Aug
14
Speech is the Future of Search
Filed Under B2B, Big Data, Blog, data protection, Education, eLearning, Financial services, healthcare, Industry Analysis, internet, mobile content, news media, Publishing, Search, semantic web, social media, STM, Uncategorized, Workflow | 1 Comment
Here is an extract from a story carried by the International Herald Tribune on 14 August 2012;
august 14, 2012
Quaid’s speech calling for religious, ethnic tolerance missing from Radio Pakistan’s archives.
Muhammad Ali Jinnah’s landmark speech at the Constituent Assembly’s
first meeting on August 11, 1947 in Karachi has been missing for decades
and all recent efforts to retrieve it have so far been in vain.
These days, Radio Pakistan runs an Urdu translation recorded in
somebody else’s voice of the same speech. Where the original speech
disappeared, and whether this was deliberate, remains an unanswered
question.
It may be no coincidence that the missing speech has these famous
words in it: “You are free; you are free to go to your temples, you are
free to go to your mosques or to any other place of worship in this
state of Pakistan …You may belong to any religion or caste or creed—that
has nothing to do with the business of the state.”
It was also in this speech that the founder had said that the first
duty of a government was to maintain law and order, “so that the life,
property and religious beliefs of its subjects are fully protected by
the state.”
Important words, and never more so than at this time. And almost certainly not lost – mislabelled, misfiled, disguised by inadequate metadata, maybe, but it is worth a small wager that the speech turns up one day – so very many things do. The world is full of attics and libraries from which lost symphonies, early works of poetry, the juvenilia of great writers reappear with monotonous regularity. This speech will be found.
But the story interested me because it provided a graphic reminder of the problem and potential of speech, and returned me to a long held conviction – that voice is the real future of search. However, I no longer believe that the route to this is through Google and the other search engines. though I am aware of Google’s downloadable voice search app, and I am sure that the pace set in this area will accelerate as Google get even further invested in the future of the smart phone. However, we need semantic technologies that can treat text as voice and vice versa for search purposes, and while we have evidence of many attributes in this pipeline, we seem to be a long way away from finding a universal solution, and one that resolves the legacy content issues as well.
I am coming to believe, however, in the explosive growth potential for voice in business, partly because of Siri and its ilk, partly because of the need to get more functionality into the phone than the keyboard will allow, but mostly because I can now see a group of very relevant business areas where being able to move seamlessly from voice to text, to be able to search both using either, unlocks productivity gains that cannot be attained in any other way. My conclusions on this were formed by following two companies quite closely. One was Aurix (www.aurix.com), a former UK Defense establishment company privatized within QinetiQ and now owned by Avaya. No prizes for guessing where their voice interests began, but their interests now lie far beyond security and intelligence. The other is BigHand (http://uk.bighand.com), bought earlier this year by my colleagues at Bridgepoint, where I do some media advisory work. It happens that both of these are UK technology companies, but I am sure that we could find equivalents for them in the USA or Israel.
Clearly a vital sector for voice search remains security. Searching voicemail alone when required is a major undertaking (not even the News of the World in its prime had the right technologies). Beyond this, media and broadcasting is surely a primary market. No one who drowned in the ocean of superlatives surrounding the London Olympics can doubt that heavyweight voice technologies of the sort that Aurix deploys will be as critical as the major investment put together by the BBC, partnering with MarkLogic, to put in the text-image-video handling platform that sat behind the BBC Sport website. And then we put together the fastest growing sector – monitoring and searching voice messages, recorded conversations and realtime calling in direct selling and customer service contexts. The productivity gains are as large as the range of uses is wide – checking compliance and script adherence, learning from common complaints, measuring call centre workloads, analysing trends in customer response etc etc.
And there are two marketplaces where voice records, the ability to attach them to text records, and to search both at the same time, has always been important, even when it wasn’t possible. One is the law practice market, and the other is the health market. Here there are solid traditions of voice recording, but real productivity gains to be made (for example, in legal eDiscovery) by using effective voice search. BigHand are market leaders in legal dictation and have the exciting prospect before them of what seemed a limited market a few years ago now opening out in an interesting way to embrace technology change which will then move into education (voice reports?), surveying, engineering and then some of the science research disciplines. As a law database publisher in 1982, I now have the delightful prospect of seeing another wave come ashore in the same market with very similar productivity, and compliance, advantages.
So here then is a brief sketch of a demand-led digital voice revolution. In 2020 we will ask our screen for research results, and define if we want them by ear or by eye – bearing in mind that some of the results will be transcripted voice turned into text, and others will be text turned into speech. Around then we shall find the missing speech in this news story – and admire again the wonderful sentiments of the speaker.
Aug
5
In Search of Enduring Value
Filed Under B2B, Big Data, Blog, data protection, Financial services, healthcare, Industry Analysis, internet, online advertising, privacy, Reed Elsevier, Thomson, Uncategorized, Workflow | Leave a Comment
Yes, I remember Dun and Bradstreet. In the old UK headquarters in High Wycombe, the “white elephant” building that was intended to become the global data centre (in the days when you concentrated data instead of distributing it) had a waxwork figure in the foyer depicting a frock-coated Mr Dun (or was it Mr Bradstreet , or Lewis Tappan , the real founder?) collecting together the vital credit rating clues of the 1840s, as well as a discreet reminder that Abraham Lincoln had acted as as a data collector in Illinois in the 1850s. But the company I knew was a large and prosperous portfolio player, the very demonstrator for the theory that markets never go all bad at once, and that change in one can be nurtured from sustained growth in others. Under the portfolio, if I recall correctly, there nestled Nielsen in market research , IMS Health in medical market research, Donnelley in directories and market-leading marketing services, Cognizant in technology, Moodys in global rating services and Hoovers in company profiling. Never, you might say, was there a better example of a company with a portfolio of related interests who could interconnect these data collections to create fresh value in wider marketplaces – and take it all global as D&B itself was already going global. What a huge opportunity that now seems to create value through connecting hitherto unconnected data values and effect the type of transformation that Thomson Reuters are now attempting.
But the voices that D&B listened to were not the voices that said things like this. They were the siren voices of the market, who said that short term values could be increased by selling off all these allied companies, organizing the buyback of shares on a major scale and creating a greater value in the parent than would have been possible if it had remained in the group. So all of these companies went, and mostly to private equity buyers. But this was still not enough in terms of value creation, so the majority of the overseas subsidiaries were franchised to local operators , with valuable operations like China, Russia, Australia, and Germany having their data leased out to previously competing market players, who would then pay fees and royalties and contribute to the global data holding (now around 200 million companies and 53 m details of directors) in return for local re-use.
Markets and managements change, and over time D&B have bought back Hoovers ( revamped and without its research services ), and bought out their local franchise holders in places like China (where they now face a local data privacy infringement case) and Australia. No one adds the loss of value from these buybacks to the long term calculation, but presumably at some point the company became aware that by paring itself to the operational bone in search of value, it was actually losing opportunity. Now we gather (Wall Street Journal, 31 July 2012) that the company has been seeking a buyer for the past year, and has now appointed financial advisors to “explore opportunities” that may or may not lead to a sale.
D&B know all about creating value. In 1963 they created the DUNS number , forcing consistency and their own metadata on a market they meant to dominate globally. In just the same way IMS Health created its proprietory BRICS system for measuring medical activity in a community. Here were the forefathers of dominant metadata systems, whose value creation (think of the recent Thomson Reuters argument with the European Union over its RICS metadata nomenclature) is the bedrock of value add in data driven systems. Given its birthright, D&B might have been the dominant player today in value-added workflow services and systems offering solutions in areas like procurement and customer profiling. Question: has it been competitively outflanked by Experian (compare performances in Brazil, for example, where D&B have been since 1933) and lost touch with a value growth plan beyond buying back the franchises it once leased out?
It seems to me sometimes as if value in the sense that markets use the word is in fact a bell curve. It is clear how the asset sales drove D&B’s valuation up one side of this and how it has peaked through an inability to add fresh value in the narrow front on which it now operates, without the advantages of platform integration and Big Data-style development. It is possible that in places where these factors have come together (insurance risk in the US would be an example, where Lexis Risk use these elements to dominate in a related but consumer-orientated marketplace) it may be very difficult, without very extensive strategic partnerships and joint venturing, for D&B to prevent itself from losing ground.
So does this mean sale at a discount to a private equity player, or are there trade buyers who would offer a premium. Before Sanford Bernstein suggest again that Reed Elsevier should sell Lexis Law and buy this, let me just say that, in my view, the only real potential fit is with Thomson Reuters, and they probably have enough on their plates without trying to absorb a $1.7 bn revenue player, or Bloomberg. Competitors would all face anti-trust issues, but enterprize software and systems players might be interested – and D&B already has good links with Oracle.
A friend reading the last two pieces on this blog – a sort of odd trilogy on valuations – kindly asked how the UBM announcement that it “might” sell its “data services” fitted into all of this. Surely, as with D&B, we do not sell data at the moment: instead we try the alchemy of value add. So I have looked at this too, and am now even more confused than when I started. For example, by “data services” UBM appear to mean the databases from which they once sourced their print directory products. Apparently they have found that advertising online earns such diminished CPMs that it is very difficult to sustain the services. Similarly with Tech Insights, which they acquired and seems to suffer from the same problem. Is this surprizing? Not at all, since unless that data can be recombined with other internal or third party content there is no real hope of getting a subscription value from it. Advertising online is always going to be dodgy territory and at best a subsidiary income source.
And what does all this demonstrate ? Is the portfolio model broken for good and cannot ever be mended ? Or maybe D&B were right fifteen years ago , as Thomson Reuters are now : you can build portfolio if the players you buy are data-related and if you have platform and distributed search going for you . When D&B lost faith in their original model they did not have the technologies to do the job . So they followed the equity market view of value , and the chronic short term thinking that results from that has brought them to this . Now comes a more interesting question : what is credit rating and how do you reconstruct the service future of this marketplace ?
« go back — keep looking »