I have been tracking the number of publicly announced speech recognition deals for the past two years and the results are midly disturbing (for those in the speech market). In 2005, there were approximately 160 publicly announced customer wins; in 2006 there were only 95. That's almost 50% less. Now here's the half-full, revenue figures and port shipments from vendors clearly indicate that the speech market has grown in 2006. In 2005, I projected that the number of publicly announced customer wins would eclipse that of those in 2005 but I was wrong. Now as an analyst I try to be accurate in my assumptions and methodologies but I know that I cannot divine the future with precision (as much as I would like). So I'm left to ponder and analyze what has happened in the market in retrospect and analyze that data and apply that to future forecasts. Here are my conclusions for the lack of PR for speech in 2006...

1) The companies that have deployed speech are expanding their investment in the technology and deploying across other areas of their business. These companies will not typically engage in additional PR with vendors. And vendor marketing departments are not effectively incentivizing customers to engage in PR / marketing campaigns.

2) Speech investments outside of server systems (such as embedded and PC-based) is on the rise. Embedded deals are typically very large and, as a result, there won't be as many deals as there are for network-based speech. PC-based deals are typically small and consumer focused (outside of healthcare transcription) and may be looked at as not 'PR worthy' until a certain number of licenses have been sold or large name companies become customers.

3) Companies are more tight-lipped about publicizing their investments in speech due to tighter regulations around information dissemination. There was a lot of activity among tight-lipped Tier 1 and Tier 2 financial services institutions in 2006.