The annual Cisco & Dimension Data Speech survery 2009

Never one to rush into comment, the blog is pleased to note that the annual Speech Survey from Cisco, Dimension Data and (new this year) TellMe. For those interested, a PDF of the report is available from Cisco.com here, and I recommend it highly.

I covered the report last year (in the post "It's time again for the Cisco/ Dimension Data Speech Survey"), so I'm please to see that the 'annual' report is continuing to be annual! This may seem a bit silly, but in the downturn many companies have cut previously committed marketing programs and so it's good to see that Cisco, Dimension Data and Tellme are continuing to invest in this research.

There are three key findings of the report that I found quite interesting:

1. Online self-service is critical when thinking about IVR: Very interestingly, the report found that 44 per cent of consumers use online self-service first for their customer service requirements. Some 52 per cent of consumers between the ages of 16-34 fall into this category. As the report says “The increasingly multi-channel nature of customer service creates a real challenge for enterprises and vendors to design and deliver service experiences that are consistent across channels,”. It's clear that the voice portal and re-use of presentation components is likely to be the future way to go for IVR development.

2. Customers don't like most speech implementations: Some 41 per cent of consumers say they would prefer to use speech recognition as little as possible, while 15 per cent of enterprises have this position. Only 3 per cent of vendors have this negative response. Reinforcing consumers’ dislike of speech recognition, 40 per cent of respondents said they avoid using speech systems “whenever possible”. Only 25% of consumers say they would be happy to use speech solutions again.

3. Customers and Enterprises don't see speech as bringing the same benefits: While vendors and enterprises largely viewed speech recognition’s ability to reduce waiting times as an improvement in customer service, consumers did not, the report found. In fact, he number of consumers who perceived no benefit to using automated services had grown from 20 per cent in 2008 to 31 per cent in 2009. “It clearly shows a failure on the part of vendors and enterprises to explain the benefits to consumers and highlights an area for improvement,” the report said. “To achieve a shift in customer perception, vendors and enterprises need to actively invest in delivering and promoting the perceived benefits and educating consumers on the additional, less apparent benefits.”

I think the key issue here is the user experience. If you deploy speech for inappropriate functions and do so purely to cut costs, then (oddly), customers do not feel valued or that they have had a good experience. I found it very interesting that the report also found that one reason for the high level of consumer dissatisfaction with speech recognition systems came down to poor or non-existent levels of integration with the next stages of customer service which resulted in callers having to repeat all their information again to an agent once they got there. It strikes me as elementary, but if you are going to spend all that money on a speech implementation, then ensuring that the data you capture in the self-service environment is passed on to agents is surely a basic step!

All very interesting and highlights that some of the success challenges that Speech faces are not the ones that vendors tend to think of!