There is actually a really good article over at news.com over the usability testing travails of hotmail 2.0. . I can’t tell you how many times I’ve been in exactly the same situation . . . trying balance innovation, strategy, and immediate user feedback.

The program was too slow to load, too different and, well, just not like the old Hotmail it was intended to replace.

It was a painful realization for the more than 100 managers and developers on the project. In banking on a snazzy Web 2.0 application to try to catch up to rivals Yahoo and Google, Microsoft had dramatically overshot its audience.

Classic mode wasn’t the only bitter pill the development team had to swallow. Even in the full version, it turned out that many customers still wanted to select messages using check boxes rather than a mouse click or keyboard shortcut, much to the dismay of Microsoft’s programmers.

“They were digging in their heels,” Sim said.

In the end, users hates any types of change (good or bad) . . . they will eventually adopt changes that are clear improvements, but that will still take a while. . . there is no such thing as a magic switch in the first place.

People that believe being “customer focused” is the ONLY goal of any product manager doesnt really understand innovation and strategy (ie innovator’s dilemma). On the other hand, product managers that are married to pushing the envelope on innovation and un-reasonably attached to their frameworks (information architecture, product strategy etc) really belong on Sand Hill road instead. There is a middle ground but its a case by case basis . . . knowing when to turn back (like the hotmail team) is a good first step. For all of MSFT’s faults, I’m actually quite impressed by this story.

Another random note, M&A teams of major companies are usually completely clueless on stuff like this. They believe that 1+1 = 3 while in reality 1+1 = 1.5. For example, taking a popular rumor over the weekend, if Microsoft and Yahoo merged what would be the combined userbase of their instant messaging product in 12 month if they tried to merge the two clients into one product? It would for sure be less than the combined marketshare of the two clients currently. The disastrous Sprint + Nextel merger was another perfect case study . . .

The blogosphere likes to report rumors as facts and edit post title accodingly in real-time as information changes. Om has some coverage (with the required hedging) while techcrunch is not hedging at all.

I have some random thoughts, not very coherent at all, but all pertaining to the transaction.

First of all, I love stumbleupon. I installed it about 4 years ago when I read about it in PC Mag, and I sincerely believe “browsing” is beginning to take shares away from search (more here and here) as a discovery methodology. So I say this is a smart move.

However, I did hear from my Yahoo buddies that they took a look at StumbleUpon (for an acquisition) and found that the active user base to be much lower than downloads. (sour grapes? ) For whatever its worth, I would think 3-6M users would be a $xxxM transaction but the rumored price is a little bit lower.

I also talked to some smaller websites and they have been telling me that StumbleUpon has been increasingly becoming a major source of traffic. (biggest ramp was 6 month ago, and now its a little flat). However, the traffic themselves are not very sticky (just stumbling through).

As Om mentioned, eBay beat out Google for the the acquisition. I wasnt surprised at all. Josh K. of First Round Capital is an eBay alumni (Half.com founder) and the driving force behind StumbleUpon’s board. David Feller (co-worker and friend from eBay) is StumbleUpon’s VP of Marketing and is also an eBay & Half alumni. Josh and Dave knows eBay inside out and have access to the very top of the company as well as the very buttom. Transaction among friends are usually more honest and straight forward . . . who wants to deal with Google’s cocksure M&A team? and un-certain integration strategy?

Ok so lastly, what can eBay do with StumbleUpon . . .

- eBay auction toolbar is one of the stickiest applications eBay has. This could potentially give the functionalities in the eBay toolbar some distribution on the SU toolbar. (wanna stumble on some auctions?)

- Phishing is THE reason user trust on eBay and PayPal is so hard to improve. Paypal and eBay are working overtime on creating industry working groups and blacklist databases to solve this problem. SU is the missing link to the end user to allow for distribution of the warning systems to the browser.

- Skype + SU toolbar are ways with which Paypal can creat a persistent wallet on the “browser-top” which would be significantly more convinient than hunting down the wallet. (as always I believe Paypal is the main benefitiary of most eBay acquistions)

- eBay can exert more leverage in the search game . Firefox generates $100M+ in revenue a year simply because google is the default homepage. Having the search box on the toolbar makes eBay in control of the user relationship while making search engines depend on it (for once!) for traffic. (there is some HBS case study on this. . . I think its called the “judo strategy”)

- I wonder how the Skype ToolBar is doing. Maybe its doing really well and StumbleUpon would be a good source of distribution for its functionalities too.

- I’m unsure what eBay is planning to do with SU’s small but growing advertising inventory. But maybe eBay does have grand ambitions on running its own ad network.

Google’s acquisition of DoubleClick for $3+ Billion was a homecoming of sorts.

DoubleClick was the original ad network. During the dot com boom, they were the pride of silicon alley and king of web advertising (outside of Yahoo! which was captive at the time).

Yet something happened between 2001-2005 . . . . DoubleClick lost it mojo. . . and its marketcap went from $15 BILLION to around $500M. And finally went private for around $1B in 2005.

Yes, $3B is an ungodly sum, but DoubleClick could have been Google; or atleast Google Adsense. But it didnt, so what happened?

Adsense kicked its ass . . . thats what. . .

DoubleClick was way too close to Madison avenue (physically and figuratively). It looked at the online advertising market with the same mindset as offline (TV, Print) advertising.

- It relied way too much on the sales force, deciding against investing in a self serve model which prevented it from scaling. Even more importantly, while the model worked in 1999 when traffic was centered around major portals and destination, as the web “long tail” proliferated, doubleclick lost a huge portion of its traffic monopoly as they were not able to serve that segment through its sales force.

-Just as important. DoubleClick looked at web targeting on purely offline terms. HH income, age, ethnicity, etc. Google, on the other hand, looked at it from an algorithmic angle and figured out that RELEVANCY was the variable which loaded the highest for click through rather than all the traditional targeting attribute. And ofcourse, relevancy and content are intricably linked. (and thus text).

MOST IMPORTANTLY. DoubleClick made a huge strategic blunder. They saw their adserving technology as something to be sold modularly rather than bundled with their ad network. They became a provider of commoditized technology and lost its ability to insert itself between the buyer and seller of advertising. It no longer controlled pricing and placement of ads. It quickly failed to become a NETWORK (with network effects) but became an ASP. There was no network based relationship . . . simply a licensing contract.

Adsense was/is brilliant in this regard. Its a black box; buyers really has very little say (until recently when Adsense has scaled significantly already) on where to place their advertising. Publishers on the other hand, has very little say in their rate card . . . EVEN MORE SO . . . both sides simply allowed Google to determine Adsense’s cut as it wishes. So long as both sides get what they want (clicks & $), the network effects kept both party from dis-intermediating Adsense AND switching to a different provider. People complain about the black box nature of the Adsense system and as a result, startups try to bring transparency to the ad network game. What they fail to realize is that the black box is the secret sauce; it gives Google the ability to increase optimization, monetization, and placement at its whim (usually for the better). The black box might seem draconian at first. . . but ultimately benefits both sides if all they care about is the all mighty dollar.

It looks like the secret sauce is going to get applied to DoubleClick soon . . .

For centuries (ok fine, atleast a decade) people (especially non-users) has been complaining bitterly about the way eBay looks (circa 1994) and works (pages of scrolling, no javascript what-so-ever). In the last 3 years as the web world went “2.0″ ga-ga the complaints has only gotten louder. One might even call eBay “Ghetto Fabulous” But the truth is that eBay grew and grew despite, and probably because, of all the quirks of the website. (case study #2, myspace).

Well, the launch of eBay Express last year gave hints to what is to come for eBay.com. And the recent (I just stumbled on it, but I think it launched last night on the 23rd) revamp of eBay Motors officially begins the march of eBay.com towards the so called web 2.0 era. Since the new interface is still in A/B testing (not everyone gets it when they login), you can get a quick education of whats new at eBay Motors 2.0 at eBay University. A lot of the new changes are born of technology first created for eBay Express and other initiatives. But the integration of all these features truly points to the fact that this is the LARGEST and most IMPORTANT revamp of eBay Motors in the history of the company.

Here are just some of the new changes:

1) Web 2.0 fonts (bigger), color (brighter), and layout (rounded edges) . . . oh ya. . . AJAX (I pray one day people would stop associating AJAX with web 2.0)

2) Moving up the life cycle of commerce from transaction consumation up stream into research (eBay used to be much more focused on the transaction it self rather than decision making)
- Product reviews are front and center
- additional licensed content from KBB, carfax and more

3) A product architecture & ontology
- listings are now grouped under a product concept (all acura TL’s are associated as the same model type)
- attributes are extrated from each listing
- cross model navigation - for example coupe, 2 doors, SUV

4) New search engine - obviously related to (3) and very much looking like eBay Express

5) New emphasis on local searches and e-commerce (this is a new market eBay did not participate in pro-actively untill recently)

As I always said, “better” doesnt always mean “better” . . . users are creatures of habit, and an “implied social contracts” drives the interaction between a company’s website and its users. For example, conversion rate goes up when Google simplifies its homepage but down when Amazon simplifies theirs; given huge overlap in userbased, there is no way to predict actual user behavior by blindly following any mantra (2.0 or not). And thus, the only way is to test repeatedly and slowly roll out any changes. Make no doubts about it, any small percentage change in conversion rate will impact the bottomline. This revamp is a neccessary and inevitable for eBay (in search of growth and fending off new competition), but it is a gamble nonetheless . . .

If all goes well (it should as the new site is amazing, my opinion) looks for rest of eBay to move in this direction in the next 12-24 month.

Not now . . . but it will get there. Last week, Skype launched the long rumored/ expected/ in development Skype Prime service(via TechCrunch). Using Skype Prime, users can charge to recieve a call.

Even before the skype acquisition, (aug 2005) I recognized the impact of Skype on payments, but this development was headed off by the acquisition (Paypal was a huge benfitiary of the acquisition eventhough Wall Street did not recognize it). This development, however, could potentially be more revolutionary.

The key to the why Skype Prime is so much more than a “1-900″ service. It is the product strategy of useing Skype Credits as the payment mechanism:

your Skype Credit is deducted by the appropriate amount that then goes to the receiver’s account. The provider does not get the call fees directly as Skype Credit — rather, they go into a special holding “box”. The provider then receives the revenue via PayPal.

As is, skype credits are associated directly to the funding currency and thus are automatically converted into real money for the reciever. However for cross currency conversions, (caller bought skype credits in US$ and reciver is charging Euros) it does not appear that this transaction could be consumated.

Because skype credits are essentially pegged to a local currency AND there is no cross border transactions, Skype credits are only fungible as part of a transaction. However if Skype begins to enable cross currency calls, and begin denominating Skype Credits in MINUTES, a liquid market for trading these minutes will likely occur. Skype will have essentially created a virtual currency/economy. (I pay you in skype minute and you charge using skype minutes, and convert them to local currency in an as needed basis)

Of course, for a Skype economy to appear, there need to be a market for skype minutes. There are one very obvious applications: Sex. And second even more brilliant suggestion by Tom Evslin - as a method to self-monetize leads.

Here’s my prediction: many people will set up Skype Prime based call services. They’ll put information in their profiles which attracts telemarketers. Telemarketers will learn who the best prospects are both from the profiles (some of which will be lying) and by accumulating lists. You’ll adjust the rate so that you’re pleased, not annoyed when you get a telemarketing call.

Of course, given the rise (and probably proliferation of virtual economies like 2nd life) , Skype Minutes could easily become the currecy of choice for these smaller vitual worlds. (and as any gamer will tell you, Skype is the best thing (IM + Voice) to ever happen to a raiding party.

(My initial title for this post was “Yahoo Panama Working?” I just changed it to “YHOO is a SCREAMING BUY!” after I did some quick number crunching.)

Ran across this press release from Comscore which claims Yahoo! Sponsored Search Ads
Click-Through Rate increase 5% and to 9% in the last few weeks.

Using the week ending February 4, 2007 as a baseline for sponsored search click-through rates (i.e. total clicks on sponsored search ads divided by total searches) before the ranking model launched, comScore studied the two subsequent weeks of click-through data to evaluate the impact of the new ranking model. comScore’s data indicate that for each of the two weeks subsequent to the launch (ending February 11, 2007 and February 18, 2007), Yahoo! Sites experienced a noticeable lift in its sponsored search click-through rate. The week .ending February 11 saw a 5-percent increase, while the week ending February 18 showed a 9-percent jump.

Furthermore,

Another anticipated result of Yahoo!’s new ranking model is a shift in composition of total click volume from algorithmic to sponsored. The “sponsored click composition” metric (i.e. sponsored clicks as a percentage of total clicks) is critical in understanding Yahoo!’s success in improving both monetization and user experience. qSearch data show positive gains in this area, with sponsored clicks representing 10.6 percent and 11.1 percent of total click volume in the weeks ending February 11 and February 18, respectively. These data represent increases of 0.5 and 1.0 points in the weeks following the new ranking model launch.

So the good news first. . . it appears that CRT for sponsored search results almost doubled WITHOUT take significant share away from algorithmic/natural search results.

Let me do some simple math to confirm. . .

PANAMA Week I

.106 = sponsored clicks / click
+.05 = sponsored clicks / search query

+.05/.106 = Incremental 0.49 total clicks per query

PANAMA Week II

.111 = sponsored clicks / click
+.04 = sponsored clicks / search query (.09-.05=.4)

+.0.4/.111 -> Incremental .36 total clicks per query

So in two weeks, the clicks per query has increase by a total .85 clicks.

To give some background, total clicks per query can approximate to quality of the results (not globally especially when the # is large it might actually point to having bad results, but its safe to assume that we are at a local optimal since Yahoo does have a decent search engine). Futhermore, the rule of thumb is that in general search engines generate 2 clicks per query. To increase clicks per query by close to 50% is freaking amazing.

Even more importantly, I was afraid that sponsored ads was taking click throughs AWAY from organic results which would mean that the effect is short lived and that the numbers would point to LONG TERM dissatisfaction with Yahoo’s organic search algorithm.

BUT to my amazement, the total # of clicks per query has increased showing that overall clicks per query has increased significantly thus organic search clickthrough has maintained (or even improved) because of the change in sponsored search algorithm (aka Panama).

These are all great news for Yahoo. I’m gonna put some money in YHOO now. . .

(CAVEAT, I’ve NEVER made ANY money in stock investments so dont listen to my ramblings)

Been meaning to do this for a long time but Shri beat me to it. She set up a ebay blogger wiki directory where past and present ebay bloggers can list and discover each other’s blogs. I also replicated it on my side bar as a seperate blogroll so my readers can find them. In fact I encourage everyone to take a quick stroll. I’ll put the eBay lineup up against the best bloggers Yahoo, Google, or MSFT can throw together (plus many of their bloggers gets paid to blog. . . glorified PR people . . . we, instead, blog secretively between meetings, hoping not to get caught)

When I started working at eBay there were only a handfull of bloggers, now there are significantly more. . . (the wiki will prbably double in short order as the word spread) . . . I remember being scared shitless that PR might fire me or ask me to shut down my blog. In fact Shri (as my boss) was the first eBay person to ever discover it . . . and she was nice enough to keep it on the down low . . .

It is well known that the advertising industry has a high Beta compared to the general economy. In short, the industry does very well during good times and very bad during bad times. There are several reasons to this, when the economy does well 1) companies forgo ROI for marketshare 2) companies believes advertising is a capital expenditure to build brand equity 3) keeping up with the jones, as advertising effectiveness is extremely sensitive to competitive spend. It is not to say these reasons are not rational, they are somewhat. But that the industry tends to go in boom and bust cycles because of them.

Google’s is in this very cyclical industry . . . so much so that the offline equivalent of ad brokers (agencies) are mostly private shops or loose but public cooperatives/roll ups to even out the cyclicality. Furthermore, Google has one large issue. . . that like the Real Estate bubble in Hong Kong in the mid 90’s, much of the spending on Adwords has been done by speculators trying to arbitrage traffic/eyeballs rather than companies productively selling a service or product.

For the past months, I’ve been tracking (unscientifically) the percent of adword advertisers with the main business model of advertising. In short, instead of selling shoes or providing a service (like translation or loan brokerage), these sites are simply trying to pay for eyeball on Google and hoping to make more than they spend on getting that visit (or visitor if the site is sticky enough). Even worse, some of these companies are venture funded so they are simply looking at whether the cost of that visit will get them an equivalent increase in VALUATION. They dont mind losing money, as long as they can flip the company to someone else in the short term.

In decemenber I tracked about 200 queries and found that about 20% of adword advertisers are arbitragers. Early this week, that # has gone up to about 35%. Again this is about unscientific as it gets and I hope some research analyst will start tracking this metric sytematically.

But I think the point is made that increasingly the revenue growth google is able to capture are significantly more risky than before. Google’s revenue beta is becoming higher and Wall Street has yet to take that into consideration. Remember back in the dot com days, all the venture money raised was poured into buying Sun hardware. when the bubble collapsed, so did Sun. Today, because the lower cost of infrastructure, this money has been poured into advertising and marketing instead (specifically into Google with a disproportionate share). If (when?) the battle for traffic is over and the casualties emerge, a significant portion of Google’s advertiers (and corresponding revenue) will also go the way of a server spending on Solaris machines . . .

But like a game of chicken, advertising spend usually drop off the cliff rather than decelerate, so there will be little to no warning when it does happen.

Yesterday, Amazon launched endless.com (coverage at techcrunch and techbeat). Obstensibly its an obvious response to the $1B that Zappos.com is taking in a year (close to 35% of the total shoe market) . But even more strategically, its an increasing sign that the first generation internet giants has reached some sort maturity and limits to scale that traditional marketing techniques like brand extensions (ie, Coke -> Diet Coke) are beginning to become popular if not standard strategic growth options.

Furthermore, as much as the web 2.0 crowd (we? me?) would like to believe, mashups are not invented by the internet generation. Using internal assets and re-aligning them to create new products our services was invented when Henry Ford (finally) launched Model A.. And continutes today, the Chrysler Crossfire uses the same chasis as the Mercedes SLK, another example. (What is interesting is that mashups has moved form assembling internal assets to mixing external and internal assets through loose couplings and arms length relationships). For endless and ebayexpress, the main assets that are being leveraged are the inventory (both share a subset of inventory from the mother site).

I blogged about this a few month back and it appears to me that Amazon is hitting the same scalability issues.

For giants like eBay, Yahoo, Google, and even Facebook and Myspace (today). . . the MARKETING FACT OF LIFE - SEGMENTATION has slowly reduced the value of network effects. By definition, network effects is a mass market play. It is in opposition to the concept of niche marketing, niche product, for niche segments - ie the better you can target your product or service to a particular niche the more likely he or her will chose your product over a competing generic solution. These giants cant no longer band-aide new site wide features and functionalities hoping to attract new segments to their website as USAGE TIME, SCREEN REAL-ESTATE, USABILITY limits the feature creep. This admission that network effects is no longer the dominant driver of their business - that segmentation is - is widely seen but rarely discussed . .

It further amazes me how the mentalities of a startup can blind someone from making the right business decisions (better, faster, cheaper /= better for the customer). For some reason, the “why’s” of endless and ebayexpress were not obvious to the pundits of the tech world. . . but for someone that works at Gap, P&G, or a HomeDepot it must seem very obvious.

Every brand has its limits, even Amazon and ebay. Ofcourse, Amazon decided to completely remove itself from the endless brand while ebay tried an extension to target new segments of users.

One size fits all is only true for baseball caps (and even that is questionable). Beyond the brand, the flexibility to create custom search/discovery experience within a category can create a significantly better conversion rate AS WELL AS re-enforce the brand itself. In the physical world, the Apple Stores and its shopping experience are a clear extention of the Apple brand.

“Enhancements” are not always better. Ajax might be cool, you might love all the drag and drop functionality of the latest website profiled on TechCrunch. But you and I are a sample size of one. For incumbent companies, by definition through the size of its userbase, late adoptors are the largest segment. In the end, the feature needs to be rigorously tested with a large enough sample size to really know if “enhancements” are good for the bottom line. . . not just kinda cool.

Re-orgs can be fun. . . really they can be! . . . part II of the series.

1. China is the next big thing, talent level uneven, and everyone gets a title inflation. Plus, Baidu/Alibaba/QQ/Netease loves to hire people from their American competitors and give them a VP/CxO title.

2. Its about time you get a new powerpoint template that wasnt made by the founders 10 years ago

3. Cleaning cubes - finding cash hidden at back of drawers, throwing out useless decks, finding your long lost garage door opener . . .

4. Headhunter who call and forces you out of your inertia to explore opportunities

5. Rumors, rumors, rumors . . . talking about rumors, regurgiation rumors, even making up some of your own. What do you care? you have millions at the bank and on first name basis with David/Jerry/Larry/Sergei/Meg/Pierre.

6. Underground re-org betting pool - when, who, how

7. You love your boss, but its time to expand your sphere of influence. Re-orgs can be a good excuse to find a new position in a different part of the company without ruffling feathers.

8. Tides turn, you were “one of the ten” left in the last re-org, but this time, you all of sudden get your own team

9. Touching or funny farewell emails

10. New asses to kiss