« September 2006 | Main | November 2006 »
October 26, 2006
The Branding Benefits from Pay-Per-Performance
Very interesting article in this week’s AdWeek about “advertisers… taking a broader view of search, buying terms they want consumers to associate with their brands, even though the searchers clearly aren't hunting for their products.”
It cites an example, “Honda has bought thousands of such keywords as part of a new campaign for its CRV. While it continues to buy auto terms as well, the carmaker is acquiring keywords related to its new "Crave" theme, like "chocolate," "banana splits" and "celebrity gossip," all designed to bring searchers to a community Web site Honda created for users to collect and share craves.”
This campaign is indeed a salient example of traditional brand advertisers experimenting with online media buys that have traditionally been in the pay-per-performance realm. I believe it gives a nod to the implicit notion that advertisers do receive “branding benefits” from online pay-per-performance campaigns, whether it is intentional (as in this case) or not.
Many online retailers that heavily use affiliate programs have known this fact for years. Surely Amazon.com receives more than just the direct gain from the traffic by having is products and links featured throughout the web on its affiliate sites. I’ve been in more than one conversation recently where someone has argued to me that affiliates (not just Amazon’s, but generally-speaking) are consistently and systematically undercompensating affiliates because they only measure direct sales generated by the traffic for the retailers, not the residual traffic that it generates in stores or the increased branding effect inducing purchases long after the cookie is expired.
In reality, though, branding benefits from pay-per-performance largely depend on the brand doing the advertising itself – I would maintain that this traffic has a greater impact if it is reinforcing and strengthening an existing well known brand (like Honda in the above example) than it does necessarily creating brand awareness for a lesser-known one. I suspect it’s the products’ brands, not the merchandisers’ brands, which will continue to experiment more in this realm, as they are the ones who are losing their pull from disruption in their traditional brand-building outlets.
The difficulty, of course, is the perennial ability to accurately measure the effectiveness of brand advertising. The AdWeek article doesn’t avoid the issue, “While brand-building search campaigns are not held to strict sales metrics, the ability to track results helps quantify their effect.” The article claims 50K visited the site in a month, while a comment on Search Engine Watch from the agency responsible for the campaign claims that the campaign has generated “nearly ½ of a million visits to the crave.honda.com.” As of this morning, however, there are only 348 user-submitted “craves” – the action the site is enticing consumers to take to further engage with the brand. Is that success? Is less than one-hundredth of one percent of people interacting in the intended way good? As pay-per-performance and branding campaigns begin to converge, these are the questions that advertisers wrestle with. Regardless, these advertisers who are on the fringe experimenting with new outlets will reap the benefits from them, whether they are easily quantified or not.
October 9, 2006
Information Dissemination and Feature Parity (or "Bells and Whistles Don’t Win the Day")
Last Friday, I was meeting with a set of entrepreneur founders who are very in their process of developing their consumer web application. And when the subject came up about features they intend to implement, they indicated that “of course” they will offer web badges for users to paste on their MySpace and other social network pages. What struck me was the tone of the comment – it was assumed that this feature would be implemented and I completely concurred – and that just around six months ago we likely wouldn’t have even considered it. Web badges and widgets aren’t completely new, but in 2006 they’ve suddenly come to the forefront of requisite features for many consumer facing apps that interact with social networks.
I’ve been spending a lot of time thinking about the implementation of and the implications of their use for a couple months (see my recent previous post on badge proliferation). In fact, I would argue that widgets/badges could become the platform mechanism which will interconnect currently disconnected and disjointed social networks. Of course it’s obvious that the YouTube player connects the two social networks of YouTube and MySpace, but with the increasing appearance of vertical social networks, the common thread that provides the glue among many silo’ed communities could be the displayed widget/badge.
However, the purpose of this post is to examine the case study of information dissemination throughout the blogosphere using MySpace badges as an example. Check out this Technorati graph which traces the number of mentions of both “badge” and “myspace” in blog posts over the past year:
The idea went from barely mentioned to commonplace in just a few months, and this illustration shows that the race to feature parity by web startups is hastened by the rapid dissemination of information through the blogosphere. How can a company compete when any demonstration of success will soon be replicated by other services? How can a company take an original innovation and run with it? It’s interesting to note that I know of numerous startups which are currently operating in semi-stealth mode to partially avoid this problem – they are completely open to the public and new users about their service, but they aren’t “screaming in the blogosphere about what they’re doing” to give them an information edge. A likely helpful strategy, yes, but certainly a temporary and fleeting stopgap.
Features – even ones that are difficult to implement – are not barriers to entry. Instead, innovation around marketing (communications and positioning, not just tactical moves), business development relationships, and instillation of network effects are and will increasingly differentiate winning web startups from the also-rans.
When I look at a consumer-facing startup, I assume feature parity of competitors within six months if not sooner. What can be accomplished in that time-period which creates lasting value beyond the next feature? What relationships are being forged with partners which aren’t easily replicated? How is the company targeting a specific demographic with understanding and expertise? How quickly will the service obtain a significant network effect which soon escalates?
Near-time feature parity benefits the web user community as a whole, as it quickly brings the best new inventive features to fruition everywhere. However, it brings difficult challenges to web entrepreneurs looking to make a mark with an innovative service.
October 2, 2006
Only A Million Dollars?
There’s a lot buzz today (here here) about Netflix’s offer to pay one million dollars to whoever can improve the accuracy of predictions of its movie-recommendation system. And it looks like there are number of people eager to throw their hat in the ring.
To this I ask: the prize is only a million dollars?
Consider that this contest is coming from a company with $830M in annual revenue (ttm), $66M in net income, and $1.6B in market capitalization. It obviously has already had the means to employ a number of distinguished people working on the existing system (a few of whom are now the judges of the contest) and surely has spent in aggregate well over $1M to develop their current technology. In addition, with a number of personalization & recommendation technology startups in the market at various stages of development, there are plenty of companies out there to acquire that would presumably command (or at least desire) a higher price.
Yet the company chose to go the route of a contest, making strides in pioneering “prize outsourcing research and development.”
Obviously, Netflix has all of the leverage here with their distribution and existing customers to which it could apply any technology. So of course it can and has offered whatever bounty it wants (which I’d argue is largely for PR purposes).
In reality, what this move does is call into question the viability of startups out there working on personalization and recommendation systems. Without the leverage of a huge market presence (read: customers) that a Netflix has, I wonder how they are going to be able to adequately monetize their offering. Surely these systems will lead to an increase in media purchases, and I’ve long been an advocate for the power of personalized predictive media. However, there is a significant distinction between being able to create value and the ability to capture it, and the power of distribution in this case appears to overwhelm. If you’re a genius who can compete “with 15 years of really smart people banging away at the problem” and it’s only worth $1M to you, then what does it say for everyone else – individuals and companies - working on that very same problem?
It Usually Doesn't Work the First Time
In the many startups that I’ve worked in/with or got to know in my experiences, there is one thing that almost always rings true: the initial idea and incarnation of the company isn’t the one that results in the end. In other words, whatever the early notion of the business, it just doesn’t work as planned.
Instead, startups after they’re first formed go through a stage of shifting and weaving, as they find the road which is the right path. Customers react differently than anticipated, revenue streams morph as the value proposition becomes defined, and costs (in time and money) of development and product vary.
Early stage startups are about experimentation. Good entrepreneurs know that.
But do the other constituents of a startup (employees, investors, advisors, customers, etc.) know it? Entrepreneurs should set appropriate expectations with others about how much experimentation is needed given their current stage. This process is difficult given that it must be weighed with confidence in the current plan.
When the experimentation works, it becomes innovation. When it doesn’t, the experimentation can become frustration. But those early frustrating experiences provide learning and market information which couldn’t have been gathered via any other means. Without the experience of small failures, a startup will not experience big successes.
Take a straw poll for yourself, and ask a successful entrepreneur if the business he set out to create is the one today s/he finds him/herself in charge of (or recently exited from). Somehow along the way s/he kept everyone involved excited about the endeavor during the riskiest stage of the company – the beginning.