunities to customers and gave them time to think things over. As a result, clients were visibly more likely to accept the offer.
- Using this method, successful agents sold 40% more than other colleagues.
Using this valuable information, the supervisors asked the training department to organize training for all agents in order to use this methodology in sales operations.
Case 6
COMPARISON OF AGENTS USING VOICE ANALYSIS
Life Insurance Company
Case 7
FIND CALLS WITH TENSION AND ANGER
Telecom operator
An IQM voice analysis analyst decided to select calls where both parties were arguing and where the tension was visibly high. First, he made a request including the words that are frequently used in these types of calls. So, he created a new IQM request and added specific words like “shut up”, “don’t talk like that” etc … for the two agent and client voice channels. After that, he got calls that are probably conflict calls.
A telecommunications operator released a new product that included free talk time but no free SMS. To measure the knowledge at the level of its agents in relation to this new offer, the call center supervisors created an IQM query which inspected whether the agents were giving customers correct information or not. A request has been created with the following words: “Free talk time” AND (“Free SMS” OR “Free messaging” OR “Unlimited SMS”)
This gave the possibility of finding agents who gave erroneous information about this product. The parameter on the frequency of uses for words has been added in order to control the number of occurrences of words with the min – max values. Only the Agent channel has been selected..
With the successful calls, they created a MyQM campaign and they realized that some agents were indeed giving the wrong information. A training action was requested from these agents using the action function of the MyQM solution.
Case 8
CHECKING THE AGENT’S KNOWLEDGE – ADVANCED WORD REQUEST
Telecom operator
Case 9
Request by specific words frequency
Finance Company
The IQM analyst made a request where the agent’s attitude towards the client was monitored by trying to spot the excessive use of words like “sir”, “no” and “madam”. The number of uses has been defined with minimum 2 and maximum 3 occurrences. Using these parameters, he was able to locate agents who used these words more than once and an example of the following results was found:
- Yes sir … It is true that your account is blocked sir … no, I have no information … no sir, you can get in touch with …
- No ma’am, you have to go to the bank and… No but, the bank is open until…. Thank you madam …
- Have you tried it sir…. no, no sir, you must connect to the web page that will open … if you just wait for the signal sir … no, the green …
- Hello madam how can I help you… no madam please wait a second… no but it is still possible to send money madam because…. no madam, unfortunately this is not the case …
The IQM analyst also used “word recognition trust” which is a value between 0 and 100 in the agent / client reliability sections. He used the values of 80 and 100, which means that the search is carried out among the words having at least a recognition confidence value of 80%. He also determined which part of the conversation the search words were in by selecting a percentage between 0 – 100 in the “Range” section. For example, selecting a percentage value between 80 and 100 searched within 20% of the end of calls and selecting a percentage value between 0 and 10 searched within 10% from the start of the calls.
An IQM analyst wanted to find the agents who had a quarrel or an argument with a client and created an appropriate request for this purpose. First, he took calls he knew were “real” argument calls. With the help of the statistics comparison module, he determined the characteristic acoustic profile of these calls.
What he realized was that the agents interrupted clients at least 5 times in quarrel calls, that the duration of the overlapping of the client and agent channels was on average 15 seconds and that the agent’s anger ratio was greater than 40%.
Second, he examined which words were used frequently in quarrel calls. The results showed that “Listen”, “absurd”, “differently”, “If you say so” were often used in these types of calls.
Finally, he executed the queries and examined the results using an audit campaign from MyQM. The first results gave 20 real quarrel calls out of the 25 calls that the request had identified.
Case 10
FIND AGENTS WITH THE STATISTICAL COMPARISON MODULE
Finance Company
Case 11
MERGER OF REQUESTS – CALLS BARRING
Internet Service Provider
First, the IQM analyst merged the existing query “Internet connection problem” with the queries “resolved problem” and “reported problem” and thus created two merged queries. The two merged requests therefore listed the clients with the Internet problem solved and on the other hand the clients whose problem had not been solved during the call with the agent.
Then, he used the IQM statistics comparison module to determine the keywords for these two queries. He discovered that calls from customers whose problem was resolved over the phone took longer than calls from customers whose problem was resolved over the phone. Also, he found that clients were more grateful to agents when their problem was immediately resolved. He also noted that restarting the modem often helped customers solve their problem. On the other hand, he also showed that these unresolved problems often occur because of regional infrastructure services. All of these deductions were made by the word differentials that were spotted by the system.
Using this data, the management of the Internet service provider saw the opportunity to reduce operating costs by setting up an IVR “self-service” service for customers. Once this was put in place, they noticed using IQM that customer calls (whose problem could be easily resolved) were fully handled by the IVR service and therefore no longer reached the agent more. The operating costs have therefore been considerably reduced.
Management also reported regional issues to the infrastructure division to give the necessary attention.
A contact center manager wanted to take a proactive approach for his clients by analyzing repeated calls from clients. First of all, using the “Non-FCR” graph (settings with minimum 3 calls within 48 hours), he noticed that a quantity of calls and therefore a huge cost due to the “non FCR”. He needed a more detailed analysis. After researching the reasons for the “Non-FCR” calls, he discovered that most of these repetitive calls were related to customers who had recently changed their mobile phone in their subscription and they called to find out if the change had been approved in the end. Based on this information, the contact center management made the decision to send a regular text message to these types of customers. By this measure, they managed to reduce their “Non-FCR” calls by more than 55%.
Case 12
NON FCR: REPEATED CALLS
Internet Service Provider
Case 1
FIND KEY WORDS WITH THE “MOST USED WORDS” GRAPH AND THE DIFFERENCE FILTER
An IQM analyst working for the operation of a bank was analyzing the “most used words” graph and realized that the word “problem” was frequently used by the Bank’s customers. This obviously attracted the attention of operations and the request to analyze this specific word “problem” in more detail. In a detailed analysis using the “Filter – 1 difference” tab, the IQM analyst discovered and realized that the word “problem” was recently often used in combination with the word “ATM”. Indeed, after an internal audit, the Bank confirmed an ATM problem. After further investigation, a software problem was involved which prompted customers to call the call center due to the problem with ATMs. The bank immediately took action to resolve the problem.
Bank
Case 2
WORD DETECTION USING GENERIC CHARACTERS AND THE CLIENT CHANNEL – MYQM CAMPAIGN.
An IQM voice analysis user wanted to find and analyze calls where the customer called for cancellation issues. So she adjusted the dates for the correct period. Then she wrote the word “cancel *” (“*” is the wildcard to replace any sequence of characters) in the quick search bar and chose the client channel to detect words or portions of sentences such as canceled, cancellation request etc… which could have been said by the customers. Once the calls were selected, she created a MyQM campaign and injected the selected calls to audit the reason for these cancellation issues.
Tour operator
Case 3
USE OF QUERIES AND DIFFERENT VIEWS – CAMPAIGN OF NEW CUSTOMER PRODUCTS
An IQM voice analysis analyst wanted to check the forbidden words spoken by agents during calls over the past few weeks. To access this information, he opened the “Search calls” menu and chose the “forbidden words” query he had created last month. Next, he considered the “daily view” instead of the “general view” and inspected the calls on the daily list. What caught his attention was the fact that the first two days of last week were the days with the highest number of calls containing prohibited words. The analyst then focused on these 2 specific days using the date period selection and used the “agent view” this time. He could then control who used the most forbidden words.
Thanks to IQM, he noticed that a group of agents was at the top of the list of “forbidden words” users. When further analysis was done, he discovered that these agents were dealing with a new product campaign. When customers called and obtained information through these specific agents, words like “I don’t know”, “let me find out” etc … were very often used. To follow up on this discovery, an immediate training program was scheduled to quickly remedy this problem.
Tour operator
Case 4
USE OF QUERIES AND A QM CAMPAIGN (FEEDBACK) CAMPAIGN OF NEW CUSTOMER PRODUCTS
Site in case 3 described above, the client created a MyQM audit campaign with the calls selected from the list of “forbidden words”. After the audit process, the coaches realized that clients were difficult to manage and that the agent was under high pressure with these types of calls. Agents were then asked for feedback using MyQM’s feedback functions.
Tour operator
Case 5
USE OF FILTER-1, FILTER-2 AND DIFFERENCE OF FILTERS, TRENDS ANALYSIS
An IQM analyst wanted to analyze the fundamental reasons for canceling a credit card for a credit company. She used a query with the words “cancel my card” and tracked the changes in the trend analysis module. Following this, she realized that the product cancellation trend had increased dramatically since last week compared to the monthly average. So she tried to understand the main root cause using the IQM statistics comparison module.
She therefore chooses last week as the date range and the “cancel card” request as the main request in filter-1 and the same request but with 2 previous weeks as the date range for filter-2. She then compared the statistics with “client only” as the voice channel to focus on the client only. She obtained the following findings:
- The word “advertising” has increased significantly in the “Filter 1 difference” column compared to the previous week’s cancellation calls.
- When she clicked on the word “advertising”, she discovered that phrases like “shopping”, “bought”, “no bonus” and “no loading” appear most often and frequently used with the word “advertising”.
By clicking on one of these highlighted words on the text analysis screen, the IQM analyst selected and listened to a few calls. What she discovered was that customers were calling the contact center because of a credit card advertisement that promised an additional bonus on purchases made at certain malls. However, the promised bonuses were not granted after shopping. Due to customer frustration, cancellation calls have increased dramatically. Based on this IQM study, the credit company took immediate action to resolve the problem.
Credit Company
Case 6
COMPARISON OF AGENTS USING VOICE ANALYSIS
In a contact center where certain life insurance products are sold by the telemarketing service, agents are rated in relation to their sales performance. To improve the sales performance of all agents, the supervisors decided to use speech analysis to see the difference between the best performing agents and the rest of the call center.
Using the IQM statistics comparison module, the comparison of acoustic data and word comparison is performed for both sets of agents during the same period. What was discovered in the “agents only” channel was very interesting:
- Successful agents used more quantitative explanations compared to the rest of the call center, as highlighted in the filter-1 difference report.
- The performance agents’ silence rate was higher than the others.
- The performance agents’ interruption rate was lower than the rest.
- The monotony and speed of the performing agents was lower than the others.
Based on this information, the supervisors created a MyQM campaign and started listening to and auditing some of the calls again and discovering the following:
- Successful agents are more likely to use exact numbers to indicate customer earnings / savings based on customer specific profiles.
- By speaking slowly but more effectively, successful agents paused just after presenting their offers or revenue opportunities to customers and gave them time to think things over. As a result, clients were visibly more likely to accept the offer.
- Using this method, successful agents sold 40% more than other colleagues.
Using this valuable information, the supervisors asked the training department to organize training for all agents in order to use this methodology in sales operations.
Life Insurance Company
Case 7
FIND CALLS WITH TENSION AND ANGER
An IQM voice analysis analyst decided to select calls where both parties were arguing and where the tension was visibly high. First, he made a request including the words that are frequently used in these types of calls. So, he created a new IQM request and added specific words like “shut up”, “don’t talk like that” etc … for the two agent and client voice channels. After that, he got calls that are probably conflict calls.
Telecom operator
Case 8
CHECKING THE AGENT’S KNOWLEDGE – ADVANCED WORD REQUEST
A telecommunications operator released a new product that included free talk time but no free SMS. To measure the knowledge at the level of its agents in relation to this new offer, the call center supervisors created an IQM query which inspected whether the agents were giving customers correct information or not. A request has been created with the following words: “Free talk time” AND (“Free SMS” OR “Free messaging” OR “Unlimited SMS”)
This gave the possibility of finding agents who gave erroneous information about this product. The parameter on the frequency of uses for words has been added in order to control the number of occurrences of words with the min – max values. Only the Agent channel has been selected..
With the successful calls, they created a MyQM campaign and they realized that some agents were indeed giving the wrong information. A training action was requested from these agents using the action function of the MyQM solution.
Telecom operator
Case 9
REQUEST BY SPECIFIC WORDS FREQUENCY
The IQM analyst made a request where the agent’s attitude towards the client was monitored by trying to spot the excessive use of words like “sir”, “no” and “madam”. The number of uses has been defined with minimum 2 and maximum 3 occurrences. Using these parameters, he was able to locate agents who used these words more than once and an example of the following results was found:
- Yes sir … It is true that your account is blocked sir … no, I have no information … no sir, you can get in touch with …
- No ma’am, you have to go to the bank and… No but, the bank is open until…. Thank you madam …
- Have you tried it sir…. no, no sir, you must connect to the web page that will open … if you just wait for the signal sir … no, the green …
- Hello madam how can I help you… no madam please wait a second… no but it is still possible to send money madam because…. no madam, unfortunately this is not the case …
The IQM analyst also used “word recognition trust” which is a value between 0 and 100 in the agent / client reliability sections. He used the values of 80 and 100, which means that the search is carried out among the words having at least a recognition confidence value of 80%. He also determined which part of the conversation the search words were in by selecting a percentage between 0 – 100 in the “Range” section. For example, selecting a percentage value between 80 and 100 searched within 20% of the end of calls and selecting a percentage value between 0 and 10 searched within 10% from the start of the calls.
Finance Company
Case 10
FIND AGENTS WITH THE STATISTICAL COMPARISON MODULE
An IQM analyst wanted to find the agents who had a quarrel or an argument with a client and created an appropriate request for this purpose. First, he took calls he knew were “real” argument calls. With the help of the statistics comparison module, he determined the characteristic acoustic profile of these calls.
What he realized was that the agents interrupted clients at least 5 times in quarrel calls, that the duration of the overlapping of the client and agent channels was on average 15 seconds and that the agent’s anger ratio was greater than 40%.
Second, he examined which words were used frequently in quarrel calls. The results showed that “Listen”, “absurd”, “differently”, “If you say so” were often used in these types of calls.
Finally, he executed the queries and examined the results using an audit campaign from MyQM. The first results gave 20 real quarrel calls out of the 25 calls that the request had identified.
Finance Company
Case 11
MERGER OF REQUESTS – CALLS BARRING
First, the IQM analyst merged the existing query “Internet connection problem” with the queries “resolved problem” and “reported problem” and thus created two merged queries. The two merged requests therefore listed the clients with the Internet problem solved and on the other hand the clients whose problem had not been solved during the call with the agent.
Then, he used the IQM statistics comparison module to determine the keywords for these two queries. He discovered that calls from customers whose problem was resolved over the phone took longer than calls from customers whose problem was resolved over the phone. Also, he found that clients were more grateful to agents when their problem was immediately resolved. He also noted that restarting the modem often helped customers solve their problem. On the other hand, he also showed that these unresolved problems often occur because of regional infrastructure services. All of these deductions were made by the word differentials that were spotted by the system.
Using this data, the management of the Internet service provider saw the opportunity to reduce operating costs by setting up an IVR “self-service” service for customers. Once this was put in place, they noticed using IQM that customer calls (whose problem could be easily resolved) were fully handled by the IVR service and therefore no longer reached the agent more. The operating costs have therefore been considerably reduced.
Management also reported regional issues to the infrastructure division to give the necessary attention.
Internet Service Provider
Case 12
NON FCR: REPEATED CALLS
A contact center manager wanted to take a proactive approach for his clients by analyzing repeated calls from clients. First of all, using the “Non-FCR” graph (settings with minimum 3 calls within 48 hours), he noticed that a quantity of calls and therefore a huge cost due to the “non FCR”. He needed a more detailed analysis. After researching the reasons for the “Non-FCR” calls, he discovered that most of these repetitive calls were related to customers who had recently changed their mobile phone in their subscription and they called to find out if the change had been approved in the end. Based on this information, the contact center management made the decision to send a regular text message to these types of customers. By this measure, they managed to reduce their “Non-FCR” calls by more than 55%.
Internet Service Provider