The Software Abstractions Blog

Does Web Advertising Work?

Do Web Ads work? Does the vaunted Freemium model actually produce meaningful revenues for anyone but the largest services and content sites? How should Web 2.0 be monetized?

Many Perspectives

There has been a lot of discussion about these and related questions in the past few days among the Techmeme crowd. Here is a cross-section of perspectives from some of the thought leaders in this discussion:

• Unsurprisingly, this discussion was kicked off by Tim O'Reilly, the same ultra-smart observer, blogger and publisher, who originally coined (or at least, evangelized) the term "Web 2.0".
• The Financial Times ran an article provocatively titled: Web 2.0 fails to produce cash , in which the authors question whether popular Web 2.0 companies have actually produced any revenue, and predict a big shake-out in the next year or two.
• The always-interesting Robert Scoble led a widespread discussion, referring to posts by Om Malik and Dare Obsanjo, on whether Twitter (that highly popular micro-blogging service, and train-crash-currently-under-way) should be charging super-users like him, while continuing to keep the core service free for most casual users
• Alexander van Elsas has a wonderful post, referring to an earlier post by Scott Karp of Publishing 2.0, explaining how advertising does not usually provide the user with any value; van Elsas feels that the real value should be created outside the network.
• Coming up on May 30th is a discussion on this topic at TiECON East: "Web 2.0: Viable Business Model or Bursting Bubble?". The discussion includes two prolific and highly popular bloggers among its participants, both of whom have (inevitably) blogged about it - Fred Wilson (Discussing Web 2.0 ) and Don Dodge (Web 2.0 = Bubble 2.0? ).

When Ads Attack!

Clearly, web ads do work under certain circumstances. The poster child, of course, is Google - in addition to ads on its own search results page, Google commands revenues from millions of long-tail publishers via AdSense. Although most of the small publishers in this group actually make very little money, the total revenue in the aggregate for Google is very high.

Web Ads work better in certain situations than in others. At a VLAB event last week, Bill Reichert, Managing Director of Garage Technology Ventures, summed up perfectly the virtuous cycle of content and advertising:
    "If the paid content enhances the product, as in the case of Google, then that attracts more users to the site, which in turn enables more income, which allows the site to further enhance its content. Thus, the ad content should fit the users' actions to be effective; otherwise, as is the case with Facebook, the ads just get in the way."

Monetization Options

Once a web site or service gains traction and gets a sizeable following, there are potentially many creative models for monetization. In a much older post, Fred Wilson wrote about the The Long Tail Of Business Models , working off an initial post on Media Business Models by Chris Anderson (and I've added my take ).

Web 2.0 monetization works especially well in specific situations, and it's likely that totally new models will emerge in the future. One such model is advertisers who create meaningful content for web users, a by-product of which is to drive sales; this type of content can include informative articles, useful services or immersive experiences. Examples of this type of model are:

• Pharmaceutical companies that educate doctors on how to utilize new types of medical devices
• Lawyers who explain the law or provide status updates on government activities
• Home improvement stores that sponsor how-to seminars and then provide discounts on materials

Clearly, the web is a different medium, and simply transposing print ads to the web will not work; the big winners of the future will be those companies that can take some of these radical new models for monetization and make those work for their own use cases.

---

Two Emerging Memes: Freenomics and Crowdsourcing

Silicon Valley has always been fascinated by new technology memes, radical new ideas and business models that take off and capture the popular imagination - often changing the world of technology in the process. Valley memes quickly become the jargon of leading-edge VCs. For example, the phrase "it's a Freemium model, based on UGC and Wisdom of Crowds" requires no explanation in certain circles; it's as meaningful to these players as the phrase "it's a RESTful Web Service based on a 3-tier J2EE architecture" is to a Web Applications Architect.

(For the uninitiated: Freemium is a term popularized by VC Fred Wilson to mean a free service with an upsell to paid premium subscriptions; UGC stands for User-generated content, a relatively new concept in which the audience or readership helps to craft creative or original content; and Wisdom of Crowds is a term popularized by James Surowiecki's book of the same name.)

How does a technology meme start? When does an interesting new idea or concept tip over into a meme?

Disambiguation of Search Results? Yup, Google's got that

Just last week, in an email exchange with another search blogger, I wondered when Google would provide options for disambiguation of search results.

When you think about it, that's an obvious requirement for the Results page of any serious search engine. If I query for the search term "Java" - does it mean that I'm looking for results about the programming language, the coffee, or the island in Indonesia?

There's no way for the search engine to be able to tell, although personalization could provide clues. The easiest solution, as I wrote back in 2006, is for the search engine to just ask - which is why Wikipedia offers this page: Java (disambiguation) . Alternatively, the results can be grouped into various categories for the user to choose from, which is another way of doing the same thing.

Until now, Google has been mostly following a third option, which is to simply pick the most popular category regardless of the user's real preference; this can lead to some strange results, as highlighted in my earlier post on deconstructing real Google searches. But this approach doesn't really cut it, since it ignores all the unpopular search results - it's very possible that the long-tail searches can collectively make up a market share that rivals or exceeds the relatively few "popular" searches.

There has also been a limited amount of disambiguation offered by Google's "related searches" feature.

Well, no more. Google appears to be experimenting with offering disambiguation directly by grouping search results into categories. See the screen shot below, that shows Google search results for the query: "freebase" . Effectively, the results page seems to be asking: do you mean, the free semantic web database, or the other kind, associated with drugs? Or a third alternative: FreeBase - a free Windows software program to configure the Apple AirPort Base Station.

The use of horizontal ruled lines to separate the sections, is a nice touch!

Obviously this is some type of test; I certainly hope it's successful. I can't wait to see this feature become mainstream among the major search engines. It will be a big step forward in Search!

---

What is "Long Tail"? A guide for small business

If you're involved with online sales, you have almost certainly run across
the phrase "Long Tail" recently; if you haven't, you soon will. This is a
hot new topic - and something of a controversy - in the blogosphere,
especially among the influential bloggers who write about Internet
Marketing.
Proponents claim that this concept will completely change the way we
think about how to sell online, specifically in terms of product mix and
user targeting. Opponents dismiss these claims as exaggerated, and
dispute the data used to make the claims.

First proposed by Chris Anderson in a landmark article in Wired magazine
(he later followed up with a book), the Long Tail theory has been
generally accepted by the online community.
Basically, "Long Tail" is an economic concept, whose main propositions
are that
(a) the Internet has dramatically reduced the costs of distribution,
storage and retail sales of products; combined with real-time
information about buying trends and public opinion, it is now possible to
offer very low-volume items profitably to users, and
(b) the total sales revenue of these low-volume items ("misses") that
make up the Long Tail, taken together, could be larger than that of
the high-volume items ("hits"), that make up the other end of the
spectrum, the so-called High Head.
The rationale is that although the volume per item is very small for
the fringe items, the number of those items is very large, leading to a
large total volume. Wikipedia has an informative article about the long tail,
which includes a typical distribution graph that highlights the Long Tail
and High Head in different colors.

No one seriously doubts the first point. A tiny online retailer - with no
physical showrooms or retail space, centralized distribution warehouses
and on-demand manufacturing - can now manage a global reach with a
miniscule budget. The costs of making, storing, distributing and
shipping products are a fraction of what they used to be, so that virtual
inventory sizes can explode to hundreds or even thousands of times
what they used to be, without significant impact on cost. Marketing
costs are another matter - whether they are significantly lower is
arguable.

The second point - whether the total sales of the low-volume items
taken together can be greater than the total sales of the high-volume
items - is currently being hotly debated. Anderson says that it is
strongly supported by the data, but others disagree. The venerable
Wall Street Journal has jumped into the fray, with an article by
Lee Gomes that tries to debunk the Long Tail theory; paradoxically, this
has the effect of giving additional legitimacy and weight to the theory,
driving it even more into the mainstream. In true blogging spirit,
Anderson has responded with a touch of class, with better examples and
a clearer explanation, on his blog.

Other industry luminaries support Anderson's position. The esteemed
Tim O'Reilly suggests that data from O'Reilly's Safari Books Online service
and from Google Book Search   supports the Long Tail theory. Others,
such as LibraryThing and Zillow , poster children of Web 2.0, routinely
refer to the long tail in explaining their strategies.
Industry heavyweights like Google are also keen to leverage the long tail;
CEO Eric Schmidt says that Google is addressing both ends of the long tail
spectrum. Others, like Yahoo and Amazon are not far behind.

If this theory is correct, then it means that (i) you could build up your
small business pretty significantly by addressing the unmet need for
unpopular items which are not easily available from the mega-stores,
and (ii) if in fact, you are the mega-store, then by focusing on the hits
and not addressing the need for unpopular items, you could be missing
an opportunity with significant profit potential.
This is especially great news for small business, for whom the long tail
has traditionally beean an area of strength. In a specific niche,
it is possible for a small business to compete very effectively with a
much larger player, and win!

Given that the experts can't yet agree on the validity of the Long Tail
theory, though, does it matter?  Would it be better to wait on the
sidelines until the data is all sorted out and the academic debate is
resolved?

For the small business with an online presence, it is becoming clear
that the long tail does matter, and it could change the way you
approach online sales. Regardless of how exactly the numbers come
out - whether the total sales for the Long Tail are indeed greater than
those for the High Head, or vice versa - there is mounting evidence
that it is already having an impact. Apart from the obvious media
industries like books and movies, where the long tail effect is
well-documented, examples can be found in a wide variety of
industries, including maps , healthcare , Canadian music and even
poker .

An increased understanding of the Long Tail can help the small business
owner in several ways:
  - It allows you to leverage low-volume items to boost your overall sales
  - It allows you to compete more effectively with bigger players - the
      Amazons and the Wal-Marts - by focusing on specific niches
  - It has a significant effect on your SEO plans
  - Most important, it allows you to create separate strategies for the
      two ends of the spectrum, tailored for the hits and the misses

For now, the discussion and debate about Long Tail continues. For hard
data and long-term trends, we have to wait and see; empirical evidence
suggests that this theory is certainly true in certain market segments. If
it turns out to be pervasive across a majority of vertical markets, then it
will have very broad implications for all internet retail sites, especially
regarding decisions about product mix and target users.

Others have written about the effects of Long Tail and the ongoing
debate; e.g. here's Danny Sullivan's excellent article in Search Engine
Watch. The companion website, www.SoftwareAbstractions.com, has
a list of some of the better resources and examples about "Long Tail".
This list is regularly updated as the debate continues.

---