There’s a well-known saying that all publicity is good publicity, but you have to wonder whether the team over at Radar Networks feel that way this morning.
The company is behind Twine, which was unveiled to much fanfare at the O’Reilly Web 2.0 Summit last November. Having just closed a $13 million Series B investment round intended to carry them toward full public access, Radar Networks recently began the process of expanding their early beta access programme. Read/WriteWeb Lead Author Marshall Kirkpatrick and I were amongst those invited in during this current expansion.
Marshall blogged his first impressions to Read/Write Web yesterday, and in amongst his reporting of Twine’s strengths he also identified a number of issues that he felt could impinge upon wider adoption when Twine leaves beta later this year. Given the attention that Twine has been receiving, it was wholly unsurprising that this review, written after a few days inside the application, should attract such attention. Semantic Web nay-sayers gleefully missed the point, screeching ‘I told you so’ from the rooftops. Twine (and Semantic Web, to a degree) groupies actually missed many of Marshall’s key points too, in their haste to weigh in and lambast his methods, his approach, and his (apparently unforgivable) failure to fall head-over-heels in love with the application.
Radar Networks’ CEO Nova Spivack responded in a more considered fashion, reminding all concerned that the application is still in beta, and noting that resolution of many of Marshall’s issues is already in hand within the planned programme of activity ahead of their public launch.
As I said, I was also invited into Twine this month, and I’ve spoken with Nova about it on several occassions. Twine is - and never claimed otherwise - an early beta. Of course there will be teething problems, and Marshall found some of those. Marshall knows it’s a beta, too, and also knows that some areas of the application aren’t as polished as others. That knowledge isn’t a reason to sit silently by, waiting for the big launch. Invite a prominent blogger like Marshall into a beta programme, and you really do have to expect him to talk (honestly) about what he sees. I’m sure Nova and his team at Radar would expect nothing less. I want to spend far more time pushing the boundaries of Twine before offering a review of my own, but having seen more than my fair share of beta applications in the past, my early impression is that Twine is more or less exactly where I would expect it to be at this stage in the process; plenty of potential, plenty of good bits, and plenty of places where some tweaking may be required before I’d happily point my mum at it.
The whole affair raises a number of linked issues, from a fledgling company’s need to build expectation in order to attract investment, through to the confusion engendered by associating too strongly with a meme or technology at the expense of being associated with solving a problem.
First, then, expectation building. Raising venture capital in order to build an application such as Twine is a perfectly valid - and very common - way to get things done. It does bring with it the need to hype yourself though; you need to paint a picture in which you, your company, and your product are the best thing since sliced bread in order to attract the attentions of venture capitalists and their open cheque books. Those investors need a return on investment, and will often want to see that within a relatively short period of time. So as well as promising an awful lot, you probably have to appear to deliver on it more rapidly than you would if your developers had their way.
Radar Networks promised a lot with Twine, and they’ve been rewarded for that by investment. Having promised so much (and I’ve no reason to believe they won’t deliver on their promises in time), that first look at the beta is almost inevitably an anti-climax. That doesn’t mean there’s anything wrong with the product. It does mean that the Radar Networks media machine, which so successfully built the hype in the first place, has a tricky job to do now in managing expectations as people actually begin to see and use this early iteration of their product.
For companies such as my own, re-investing revenue from a well established suite of products in order to build something new, the pressures are very different. There is no big-bang injection of cash, and there is no flurry of 'easy' attention on the back of that. However, there is steady, planned, and controlled growth at a pace of the company’s choosing. Both approaches can result in eventual success, just as both can fail. Neither is necessarily more ‘right’, but the market conditions, prioritisation and delivery timescales differ substantially.
And secondly, association with a meme. Long before Twine saw the light of day last November, Radar Networks CEO Nova Spivack had worked hard to associate himself and his company with the Web 3.0/ pragmatic Semantic Web meme. Any article on the topic could be guaranteed to include a sound bite (or more) from Nova, and he became strongly linked to the growing realisation that semantic technologies were (finally) escaping the laboratory and offering real value in the consumer and enterprise environments. That association was of value to Nova and to Radar. It also helped the rest of us, as Nova’s messages clearly resonated with those in the mainstream media that everyone was trying to reach. As Nova - and Twine - became more visible, a host of other ‘semantic’ applications gained credibility and visibility by association.
The danger, now, is that over-analysis of the early stages of a single application like Twine is mis-applied to other very different applications that happen to use some of the same technologies. As I alluded in an earlier post about Twine, raising awareness of the underlying trends and technologies that make something like Twine possible is one thing. To then market (if that’s what they’re doing) a consumer product as ‘a Semantic Web application’ rather than focussing far more on the problems that the application is intended to solve is something else entirely, and far more risky.
Radar Networks, Adaptive Blue, Talis, Hakia, Powerset, ClearForest, OpenLink, Garlik, Joost, Metaweb, TrueKnowledge, NetBase. All these (and more) are building applications that depend heavily upon Semantic Web specifications and other semantic technologies. They’re not ‘just’ building ‘a Semantic Web application,’ though. They are building applications that solve specific problems, and that use semantic technologies as part of the toolkit that makes their application (a travel portal, a social bookmarking application, a community information product, or whatever) better than its competitors. Few, if any, of these companies are competing with one another today. The companies they’re competing with are those within their chosen market segment, and that is where the semantic technologies they use should be judged. Is the application better/ cheaper/ faster/ more scalable/ whatever, and did semantics help that?
Radar Networks received some honest criticism from one well connected technology blogger. That’s no reason to write off Radar, Twine, or the Semantic Web.