Marketing the Semantic Web, My Article on Nodalities Magazine

Talis has just released its latest issue of Nodalities Magazine, including an article I wrote on Marketing the Semantic Web. You can find it at www.talis.com/nodalities, and I copied the article below (together with URLs that were not included in the pdf version).

My article notwithstanding... this is a very strong issue of Nodalities, with an emphasis on the business applications of the technology, and my personal favorite, the quest for the "killer app". For those of you interested in the intersection of business and semantic technologies, I invite you in particular to check out the great articles by Andraz Tori, "New Web Cambrian Explosion - Killer Apps?"  and Eric Hoffer, "Semantic Technology in the financial industry". I won't mention that of Alex Iskold on search as I know you don't need an ego boost, Alex, after all those offers for financing pouring in at the semtech conference... ;)

So, here is my article in PDF version, which you're warmly encouraged to redistribute (as provided):

Download Nodalities_Greg_Boutin.pdf

And the HTML version:

Marketing The Semantic Web

Time To Deliver

The so-called Semantic Web is a set of technologies intended to facilitate data analysis and transformation by microprocessors, thereby extending the functional reach of machines. As such, it’s capable of enhancing a wide range of existing applications, and enabling new ones we have yet to picture. Given the breadth of possibilities and interests the semantic web covers, it will come as no surprise that there are many ways it can be marketed.

Building general awareness of the possibilities offered by those technologies is critical to getting to the so-called tipping point. In that respect, the semantic web has so far done a pretty good job. With the endorsement of the W3C, and Tim Berners-Lee as its top rep, it had the perfect seal of approval to start with. The industry has used this as a stepping stone towards enhanced awareness among technologists and early adopters. Just in the past months, there were articles about the semantic web in the tech sections of The Economist, the International Herald Tribune, the New York Times, the Financial Times, and in Scientific American.

Driving user adoption has proved to be a bigger challenge. Invented by the web equivalent of rocket scientists, the initial offering was too complex. For the most part, it remains quite daunting even today: simply including semantic capabilities on one’s website remains overkill for most webmasters, for example. Semweb technologies have managed to buy time by delivering a few narrow benefits, first in vertical B2B markets (see W3C semantic web case studies, or the offering of Zepheira), and more recently, in the consumer market, through services like Powerset, a semantic search engine for Wikipedia; UpTake (previously Kango), a semantic travel search engine; and of course Twine, the most imaginative semantic service so far, but still a semantic mashup between social networking, RSS feeds, and del.icio.us, whose key benefit appears to reside in its bookmark recommendation engine. Among larger corporations, Reuters has been the first one to buy into the concept with its OpenCalais service.

In a positive development, semweb technologies are increasingly infiltrating existing value propositions: Yahoo is testing the waters with its Search Monkey platform and, having acquired Powerset, Microsoft is expected to rapidly incorporate semantic capabilities in its search engine. I was told that Google already makes use of some sort of semantic indexing capabilities in its search engine, although I’ve seen little evidence of it so far. I’d make a wild guess here based on limited evidence I collected, and say that the search gorilla does not have a search monkey in the making, and instead is taking a statistical approach to tackling semantics, which might lead to some interesting development.

All in all, the semantic web has had difficulty delivering on all those promises it made. And so everyone is now eagerly awaiting a killer value proposition to catalyze market adoption and bring those technologies into the mainstream. Corporate search, travel search, wikipedia search, recommendation engine: those are all nice-to-have apps, but not killer apps. And so, at this stage, marketing the Semantic Web is not so much about PR and nice brochures as it is about building marketing right into the products, to get that killer app out and silence the cynics.

But how do you go about building a killer app for such complex lab technologies?

My Personal Recipe For Killer Value Propositions

For all its long history and numerous gurus, marketing surprisingly remains much more of an art than a science, and it is especially wasteful and ineffective at producing killer value propositions. In my previous roles, I have been grappling with a number of theories and concepts supposed to achieve just that. Most recently, I used the Pragmatic Marketing framework, which I implemented through a modified Agile Programming process adapted to fit our marketing needs.

Quickly, the Pragmatic Marketing framework became just a nice-to-have checklist to assess whether we were covering all the basics, and to oversee the interactions between the different marketing pieces I was putting together. I found that Pragmatic Marketing was not designed to come up with killer value propositions. In fact, it itself claims to be dedicated to “managing and marketing technology products”, implying the product already exists. Ultimately, what I learnt through all this is that common sense and logic, supported by some analytical firepower, a strong sense of the big picture, and the intellectual readiness to let the market show you what works and what doesn’t, beat fancy theories and jargon every time.

With this in mind, the first piece of advice I have for anyone seeking to find the semantic web grail is to keep things simple. As I wrote last month on my blog, that generally involves simplifying and cutting things to the chase. Clarifying and focusing on simple benefits. This is notoriously difficult for many technical founders, and yet they must learn to let go, or risk falling into market oblivion. For more on this, I invite you to read the chapter on the Knowledge Curse in the excellent book Made-to-Stick.

A second piece of advice is to not define yourself too narrowly as a semantic web company. For one, you want to avoid marketing myopia, as the killer app may be one powered by “semantics + some other technology”, and for two, it is generally more sustainable to define yourself by the benefits you provide rather than by the technology that supports you. If superior competing technologies come along, the brand that you have built by serving your customers will generally still grant you the best sit at the table.

To complete that work, finding an objective marketing lead and empowering her or him to deliver is essential. That person should be eager to do four things:

• learning quickly about your potential customer segments and finding out what they do;
• learning quickly about your own competencies and resources and finding out what gives you an edge. You need to have a perfect understanding of your competences, i.e. what will make you better than others at addressing those problems;
• learning quickly about your competitors (and potential partners) and finding out what could give them an edge.

And the fourth, having your marketing lead continuously synthesize this into a benefit analysis that details out the main actions you intend to help users with. The key to success is to conduct all four marketing activities in a quick iterative manner so that, when you learn something about your customers for example, it is immediately reflected into your view of the competitive landscape, in the benefits you are working on delivering, and in the competences you need to that end. Let me emphasize that further: the quest for a killer value proposition is about quickly iterating between knowledge of your capacities, those capacities of your competition, the needs of your consumer, and harvesting new insights from it every time into a coherent benefit analysis feeding your product development process.

Once the core benefits you plan to deliver are defined and prioritized, you need to turn them into features, bundle those features into products, and plan for the product development by having the Product Manager use those inputs to develop a first draft of the product requirement document.

The key idea there is to sequence your launch(es) as much as possible, so that you test your proposed value propositions in the most cost effective way, rapidly seeing what sticks and what doesn’t. Sequencing also helps by ensuring that your development always has an immediate deliverable to work towards. Lastly, and I’ll keep the detailing of that for another time, build the product so people talk about it to each other (I found the book Word Of Mouth Marketing very useful in that regard). That means having emotional benefits too, something web start-ups often disregard at their own peril.

I have summarized the above in the chart below. C’mon, it’s not that fancy-pants, is it?

[note: the jpg doesn't turn out too nice on the web and the pdf, so feel free to ask for a clean version by email if you need it, by sending me a comment to this article - I won't publish it]

 

You’ll see references to the concepts of Organization, Culture and Processes in the chart. Indeed, building a killer value proposition is not a process in isolation, it needs to be managed from the broader context of motivating people, focusing their efforts on the right task, organizing handovers and interfaces, and learning from it all, so the organization gets better and better at it over time. This is often a difficult one for start-ups with limited resources, and yet a key success driver. Taking an even broader look on the shared drivers of success start-ups, the book Blueprint to a Billion furthers the debate to board organization, alliances, management teams and other things important to crafting and launching killer value propositions.

Digitizing and Monetizing Existing Needs

One of the long-lived myths I’d like to dispel, and which will take us back to the semantic web, is one that I often hear from venture capitalists: “don’t try to build a new market from scratch, go where the money is”. Venture capitalists are not wrong, it’s just that the way they express this idea is rife with potential misinterpretations. It leads too many start-ups to try and beat Google at the search game, for instance.

Creating new needs is close to impossible. Going after existing needs addressed by online players, too, is extremely tough. Common wisdom says that your technology needs to be 10 times superior to the current incumbent’s technology if it stands a chance to overcome it. Incumbents have had time to look at the problem from all angles, and they have normally prepared themselves to all potential direct attacks.

Instead, take existing needs that are neither digitized nor monetized, and monetize them by improving on the ways they are currently met. I didn’t invent this; it is a key lesson from experience, also highlighted in a book called the Innovator’s Solution (one of a few non-fluffy recent marketing books!). You’d be surprised by the number of tasks today that are still a real pain to accomplish and would greatly benefit from online assistance!

One such need I uncovered in my past explorations, to illustrate the idea, is that of synthesizing information. There are a lot of applications out there to search information, much less so to put it together in a coherent manner once I have found the data I was looking for. How do I do that today? I paste things from web pages into MS Word, or save them onto my hard drive. Then I review them, one by one, taking what I need and leaving the rest. I use my brain for the most part, and a few basic tools that are only a marginal improvement over paper. There are no tools yet that compete with my brain for those tasks. And what if I wanted to find out at a conceptual level where we agree and where we disagree on the results of the US policy in Irak? Again, the optimal way today is to discuss it directly with you, or a long winded conversation online. I’m sure the semantic web could enable more productive ways of doing this. What if I wanted to connect on LinkedIn to all the bloggers having blogged somewhere about the semantic technologies conference? Or to get a timeline summary of all the dates and event present in a particular set of websites, together with a number of new visualization possibilities for existing web information. Pretty much impossible today. Wow.

Those are all existing needs that are not addressed optimally, and that semantic technologies can solve. Marketers can review existing processes to see where people spend time, and improve parts of those processes. Much easier than vying to take dominant companies out of their thrones. The bigger and more ubiquitous the need, the more “lethal” the value proposition. Just look beyond search.

In sum, to get to a killer value proposition, take a technology, simplify it, focus it (on clear benefits), structure it, iterate it, sequence it, crashtest it, sample it out, evangelize it, and watch it grow! Hopefully those thoughts will help a little to build the semweb champion which I strongly believe will be instrumental in taking our technologies to the masses. The “semantic web” concept is complex and reflects a wide brochette of technologies and solutions: what we need now is a leader who will dispel that fog and finally leave the industry with a clear web forward!