04 Mar

Guaranteed Google Acquisition of the Year: Splunk

You read it first: Google is going to acquire Splunk in the next 18 months, and it’s going to cost them. They’ll ship it as an appliance for enterprise customers as well as give no-nag freebies to (geek) consumers. Well, I don’t know that they will, but they should. Splunk 3.2 just got released and looks quite promising. We started using 3.1 about a week ago and we’re just noobies so far, but it’s already obvious this is essential software for system administration, security and QA purposes. I can see ahead the same kind of learning journey I started with Google Analytics last year. GA by the way still resists me stubbornly in some areas, with campaign tracking and site overlay only partly working. OK we have some sites with three different GA profiles, goals, filters, transactions and whatnot, but we have relatively small sites and this shouldn’t be rocket science. The problem is, web analytics implementations are tedious and slow to debug, especially Google because it lags by several hours (last year it took 24 hours for anything to show in reports).
This type of software can be frustrating to master, but the level of visibility and understanding they bring is a key component to running an online business. I know you’re not supposed to build lasting competitive advantage on operational efficiency alone, but I do believe there are going to be winners and losers in this race. How well do you know what’s going on with your online business? How fast and how decisively can you act? Ourselves, we’re just getting started, but I’m sure it’s an effort that will be well worth sustaining. Right now I’m mainly assessing our level of non-quality, and it’s not a pretty picture. But at least I know what’s not running properly with a tool that scales (though from my limited experience, be careful with data exports with Splunk running on a production server).
PS for the Splunk team: please redirect the old Splunkbase links to their new URL.

23 Sep

Stemming Gone Wrong

Look, Google, I actually know how to spell Chile, ok? Stemming “Chile” into “Chili” and polluting my queries with recipes and rock bands that have NOTHING TO DO WHATSOEVER with the country of Chile is not helping. I am not writing my queries in French (Chile is spelled Chili in that language). I seem to remember Google used to suggest alternate spellings and link to them. Now Google has decided they know better than the user and mix their braindead second-guessing into top 10 results. Way to go to destroy the product.
Anyway, back to using Yahoo as my primary engine, though in fairness it’s about as bad. It’s amazing how utterly useless many Google SERPs have become. Maybe I have higher expectations. Look for country-specific information (not just hotels and resorts) and you’ll quickly see what I mean. The current state of search is about as bad as Altavista circa 1998 I guess. It’s time someone pulled a Google on Google. Jerry Yang, I just gave you your roadmap. And do something useful please, let’s not waste our time on jokers that pretend hand-writing 10,000 pages is a substitute for a real search engine.

15 Mar

What’s Going On with Adwords and Adsense?

I can only describe my attempts at improving our performance with Adsense as intensely frustrating. First there was the poor, canned customer support. The incomplete reporting (want to track ad channel performance over time? Tough luck, gotta buy third-party software). The spammy gateway advertisers with absolutely no content on their site (automated Adwords/Adsense arbitrage would be a nice way to describe what they do). The irrelevant ads because, let’s face it, automated keyword targeting can be quite dumb. Then an apparent lack of understanding of how a publisher works. If you’re serious about making money, in most categories Adsense is only going to be a convenient way to monetize unsold inventory. This can’t be your primary money maker by any means unless you’re really lazy and love to leave a lot of money on the table.
But despite the fact most of their current content partners are utter crap (I know, I tried to find interesting finance sites to buy ads on), the Adsense team seems to hold the weird opinion that without them you wouldn’t be able to make any money at all. Hello, not everyone is handling a blog trying to get extra pocket money here. Not to mention cheeky decisions like the introduction of site targeting and site representation on a sneaky, opt-out basis (do no evil, right).

Read More

02 Jun

What Comes After a 10:1 GOOG Stock Split?

This graph is definitely impressive. Kudos to the Google team for creating so much wealth, and to people buying at IPO time. I’m not one of them, I thought it would rocket and crash, so I’ve given up on pretending I know anything about what I’m talking about in this industry. Or is the stock market delusional and I’ll be proved right in the long term? For all those years I felt Yahoo could make a big "software comeback" if they wanted to, and it’s exactly what’s happening, so maybe I’m only half a buffoon. That and I did write before the IPO it would take time to burst the Google bubble.
In any case, I’m now using Yahoo as my primary web search engine (still using Google for news). Results seem slightly better. A query on, say, "gaming mouse" doesn’t put something written six years ago in the first spot, as is the case on Google. And there’s less spam. As I’m using Firefox to initiate queries the barrier to exit is zero and retraining is minimal, as Yahoo mimicked many of Google’s tricks. With Firefox I’ve given up on the Google toolbar and never looked back.
Google is to web search what Hotmail was to webmail a couple years ago: not a major innovation in a long time, and more spam by the day. Ripe for replacement, and you don’t even need to change your email address. It seems obvious Google hasn’t got a clue about how to properly handle commercial sites and tell apart those with relevant content from the rest.
Now that I have a whole omelette on my face with the GOOG explosion, I might as well reinstate my bearish outlook on the company! The lasting, defensible position? I don’t see it, and the market is becoming big enough to warrant more focus and resources from competitors. Contrarily to the slavish reporting and blogging around, Google is not the first company run by smart engineers each doing their own thang. The paramount example was Bell Labs within AT&T, an organization that has delivered orders of magnitude more innovation and hardcore research than Google probably ever will. Or think of Xerox. Look at where they are after market conditions changed under their feet, it’s not a pretty picture.
Many companies have too many marketers and too few engineers in charge. Google will prove to have tipped too far in the other direction.

21 Jul

A Critical Mass of Advertisers Makes Contextual Ads More Relevant

Andrew Goodman:

"I can attest to having been extremely frustrated doing web searches trying to find things like software, templates, standard documents, and legal agreements to help me grow my business.
But when advertisers of such services start to put their little classified ads up next to those keywords, the process gets a lot faster. I still use PDF Factory Pro, for example, because I saw their ad on Google next to a jumble of useless results for my web searches on PDF publishing."

Hear, hear! I find myself actively looking for the ads when I’m doing product/shopping-related queries, as sometimes all I want is to know where to buy the freaking stuff and how much it costs. Someone explaining why they offered the thing for their grandma for Christmas, or an obscure academic paper which happens to include my query’s keywords, is sometimes just noise. It’s all about context.
01/14/05 update: Google search within sponsored links.
05/30/05 update: indirectly related: Yahoo Mindset.

03 Jun

20 Questions with Gary Flake, Head of Yahoo Research Labs

In an interview with Gary Price:

"The beautiful thing about a relational database is that its structure tells you a lot about what is important. Database designers have been brilliant at optimizing databases (both the organization of the information as well as the algorithms) to best exploit this regularity. When you flatten out a database, those paths towards optimization often aren’t available.
A middle ground — which is not perfect, but adds a lot of utility — is to convert structured into a semi-structured form. Today, we treat documents as a big bag of words and index those words. In this semi-structured approach, we take structured information (say, the value of specific fields) and synthesize fake words that represent the fact that “document X has field Y with value Z.? Now, clearly I can’t run a SQL query on this representation; but at least I can search for documents with specific field:value pairs.
I’d like to tell you that we will be able to make an unstructured database as powerful as a structured database; but that simply is not the case. Nonetheless, the fusion of structured and unstructured data and approaches will add a lot of utility to the lives of most users."

Emphasis mine. On one hand it’s still frustratingly inefficient to look for information on the open, unstructured web. On the other hand perfectly structured, metatagged content is a dream that’s not going to happen, so I fully concur with Gary Flake’s statement. Structure and tag what you can, know where to stop, and let the rest self-organize.

14 May

Google Turns Usenet into “My Groups”

The new Google interface to Usenet has Atom feeds for every newsgroup (add /feeds to a group URL) and features membership more prominently, with shortcuts such as favorites groups and bookmarked threads. Some of the display modes work better than others, but where did the ability to sort search results by date rather than (dubious) relevancy go? This just after the new, very consumer-friendly (as opposed to power-user oriented) Blogger. Waiting for Froogle’s facelift…
Update: I missed the most important: it’s not just Usenet, and you can create your own groups.
06/05/04: I had forgotten or missed the Neotonic acquisition in April last year, which is relevant to this (and possibly Gmail?). Brandon Long and David Jeske worked for Egroups (which became Yahoo Groups) then Neotonic and now Google. I also just learnt that Mark Fletcher who founded ONElist/eGroups is the founder and CEO of the company behind Bloglines. And I don’t think I knew about the templating system Clearsilver (used by Techdirt among others) either.

15 Apr

Obligatory Comments about A9

Since John Battelle is hyping it like the new slicebread and everybody is getting excited, let me throw a few links from my archives, just follow the trail:

At the end of the day, I reiterate that Google and Amazon have more interest in, and are more likely to, ramp up their partnership than butt heads, and I believe A9 is a step in that direction. Yahoo, MSN and Ebay are strong enough that these two don’t need to invent themselves new enemies, when the reality is, their strengths are perfectly complimentary from both technical and business standpoints. Time will tell, and I’m ready to eat my words if I’m proved to be wrong, but I think Battelle has it exactly backwards when he claims this is a somehow a move from Amazon against Google. I expect them to move even closer to each other.
Update: evasive interview. Right now there’s not much to the site to justify the buzz (as opposed to Gmail), let’s see what else Amazon has up its sleeves. Remember Amazon already embedded Google results in its main site (though if I must believe my referer stats, that doesn’t drive much traffic).
Update: Rex Hammock: "Battelle’s whole Google vs. Amazon riff is a red herring […] Amazon’s amazing skills at such application of collaborative filtering in providing search results will be A9s’ true secret sauce."
I believe John Battelle lets its media background transpire by framing this as an umpteenth "clash of the titans", not unlike, say, 67.39% of CNet’s articles. Again, more hype than analysis, but from a PR perspective, as Jason points out, good mileage.