It's Rishi

Google surprised everyone this week with the unveiling of Google Latitude:

What’s most interesting about Latitude is thinking about why Google prioritized this feature over others for Google Maps. To kill off Loopt? I just don’t see that. After all, how much of a threat is Loopt to Google? Not much. Google Maps is the dominant mapping application on every mobile platform. Ultimately, tracking the location of your friends is a feature not an application. The idea that users were going to launch Loopt on their phone for the sole purpose of checking their friend’s location is just not realistic. It’s natural that this feature was eventually going to come to GMaps. But in the meantime, I doubt Google was losing sleep over Loopt.

So why now then? My guess is amassing more user tracking data. Note that when you enable Latitude on GMaps, Google now ties your Google login to your GMaps installation. This was a link that did not exist before. As a result of all this Google can now record data points such as:

1) Where you go, spend time at, etc.
2) Where your friends go, spend time at, etc.
3) What types of POI’s you search for
4) What routes you take

All very interesting (and very scary) primary data points. It’s not clear if Google can do anything meaningful with this data yet. However, it’s reasonable to expect that that in a year or two, once a lot of this data has been collected, either Google’s search or ad targeting technology will incorporate this info. However, if I am correct that Google is recording this data, they really should make the user aware of this. Of course you can argue that none of the ad networks, including Google, makes online users aware that their every click is being recorded. However, I think there’s likely to be many users who are concerned about their physical location being tracked!

Finally, Latitude is a fantastic way for Google to get users to formally “friend” their contacts. In order to use Latitude, you must have friends on Google’s social network. I, like the vast majority of Google users, have plenty of contacts (via GMail and GTalk) but never formally friended them. Latitude forces users to do this in order to use the service and should help Google build up its social network. This will pay dividends for Google as they continue to expand their social product offerings.

I must say, Google getting serious about mobile is exciting. It’s forcing all the players to step up their game. In fact, my hunch is that Android’s ability to run background tasks had something to do with Apple killing off the push notification system in favor of a more robust background task function for the iPhone. In the end, the consumer wins. I love it!

The answer: 50% of all Internet users! Hitwise released a report comparing the traffic of the four leading map products: MapQuest, Google Maps, Yahoo! Maps, and Microsoft’s Live Local. What didn’t surprise me is that Google Maps is the only one of product of the four that has gained significant year-over-year traffic. The others are flat to down slightly. However, even at GMaps’s brisk growth rate, it still won’t likely catch MapQuest for atleast 18 months.

I don’t know a single person that uses MapQuest. Even after a recent redesign, MapQuest is still a poor, uncompetitive offering. The only times I use MapQuest is when it is integrated into a third-party website (most often store locators). Even in such cases, I often copy+paste the address into GMaps. MapQuest is that horrible and I am shocked to learn that half of Internet users still choose it. Above all, this news serves as a great wake-up call that even a compelling product from a market leader still can take years to cross the chasm.

Back in July 2006, I wrote a post titled Predicting the Future With Google Trends in which I described how Google Trends, which measures the relative search volume of keywords on Google, could have been used to reveal the relative popularity of real-world phenomena, such as who will win American Idol.

Out of curiosity, I pulled up 30-day trailing data for both the republic and democratic presidential candidates and compared it to the actual Iowa Caucus results yesterday.

Is it just me or was Google Trends a remarkably accurate predictor of yesterday’s result! I tried specifying trend data for only Iowa however it seems as if there isn’t quite enough data to draw any meaningful conclusion.

Google introduces PPA

The big news of the week was that Google released a pay-per-action (PPA) advertising product. This has been hinted to for a while by Google. In fact, during their Q2 ‘06 conference call, I noted several references hinting to a future PPA product. Long-term, Google is definately going to increase advertiser’s confidence in Google’s ad networks with PPA. Most importantly, advertisers concerned about click fraud should be very happy about this. I also liked how Google will be ranking ads by profitability per click. It’s a simple and very clever strategy. The big problem with PPA is that you are relying on the advertiser to effectively convert clicks into actions. If Google deliver clicks to the advertiser but the advertiser does a poor job of converting those clicks, then Google would lose. However, by giving advertisers who deliver a higher rate of profitablity to Google, Google, in effect, forces advertisers to focus on conversion.

This focus on ad profitability will also encourage Google AdSense publishers to get on board with PPA ads. With traditional PPC ads, an AdSense publisher depends on Google to deliver relevant ads that will result in good click-thru ratio (CTR). However, with PPA ads, in order for the publisher to get paid, the publisher also depends on the advertiser to convert the click into action. That might be tough to swallow. Ranking ads by profitability should help AdSense publishers trust PPA ads more.

Reduction in the number of ads on search pages
Scoble posted an interesting tidbit about how he had been noticing that search result pages on Google have been showing fewer ads than before. He got confirmation from an unnamed Google employee that Google has in fact reduced the number of ads displayed.

Google has done a lot of research with users and found that fewer ads mean less revenue SHORT TERM. But long term the advertising revenue actually goes up. Why? They found their users started trusting the advertising more and were more likely to click on ads.

Interesting huh? Scoble points out that this change will help users trust ads more. Ads will be seen by users as aiding in their search, rather than introducing noise. As Scoble also points out, it is conceivable that Google’s top-line will take a small hit as a result. However, if you think about it, all they’ve done is set the relevancy bar higher. Only the more relevant ads are being shown. Less relevant ads are simply being removed from the page. Well, by definition, less relevant ads have low CTR, deliver few clicks, and generate small total revenue. I doubt we’ll see an impact at all.

Google (quietly) introduces text link ads

And now for the bad. Along with the announcement of the PPA ad program, Google also slipped in an introduction to a text link ad unit. Text link ads suck on many levels. First of all, they are confusing to the user. Second of all, they encourage publishers to include certain keywords and phrases that would result in lucrative ads. I’m sure Google will take steps to curb this. Plus, if you think about it, if a publisher sticks in random irrelevant words in their copy, users are unlikely to be clicking and the distraction will discourage users from returning to that site. But the real issue is that up until now, Google’s ads have always been clearly separated from page content. With text link ads, this is no longer the case. Ads can now be intertwined with the page content. A couple forum administrators that have implemented text link ads from other ad networks have since removed them due to poor conversion. It’ll be interesting to see how this plays out for Google…

I’ve gained incredible amounts of knowledge from the Web. Being a self-described “knowledge whore”, I’ve spent countless hours on sites like Wikipedia and howstuffworks.com as well as other sources of knowledge like blogs, newsgroups, and forums. Almost without fail, though, every time I find myself digging deeper into a topic, I quickly realize that the topic is WAY more complex than I had imagined it to be. Try it sometime. Pick a topic and Google it. For even the most obscure topic, the sheer vastness of relevant information on the Web is mind-boggling.

I realized today that even though the Web has given me volumes of knowledge and wisdom, above all, the Web has taught me this:

You don’t know what you don’t know.

The Web bombards us with this lesson because it’s so damned efficent at information retrieval. In minutes we can open gateways to knowledge sources that might have taken hours or days before. More importantly, though, the highly-linked nature of the Web supports a breadth-first search pattern of knowledge gathering. You might be reading about sub-topic A and in the middle of a paragraph follow a link to sub-topic B and so on. I’m sure you’ve done this plenty of times. While your initial intent may have been to perform a linear search to ascertain information on a specific topic, before you know it, you’ve spent an hour reading about 10 different sub-topics. In one hour, you’ve gotten a broad, but relatively shallow understanding of several sub-topics.

If you had been performing research thru offline methods, you would have found an information source (a book, news article, thesis, etc.) on a single sub-topic and digested it thoroughly before continuing on to the next source. This pattern of information gathering is more similar to depth-first search. Using this method, in the same time as above, you may gain relatively complete knowledge of 2 sub-topics, but not even realize the existence of the 8 other sub-topics that you would have encountered if you had followed a breadth-first search pattern. In other words, you’ll know more about less. With the Web, you’ll know less about more. The curse of the latter is that you will have learned of the existence of many more topics which only further increases the magnitude of how little knowledge you have.

Anyway, it’s late and it’s likely that I’m just rambling, so I’ll cut this post off now. In conclusion, even though the Web has given me tons of knowledge, the most valuale knowledge it’s given me is the realization of how little knowledge I actually have. My guess is that by the time I’m an elderly man, instead of feeling old and wise, I’m going to feel old and dumb. Very humbling…

Read transcript of call

Some interesting tidbits:

on Google Checkout and the topic of CPA which came up several times

Brin: “Studies show that about 63% of users abandon their shopping cart before they complete the buying process.”

Schmidt: “We did Checkout to solve the problem that Sergey described, which is we simply want to make the whole process of buying just quicker, more foolproof, more likely to conclude. We think that that ultimately will translate into higher value for the advertiser, which should ultimately be reflected in our overall revenue.”

Schmidt: “all of our testing indicates that people are much more likely to purchase from our advertisers if they have enabled Google Checkout”
Kamangar: “Our goal is long-term to increase the amount of searches and commercial searches on Google, increase the click-through rates to our advertisers, and increase the conversion rates that these customers get because of increased convenience.”

Page: “I know there’s a lot of talk about the CPA, or paying per purchase. One of the issues with that is people look around a lot before they buy something. So it’s probably not the only information you want to look at when you’re paying for advertising. But we’re also excited about using more data like that in our models.”

Kamangar: “That [conversion] information sits in the hands of the advertisers, and they can adjust their bids if they like. Conversion rate signals are not affecting the ranking.”

So I guess the official partyline about Checkout is to increase conversion for advertisers by making the checkout process more convenient. However Checkout isn’t a revolutionary improvement in convenience. Services like (the now defunct) MS Passport, Y! Stores, and even many browser toolbars offer the ability to automatically recall a user’s relevant information (addresses, cc info, etc.). I’m surprised that their touting this functionality so much.

It was emphasized that through Analytics and improvements to the AdWords tool, advertisers can efficiently monitor conversion themselves and adjust their bidding appropriately. I’m not sure how realistic that is. It seems like it would require the advertiser to be fairly sophisticated or pay a SEM optimization firm to regularly analyze the conversion vs. bidding ROI data and adjust correspondingly.

Ultimately, Schmidt says that the idea that Checkout is the first-step in a move to CPA is “a very interesting idea” although Kamangar says “we don’t have any plans right now to use Checkout information for anything outside of Checkout”. Eh, who knows…

On ad quality

Kalamangar: “beginning in last year, we began incorporating signals from the landing page into the ranking method and the final quality score…It has produced a small number of advertisers that were engaging in arbitrage and other methods that resulted in lower-quality ads, and as a result, has allowed us to show more ads from higher-quality customers.”

I had read here and there about Google’s initiative to weed out advertisers who are essentially arbitrage plays, but never any official confirmation that Google was actually doing the landing page analysis and booting certain advertisers from the system.

On net neutrality

Brin: “we really care about net neutrality, not for Google as a company, but rather for all the small Internet companies out there”

Page: “I’d also point out that it’s largely been driven in countries like the U.S., where we have very poor last-mile connections for users, much slower than in countries like South Korea, which has spent less money and has much better connections, and haven’t been pushing for elimination of net neutrality.”

Even though what Brin is saying is true, let there be no doubt that losing net neutrality will cost Google money. It may not have a material impact on the business, particularly because of their existing relationships with ISP and Google’s significant leverage because of their huge brand/presence, but it will hurt the bottom line.

David Leonhardt wrote an interesting article about Google Trends titled “The Internet Knows What You’ll Do Next” in the NY Times today. He references John Battelle’s idea of the Database of Intentions that he introduced a couple years ago. The concept is quite simple. People use search engines to find information about things they desire, intend to do, want more knowledge in, etc. Essentially, it’s the person announcing their intentions. If all the searches performed were recorded and then made available through a database, it would represent the intentions of all the people who use the Web…which is so many people that it basically represents all the people in our society.

I had played around with Google Trends just a bit and did find it pretty intriguing. The bummer is that while you get a graph representing the trend, there is no y-axis. In other words, you see relative changes in search popularity but not absolute numbers.

Even still, this data is worth its weight in gold. The data essentially allows you to predict the near-term popularity of anything. In the article, the author references the fact that Google Trends data accurately predicted the order of the top 3 finishers in American Idol last season. An apparel merchant claims to have altered their inventory, favoring one brand over another based on Google Trends data between the two brands.

The big question is, should you have to pay for this data? The Attention Trust guys (of which I am a member) would likely say no. After all, why should a company profit off of you by recording your search behavior? Or atleast if they do profit, they definitely should make the same freely data available to users of the service. My guess is that Google will keep the Trends service free but charge for API access to the data. I don’t believe an API exists yet but you gotta know it’s coming. Businesses from all walks of life could use this data to optimize their operations.

Note: Google Trends could also incorporate Google Checkout data to create the shopping tool that I discussed earlier.

As many of you know, I’ve spent the past couple of years in the comparison shopping space. In that time, I’ve spent endless hours contemplating solutions to help consumers be more successful shoppers on the Web. One idea that kept recurring in my mind was a site where you could browse and search what other people have bought. For example, you could search “digital camera” and instantly see which cameras were the most popular today, this week, this month, etc. What the product is, how much was paid, and where was it bought. Of course the list of cameras purchased in any given day woudl be massive and even for a given camera, the list of merchants and prices woudl be massive as well. However, the key thing is that you would see purchasing trends. The data would indicate which products are hot and where the hot deals are.

The problem with such a site is how to get the data? There really is no feasible way. At best you could do an FF extension that could intelligently scrape a purchase confirmation page on a merchant’s website. The problem with this is that there’s no incentive for the consumer to do this. I as a consumer probably would not mind this data to be recorded if it was done anonymously and both the collection process and resulting data set was very transparent. However, it’s unlikely I would actually spend even a couple minutes to make sure the data was recorded accurately. Moreover, FF represents a small minority of Web users and the % who would have this extension installed would be small (there’s no real immediate motivation for a user to install it).

Well, this all changes with Google Checkout. By acting as a payment provider for merchants who use Checkout, Google can record the details of each transaction: what was bought, what price (possibly even what coupons were used?), etc. Voila! You have the data you need for the tool I described above. And because the data is taken straight from the horse’s (merchant’s) mouth, you do not rely on scraping or the user which are flaky. Beyond just the community value, there is also value for the user in that he/she could log-in and see all the purchases they’ve made online and possibly view receipts, rebates, etc. in one spot.

Sound cool? I think so…

Yesterday, Google announced a limited availability of their new web-based spreadsheet application aptly named Google Spreadsheets. Is it nicely-designed and does it deliver basic spreadsheet functionality? Absolutely. Is this the first web-based spreadsheet application? No. Is this the best web-based spreadsheet application? Possibly. Is this going to disrupt the MS Excel-dominated spreadsheet market? Not anytime soon. Let’s face it folks, barring some special circumstances, I can’t think of anyone who would prefer to use a web-based application instead of a locally-installed application. While Web UI technologies have improved to the point where a web-based spreadsheet is feasible, it still pales in comparison to the rich UI libraries available in any modern OS.

So does this mean that Google has failed? Not at all. Every large R&D organization has tons of development projects going on at any given time. However, in just about every other industry, the majority of those projects never see the light of day. Many projects are done just to gain internal expertise, others are skunkwork projects that never had a clear market driver, and other projects are simply cut by management for all kinds of reasons. The unique thing about the web software industry is that the cost of manufacturing is nil. If you have a piece of software, you can basically flip a switch and it’s live to the public. Google probably knows full well that many of its projects aren’t revolutionary in their present form, but there is little reason not to make it publicly available. Part of the research and design process is receiving feedback and I think that’s what Google Labs is all about. Kind of like “here’s a cool project some of us (engineers) have been hacking at…we wanna see what you think and how you guys use it”.

The problem which Google runs into is that they’ve built a brand like no other. A brand that screams innovation. Because of this brand, people expect awesomely innovative products from Google. Those of us in the tech community who follow the bleeding edge of innovation are often less than satisfied with the innovation we see from Google.

As I say time and time again, I think a lot of people in the community need to get a breath of fresh air – from outside the Silicon Valley bubble – and realize that a simple web-based spreadsheet from Google is not going to challenge Microsoft in any meaningful way for a long, long time.

Some bloggers have commented that this lack of innovation is starting to hurt Google’s brand and reputation for innovation. I don’t really think so. If you were to demo Google Spreadsheets to the average Web user, they’re likely to be wowed and comment on Google’s brilliance. However, it’s also a safe bet that they’ll still use Excel the next time they need a spreadsheet.

Note: I’m not anti web-apps. I think they will be very much be a part of computing within the next 5-10 years. Read some of my vision

Alexa is an Amazon company, that among other services, provides traffic details and rankings for web sites. Everyone references Alexa’s traffic rankings. I use it regularly to get a glimpse into how much traffic (and the traffic growth) of a particular startup. Likewise, many tech journalists/bloggers/pundits use Alexa rankings to analyze how much traffic one site is getting versus another and if the site is growing or decreasing in popularity. I’ve also read on many SEO discussion boards that prospective advertisers will look at your Alexa ranking to determine how much your ad space is worth. I’ve also spoken with investors who include an Alexa traffic check in their initial look at a startup. So, clearly, if you run a web-based company, that Alexa ranking is an important number. It will likely have a direct influence on your business.

There has been much debate over how accurate Alexa’s rankings are. To understand how accurate (or inaccurate) Alexa’s data is, you need to understand how their system collects traffic data. It’s really quite simple. Alexa offers a toolbar for IE that offers up a Google search box, Alexa site info, popup blocking, and page highlighting among other features. More importantly, it records every single URL the browser visits and sends that information to Alexa. The important thing to understand here is that Alexa’s traffic data is based solely on the traffic of users who run IE and have the Alexa toolbar. For some more specifics about how and what Alexa calculates, go here.

Up until this point, you might be thinking…hmmm this Alexa sounds pretty reasonable! Rishi, why do you call it a “myth”?

Alexa states that they compute “traffic rankings by analyzing the Web usage of millions of Alexa Toolbar users.” Millions huh? Seriously, have you ever seen anyone that runs the Alexa toolbar. I sure haven’t. Google or Yahoo toolbar, sure. But never Alexa. My hunch is that it’s more like hundreds of thousands, not millions, of users. Possibly millions have downloaded it and had it installed at some point, but not all are currently running it. Check out this post from a couple years back on SEOChat. The author concludes that “that the Alexa sample size was 180,000.” Since Alexa does not post any demographic data, all of this is 100% speculation. However, the point is that the toolbar’s user base represents a very, very small percent of the Web-browsing population. And what kind of people do you think have the Alexa toolbar installed? I’m thinking it’s more Joe InternetProfessional and less Joe Teenager and even less Joe Noob. Evidence of this is that MySpace is currently ranked #5 and Google is #2. According to every other data point I’ve seen, MySpace pageviews far exceed Google. My guess is that your typical MySpace user is a lot less likely to be running the Alexa toolbar than the typical Google user. If efforts were taken to make this userbase statistically random (like a Gallup does) then the traffic data would be a lot more reliable. But as it stands today, one could not expect the pool to be random and thus the traffic data just isn’t that reliable.

Another huge problem with Alexa rankings is that it easy to manipulate. The traffic rank of ItsRishi.com is a patheticaly low 3,741,850. According to Alexa, my reach per million users for the past 3 months has been 0.1. That means that 1 in every 10 million Web users visit my site on a daily basis. Is that accurate? Who knows. Even Alexa states that traffic data beyond the top 100k is not statistically significant. Fair enough. As an experiment, I will be improving my traffic rank over the next few weeks. How? It’s simple. I installed the toolbar on my IE browser and from now on will use IE whenever I am working on ItsRishi.com. What this means is that I’ll probably end up with a few hits on average per day. If I don’t see enough of an improvement in a couple weeks, I’ll recruit some IP addresses – oops, I mean friends – to also install Alexa toolbar on IE and use it while visiting ItsRishi.com. My hope is to crack the top 100k. I’ve read many stories from webmasters who have done this very same practice with great success.

While I will be doing this for ItsRishi.com purely to satisfy my curiosity, you can bet that many companies are using such techniques to magnify their Alexa rank. The higher your rank – especially if your rank is rapidly climbing – the more people are interested in your company. That’s just a fact. Whenever I hear about a startup for the first time, I usually do a quick check of their Alexa rank to get an idea of their popularity. If the rank looks promising, I assume the site is hot and I take a deeper look. As I’ve explained, this assumption is faulty, but there really is no better (public) way of determining a site’s popularity. So while the Alexa number may not be accurate, it is a number. It’s not really much different than when Forrester publishes a report saying that an industry is X billion dollars. When you want a number, any number is better than no number. I guess the lesson here is to be careful of how much meaning you try to extract from the Alexa data.

If you’re trying to compare two companies in the same industry, the best suggestion is to find a metric that’s more pertinent to that industry. For example, an owner of a online retailer comments: “There is one competitor in particular that I watch..He is a member of Bizrate, so I can count the number of customers that take his survery each month, and compare it to the number that take our survey.” In his space, the number of Bizrate survey submissions is a much more accurate metric for him because its a nice apples-apples comparison since you can assume that at his store and his competitor’s store, the % of customers who complete the survey is about the same.

You can follow my Alexa progress here.