Conversation Analytics® is different Than What Competitors Are Doing
3 min. read
We’ve received several questions recently about how Conversation Analytics® is different from some of the call analysis methods our competitors are employing.
In this blog, we will provide a brief overview of all of the options you have for call analysis.
We’ll try to discuss the options our competitors offer without actually using competitors’ names, because, you know, that would be weird and stuff.
There are two competitors that analyze calls using human call scoring. Actual humans listen to the phone calls and actually score them. They manually fill out criteria on a scorecard to come up with a total call evaluation. These scorecards can be quite lengthy–sometimes over 20 or 30 criteria.
(We also offer a human call scoring model for clients that insist upon it).
Human call scoring is generally used to evaluate sales performance. It provides really, really good information for training companies and internal sales training departments. The information can be very, very granular because actual humans are scoring calls.
The two biggest problems with human call scoring are these:
1) It Is Expensive – Using humans to score calls get expensive. Why? Because you have to pay actual humans to score the calls. Scoring a 5 minute call will take 7-8 minutes. That means one person is only able to score a few calls each hour. And even if you outsource it overseas, the cost adds up very, very quickly.
2) Questionable Accuracy at Scale – Humans can score 1 call correctly, 2 calls correctly and even 5 calls correctly. But, if you are asking a human (or a team of humans) to score hundreds of calls each day, there will be mistakes.
It analyzes phone calls using speech recognition technology and natural speech algorithms developed by Convirza.
Humans are not needed. The cost goes down.
The calls are simply fed through the Convirza’s Conversation Analytics® tool and data come out the other side.
Conversation Analytics® provides data regarding how frequently the agent asked for the business, if they were polite, if they were empathetic, and if they took ownership of the call. These are important high-level metrics that a sales training company could use to evaluate employees.
This is a very useful tool.
One of our competitors offers a service that is basically human call tagging. They don’t do a full analysis of the phone calls, but rather, they will listen to the call and determine if the lead was good or bad and if the call converted to a sale. In other words, they simply ‘tag’ the call with a couple of useful labels.
(Again, we also offer this model).
This is a very useful tool for marketers seeking to determine the relative quality of campaigns.
Again, it does the same thing without using humans. The data is cheaper and it is more scalable. Plus, because machines are extracting the data, it is available in near real-time.
Two of our competitors have recently released tools that analyze call content with speech recognition.
On its face, this appears similar to what Convirza’s Conversation Analytics® does. So…how is it actually different?
These tools transcribe the call to text and then allow a limited number of searchable queries into the call. For example, you could search for phrases like ‘book a room’ or ‘buy a car.’ You’d then receive reporting data whenever those phrases appear in the call transcription. Generally, the number of search queries you can enter is under 10. In other words, it is limited.
Conversation Analytics® is very, very different.
Conversation Analytics® also uses speech recognition to transcribe the call…but it doesn’t stop there.
Instead of allowing a few search queries into the transcription, Conversation Analytics® automatically analyzes 45 elements of the call using hundreds of thousands of individual queries.
Our lead quality metric, for example, contains algorithms that look for several thousand phrases and words in the call. Quite simply, the same data could NOT be extracted using fewer than 10 queries. And this is assuming that one would know which queries would be most effective, which you wouldn’t unless you listened to millions of calls… as we do.
In essence, the Conversation Analytics® is fully baked. Other solutions merely give you a limited number of ingredients. They give you a box with a potato and raw steak and forbid you from owning an oven (by limiting search queries), we give you a fully plated steak dinner with mashed potatoes, green beans, and a dessert.
Conversation Analytics® is a vastly different tool than a speech transcription technology. They both have their useful functionality, but there is no doubt that Conversation Analytics® is a more complete solution.
Originally published on September 3, 2015. Updated on December 20, 2018.