jump to navigation

Beyond Vertical Search to Business Networks January 30, 2008

Posted by bernardlunn in B2B Media, Vertical Search, Web 3.0 Semantic, social networks.
add a comment

Vertical Search is one of those confusing terms that means many different things, depending on where you are coming from.To most RWW readers, Vertical Search tends to mean “the search space that Google has not yet grabbed and that does not require a major technology breakthrough such as natural language search”. That’s a good enough definition from a start-up perspective.For traditional media, Vertical Search is also about creating a space that Google cannot simply steamroll over. Traditional media may call it Rich Data or Information Services or Data Products, but the end goal is the same.For the sake of consistency I will continue to call this Vertical Search, although the opportunity is deeper than simply search. In fact, in order to build sustainable advantage against the Google steamroller, it has to be deeper than just search.As Alex Iskold explained recently, Google Custom Search is setting the bar for Vertical Search Engines. This looks like a smart play by Google and they will grab the big low hanging fruit very well. For example, search for something related to healthcare and you can see that Google already understands concepts such as Symptoms and Treatment.That still leaves a lot of opportunities within niches that may look small at first glance (and which won’t justify any focused effort by Google). You can dismiss these niches as “picking up peanuts in front of a steamroller” but when the speed and direction of the steamroller is fairly clear, there are lots of opportunities if you are agile and quick on your toes -)You may also be surprised by how much money can be made in small niches. My favorite unglamorous niche is ASI, the Advertising Specialty Institute. Their business is “promotional products” or as they put in their site: “think of all those freebies like mugs and T-shirts you see at trade shows”. It is an $18.6bn industry which is pretty small (compared to the big markets targeted by vertical search start-ups) . Yet a few years ago ASI’s revenues were well north of $50m and they were highly profitable and growing at a reasonable clip. This type of niche won’t get you on the front cover of Fortune but it could make you rich.ASI is more than vertical search, a lot more, but the core of the business is data about their industry. That data enables them to take small slices of lots of interactions within their market. They have become a “business network”. More on that later.First, lets look at who the players are in this market:

  1. Start-ups. Most Vertical Search start-ups (at least the visible ones) operate in broad, large scale markets that are primarily consumer-centric. Many of these have been covered before in Read Write Web, specifically this post in September ‘06. The big sectors are travel, finance, consumer electronics and health. These markets are the big low hanging fruit that pass the addressable market size hurdle set by Venture Capital; this also makes them the most vulnerable to Google.
  2. Traditional B2B and specialty enthusiast publishers. This market comprises lots of small niches, micro niches if you like; think “Waste Haulage” or “Pig and Poultry” (those are real titles of B2B magazines). Think Trainspotters Monthly as an example of a specialty enthusiast publisher (I made that one up). This is the long tail of business publishing. Traditional B2B publishers have two big advantages. First, their long established brand (and the trust that it represents) and second, decades of domain expertise. They also have two big hurdles. The first is that they have traditionally been weak on technology; that is relatively easy to fix by partnering with technology vendors. The second hurdle is that Vertical Search is one of many lines of business (and it is currently a small % of their total revenues) and their core lines of business are under threat; Vertical Search therefore does not usually get the senior management attention and investment capital that is needed to win.
  3. The mass market online players (Google, Yahoo, Microsoft). They are grabbing the large low hanging niche markets and butting heads with those VC funded Vertical Search start-ups; but there is no Google team going after the Pig and Poultry market (AFAIK).
  4. Vendors including:

      * Technology. This includes search engines, scrapers, tagging tools, XML databases, research workflow, subscriptions/billing and more.

        * Data research services. Many sites rely entirely on aggregating, filtering and displaying open content from other sites. Other sites like to add proprietary data that needs to be collected by traditional market research techniques.

        There are also a few small niche consulting firms providing strategic advice, usually through some mix of research reports, seminars and workshops. Financing is often provided or arranged through a number of specialist “boutique” investment bankers and Private Equity investors specializing in this market.Vertical Search sites typically provide one of three main types of data:

        1. Lead Generation (contact details with enough data to target effectively for selling or recruiting). The value proposition is obvious. The danger is that these databases simply become tools for spammers. Every publisher does “list rental”, which means providing the list of subscribers to marketers; that can easily degrade into spamming. One intelligent approach is something like LinkedIn, where the person has control over what data is provided. Another intelligent approach is to provide enough depth of data to allow highly targeted marketing. One example is Computing Market Intelligence in the UK. People usually don’t mind a pitch if it is highly relevant and targeted to their current situation.
        2. Pricing Guides in opaque markets. The best examples create lock-in by becoming the authoritative source for the industry. There is plenty of news, comment and analysis in most markets that are free to the audience (i.e. paid by advertising); but it is still difficult to get hard data that answers the question “how much would it cost me?” There are many markets where pricing is quite transparent; however there are many other markets where there is major information asymmetry that is maintained deliberately by vendors. A service that opens up these markets will save money for buyers; B2B buyers are willing to pay for that.
        3. Financial market information for investors. The base data about stocks and bonds is increasingly free and available on multiple sites. What is hugely valuable is unique insight backed by hard facts that are not commonly available (and that do not breach insider trading rules). One example is tracking a new hot consumer electronics product, figuring out early how the product is shifting at a retail level, working out which publicly traded companies make components for that product; that leads to the right mix of legal, proprietary and timely data that institutional investors will pay big $$$ to subscribe to.

        There are also three basic techniques:

        1. Search engine on top of unstructured data. Google Custom Search engine serves this purpose and sets the bar for competing engines. Many Blogs use Eurekster and traditional B2B publishers often use Convera.
        2. Apply structure to unstructured data. Structured data (fundamentally data in a relational database) enables more effective parametric searching. Structure can be added through manual or automated techniques such as tagging and scraping; increasingly it is the combination of automation and manual methods that wins.
        3. Proprietary new data. This is the world of traditional market research, with a focus on facts and numbers not news/commentary/analysis. Proprietary data is the yeast that makes the aggregated non-proprietary dough rise. The cost to acquire this “yeast” is therefore critical.

        Current monetization strategies are usually either advertising or subscriptions.If you can show ROI by saving money you can charge subscriptions. Subscriptions enable predictable, high margin businesses; subscriptions are also much more recession proof than advertising. The basic rules for getting people to pay for data are:

        1. It would cost them way more than the subscription cost to research this data for themselves. In other words, they cannot easily use Google to find a free equivalent or call 2 or 3 buddies.
        2. The data can be used to make a case, either internally for a budget or externally for new capital, to spend some serious money. This means that the data must be verifiable and the source must be trusted and can be referenced.

        There are lots of pricing options for paid subscriptions – by registered user, by concurrent user, by # of records, by time period, by product (print vs online). Print On Demand technology has rejuvenated print by making very small print runs cost effective.In recent years the trend has been to move increasingly towards making information free, letting advertisers pay the bills. Google has clearly changed the advertising landscape with Cost Per Click, as this is closer to a model with proven Return On Investment (ROI).The future may move to transaction models. There will always be a big place for brand advertising (aka “faith based advertising”) but the logic of ROI tends to drive from Cost Per Click to Cost Per Action to Cost Per Product (i.e. a % of the product price). Given a choice, most sellers prefer to simply pay a % of the sell price; in an ideal business world all costs are variable costs.These transactional business networks rely on two key planks:

        1. Data. If all the market participants are accessing a common set of data that is trusted and verifiable they are more likely to trade with each other through the network. So a vertical search database is the key enabler for transactions.
        2. Trust. This comes from having these niche communities bound together through content sharing and ratings. This assumes a registered user community using real names.

        Google is not the only steamroller heading this way. LinkedIn has the other piece and they are on a roll. The strategic imperative for media firms is to use the base techniques of search and social networking to build their own value space. The technical pieces for both are easy to assemble. It is simply a land grab game.Business Networks work best in industries with lots of buyers and lots of sellers. This is known as a “fat butterfly” industry structure. If a few buyers or a few sellers dominate the industry, there is less opportunity for an information intermediary. This may sound familiar to people who remember B2B 1.0 circa 1999 with over-hyped companies such as VerticalNet. The end goals are the same in B2B 2.0, but a lot has changed since then:

        • B2B users are now online way more than they were in 1999.
        • Vertical Search is better at creating the common pools of trusted data.
        • Social networks and ratings have accustomed people to make judgments about who they might be interacting with online.
        • The cost to develop and deploy these types of systems have come down dramatically.
        • Entrepreneurs have figured out how to enable and co-opt intermediaries in the supply chain; rather than naively assuming that they serve no purpose and will quietly disappear.

        Vertical Search companies that rely on a few jazzy features or a “we try harder” approach will have their competitive advantage inexorably eroded by Google. Companies that use search as one tool to build niche online business networks based on a transactional model can create sustainable competitive advantage.

        eXpresso takes the enterprise route to web office January 29, 2008

        Posted by bernardlunn in Uncategorized.
        add a comment
        I noticed eXpresso when they bought a little start-up called Xcellery that had a neat way to collaborate using Excel. I had used the product and it worked pretty well.

        eXpresso is different from the web office contenders that you normally hear about on RWW for three reasons:

        1. The company originated as an enterprise systems vendor that happened to use collaboration around Excel to deliver solutions. So they arrived at the market by way of solving real world problems as opposed to companies that look at how much money Microsoft makes with Office and saying “I want a piece of that”.
        2. They focus on Excel. They don’t try to replace Word, or Powerpoint or Outlook. The Office bundle was based on the logic of the OEM market for PC vendors. There is no reason to have a bundle today. A great spreadsheet can stand alone quite easily.
        3. They extend Excel rather than trying to replace it. That works in the real world. Apart from early adopters with relatively simple spreadsheets, people are comfortable with Excel. It works well, it is extensible, it is easy to use. In an enterprise setting the cost of Excel compared to the functionality is a complete non-issue. But Excel’s native collaboration tools are weak and emailing versions around gets to be a real pain when you get to 3 or more contributors; that introduces errors in what can be mission critical applications.

        eXpresso say they want to replace Powerpoint as well. I think that’s a mistake. I would urge them to go long and deep on the Excel front. There is little logical connection or synergy between Excel and Powerpoint. The whole Powerpoint paradigm is flawed and will get replaced by Screencast and YouTube type presentations, in my opinion.

        Excel really is a unique product. It could be a winner for a long time to come with some simple updating. Microsoft currently seems to want to push integration through SharePoint. Thats not want users want. The IT guys may like that and tell Microsoft thats the right way to go but Microsoft did best when the listened to users more than to IT.

        Users don’t care a hoot about SharePoint but if you try taking away Excel you will hear muttering about “over my dead body”. So many start-ups have tried to create a programming environment that ordinary business people can use and nobody has succeeded. Excel delivers on this promise and has for a long time. A novice can be productive in minutes. An expert can find endless ways to add sophistication and capability.

        Collaboration really is the missing piece. I used Xcellery first for a simple sales forecasting application where 4 people had to enter forecasts. It worked a treat and there was no set up or re-training required. Millions of people have the same need.

        Collaboration is also easy with Excel as there is an obvious unit for the locking – a cell. This is much harder with Word and Powerpoint.

        Two real world stories illustrate the power of Excel. One was the CEO of an outsourcing company who said that one of the first things they looked for was a high level of proficiency in Excel. Without that it was, pass. The other was a comment by a vendor selling high end financial trading systems about Excel replacing the browser. The IT guys in the room were aghast with horror at the idea. The users were “yep that’s right”. Excel has handled real time updates pretty well for traders for over a decade; long before Ajax.

        twittering about twitter January 29, 2008

        Posted by bernardlunn in Uncategorized.
        add a comment

        thought twitter was lame

        thought i ought to be more open minded

        resolved to try twitter

        20 other resolutions got priority

        forgot about twitter

        saw twitter post about myself

        thanks marshall

        narcissism rules

        twittered

        watched people twitter

        all week long on my cell phone

        google “twitter is lame” and you get 2 pages

        reading thomas hardy book

        dropped hardy book

        he is too prolix

        oh no i have a.d.d

        give me some ritalin

        now

        maybe a.d.d is a competitive advantage

        the future wants to invite you

        don’t be such a twit

        what would ee cummings do with twitter?

        SWITCH OFF CAPS LOCK

        seesmic sees me?

        oh mon dieu!

        i want to know what 6 billion people are thinking

        and i want it now

        well maybe 6

        how about 1?

        call her

        now

        started long blog post about twitter

        deleted long blog post about twttr

        mdium is message

        massage

        now thats a good idea

        let the cat out

        cats r cute

        yawn

        Viral + Monetizable = StartUp Magic Quadrant January 21, 2008

        Posted by bernardlunn in Web 3.0 Semantic, social networks, start-ups.
        add a comment
        Hotmail is credited with inventing online viral marketing. I am sure there were earlier examples, but the whole point of viral is that it is so infectious that it obliterates memory of earlier attempts and they certainly did that. I was an early Hotmail user. It was just so simple, obvious and useful. Then all my emails had free advertising for Hotmail. (No, I don’t use it any more, it has been left behind by better alternatives).

        Most of the Web 2.0 success stories have been viral. Apart from Hotmail, this was not true in Web 1.0. The game at that time (hopelessly flawed in retrospect) was raising tons of money to advertise (online and offline) to get traffic. Flickr, YouTube, Twitter, Facebook type services don’t need to advertise to get mass scale.

        Then use Amazon Web Services or equivalent for your infrastructure; you don’t even need to raise money for capital expenditures, you just pay as you go as a variable cost.

        That is a major revolution in business; it overturns the accepted wisdom that you needed mega millions to build a consumer brand.

        But there is a problem with all this. I don’t believe I ever clicked on a Hotmail ad. I don’t think I even looked at them. Here is the nasty law:

        “The more viral it is, the harder it is to monetize”.

        These are the three streams of the Internet – communication, entertainment and research. You go online to communicate, have fun or find some information. The viral properties of each stream are quite different:

        1. Communication. This is perfect for viral. Think Hotmail, Twitter, Skype and Facebook. The viral property is built into the service.
        2. Entertainment. Think You Tube or Second Life or any online game. I tell people about a really entertaining video. The viral property is weaker as it depends on a stream of loss-leader hits.
        3. Research. Google got viral adoption because the alternatives were weak and it was a major problem faced by millions every day. As it was free and dead simple to use, there was no barrier to adoption; but the viral spread was only possible because it was such a big problem and it was so much better than the alternatives. This happens very rarely.

        Only Communication is a sure fire viral success. It only works when it is a genuinely new form of communication (webmail, social networking, microblogging). You cannot launch a web mail or social networking site today and expect viral adoption. But when it is a genuinely new form of communication, the viral adoption is stunning in its speed. I can see one new form of communication out there today that could get mass viral adoption, which is video conversations as exemplified by Seesmic.

        When you look at the three streams of Internet services – Communication, Entertainment and Research from a point of view of monetization, the order is reversed:

        Research is simple to monetize. It leads to a database of intentions and that leads to any number of advertising models that have a) proven returns to advertisers b) proven use for searchers.

        Entertaiment can be monetized through advertising and “freemium”. We are used to the idea of ads to get free entertainment on TV/Radio and used to paying to go to the Movies or rent DVDs.

        Communication is really tough to monetize through advertising. It has to be free to be viral (and it can be free because the marginal cost is close to zero). So the only way to make money is some form of advertising. It is just really, really hard to find a good way to offer advertising around a communication service that works for both the user of the service and the advertiser.

        The big debate about Facebook’s value and their Beacon debacle is a reflection of this fundamental problem. So is the online debate about Twitter monetization and the heat that eBay got for not being able to wring the expected profits from Skype.

        There have even been attempts to offer free telephone services in return for listening to ads. They were ridiculed and have failed. Yet we assume that it is OK to do this with online communication.

        The simple fact is that when we are communicating all our attention is on communicating, so ads don’t get our attention. Entertainment can have breaks; TV has accustomed us to this idea. But try saying “we interrupt your attempt to get a date to give you this message from our sponsor”. I think the sponsor would suffer some serious brand damage!

        Facebook is trying hard with some new models to monetize the social graph. But they all hit a fundamental problem. On page 44 of “Wikinomics, How Mass Collaboration changes Everything” it says “relationships are the one thing that you cannot commoditize”. That is like the law of supply and demand, you can count on it and take it to the bank. So any attempt at making social network relationships into either an Amway scheme (I make money by selling stuff to my friends) or a Beacon scheme (Facebook makes money by me selling stuff to my friends) will ultimately fail.

        This does not mean that you cannot make good money on a new form of online communication. If you have a new form of communication and you get mass scale virally, you will get good returns on capital. Even if ad monetization rates are very weak, you make up for low rates with scale. As it costs so little to get that scale, it is still an OK business. Somebody who needs scaled-up features to add to their platform will pay good money to acquire you.

        However that is a small prize compared to a Research service that gets mass adoption.  Google is valued at over $200 billion because they got viral adoption for a Research service. They have even found a way to make email advertising effective. I now use Gmail and the ads are often bang on target and I have clicked on ads in Gmail. (They are also often totally, crazily wrong; my favorite was when I was writing about somebody called Cooper and got ads for Mini Cooper cars).

        Services can mix Communication, Entertainment and Research. However the core proposition has to be clear. A new Communication medium is initially always Entertaining just because finding new ways to connect with people is a buzz. But once that “gee whiz” early adopter fades, the service has to be useful on a daily basis for mass markets. New entertainment models have to be social to break into what is already a hugely powerful entertainment industry.

        Research is currently solitary. It is not fundamentally entertaining. I don’t see fun as a driver for Research beyond a gee whiz phase. However collaborative research, search with a communication angle, does look like the next big thing. My definition of Web 3.0 is:

        “The combination of Web 2.0 mass collaboration with structured databases”.

        If you can build a research tool that propagates virally and gets more useful with each person who uses it, you build a business with phenomenal power. That is a lot easier said than done. The purely technical challenges of creating structure out of lots of unstructured input is considerable. Much tougher is the chicken and egg problem; the tool has to be useful out of the gate, which is tough if the use derives primarily from the interaction of many people.

        This means that funding has to be substantial to build enough value before the community kicks in to create value. That is why services such as Mahalo and Freebase raise VC measured in tens of millions of $. This is not like a pure Communication service that can get viral adoption out of the gate (but where the eventual returns are limited).