free html hit counter Tim Armstrong To Lead AOL - Further Thoughts | John Battelle's Search Blog

Tim Armstrong To Lead AOL – Further Thoughts

By - March 13, 2009

This I find quite worthy of comment. More after breakfast…

AOL is getting new leadership again, just two years after the outgoing executives were chosen to turn around the struggling dial-up and content company.

Google Senior Vice President Tim Armstrong will take over as chairman and chief executive, replacing Randy Falco, said Time Warner, AOL’s parent company. Ron Grant, AOL’s president and chief operating officer, will leave with Falco after a transitional period of a few weeks.

Kara has an interview here. In it, Tim says:

“For me, it is a great opportunity to go to what I consider a top-five Internet brand…I am looking forward to taking what I have learned at Google and seeing what I can bring to really help AOL.”

One of the things Armstrong might bring is a spin out, Kara writes, and that is certainly something I’d welcome. (Type “set AOL free” into Google, and see why. I wrote that FIVE years ago).

But Tim’s leaving Google strikes me as yet another signal of how much Google as a company has changed. Good people, whether engineers or top executives, are leaving the company to do other things, regardless of how wonderful the place is, how wealthy it made them, or how hard the company strives to create a culture that encourages retention.

I’ve had several discussions with folks who’ve left Google lately. Here’s a direct quote from one of them, who is starting a new search related company: “It’s very hard to take risks at Google.”

I bolded that statement because it strikes me as poignant, and telling. I pressed as to why that is, and the founder told me that Google is probably “the best large company in the world to work for” but that it is driven by its current search and Adwords businesses.

Does that sound familiar? It should, if you are a watcher of the IT industry. Replace “search and Adwords” with “Windows and Office.”

Around ten years into Microsoft’s existence, good people starting leaving the company (and, not surprisingly, about five years after its IPO…). Not because Microsoft was a bad place to work, quite the contrary was true in fact. But because the company was too big, or too slow, or too indifferent to the ideas they had. It’s not a rap on large companies, it’s a fact of being a large company – you can’t run in every direction your employees might want to go.

It’s been five years since Google’s IPO. The company is big, and despite all of its efforts, it’s getting slow. It has to, in order to protect that which brought it to greatness. (For background, read my piece from 2005 about the worm turning. In it I write: Can the company shift its culture and avoid the fate which ultimately hobbled Microsoft? That, more than anything else, will define the next chapter in the company’s fascinating story.)

Instead of trying to retain great talent, perhaps Google should encourage them to leave. Start a different kind of founder’s award – one that seeds new startups. Given all the talent and all the interesting new companies springing up from the fertile soil of ex-Google land, I’d wager that fund would do damn well.

Related Posts Plugin for WordPress, Blogger...

9 thoughts on “Tim Armstrong To Lead AOL – Further Thoughts

  1. Bernardo says:

    Interesting. I hadn’t before wondered how even tech companies struggle with their bureaucracy. I liked this article.

  2. JG says:

    I pressed as to why that is, and the founder told me that Google is probably “the best large company in the world to work for” but that it is driven by its current search and Adwords businesses.

    Nice analogy, between search/adwords, and windows/office.

    Let me note that at least MS kept innovating in windows for many, many years, even if it could not break out of that business model. Yes, I know — it “innovated” by copying other people, by buying technology rather than inventing it in house, and by monopolistically locking in certain “innovations”, such as when they tried to claim the IE was an integral part of Windows. I still laugh about that one.

    But the point is, the OS kept changing and evolving and getting better and handling all sorts of off-the-shelf hardware and software, getting better graphics APIs, better networking, etc. Whether or not we like MS’s style of innovation, they were indeed innovating. Today, I no longer have to use Winsock to connect to the internet, as I did in the days of Windows 3.1. It’s all built in, and much easier for me to understand and manage. It is a concrete improvement, and it is very visible to me, as an end user.

    I look at search at Google, on the other hand (the analogy-product to Windows at Microsoft), and I see a slowing down rather than a speeding up of innovation in that core product. Windows really changed a lot in the 10 years between 1992 and 2002. It was drastic, it was dramatic, and it was tangible. How about Google search as a product, in the 10 years between 1998 and 2008? Not so much.

    Once every 20 or 30 searches I get a YouTube video (aka universal search?) at the top of my results, rather than a regular blue link. The spacing and layout of the page has changed slightly, to allow better emphasis of relevant information. (Google touts this as one of the wonderful ways that they’re innovating. So what? Windows layouts and fonts and spacing changed and improved a lot, too.) And I can also track UPS packages and find movie times, if I am able to remember the special command-line option for doing so. But really, how has search at Google fundamentally improved, in the past 10 years?

    Maybe it has, and I just don’t see it, because all the improvements are hidden, locked away, invisible. But even on a visible level, there are so many other interesting information retrieval things that Google could be doing, to improve the process of search itself, i.e. the user interaction with the engine. And they are doing none of those, I know that for sure.

    So after 10 years of search, it feels like Google is slowing, and getting slower, whereas after 10 years of Windows, it was picking up speed and finally getting really good.

    Someone snarky could probably say something like “Well, that is because Google search was so good to begin with, and Microsoft windows was so terrible to begin with. That’s why windows improved so much in 10 years, and search did not.”

    That is a good point. But even with search, there is still so much left to do. Windows is now 90% solved. How much more in basic operating system design is left to be done? Search on the other hand is only 5% solved. The whole horizon is out in front of us, and rather than picking up speed, Google is slowing down. Given the landscape, that’s exactly the opposite of what should be happening.

  3. Bernardo says:

    JG: “…there are so many other interesting information retrieval things that Google could be doing, to improve the process of search itself, i.e. the user interaction with the engine. And they are doing none of those, I know that for sure.”
    _____________________________

    That’s kind of vague for being your comment’s focal point. Where is this degree of innovation left coming from? I’m actually asking – I’m curious.

  4. Wow! I couldn’t disagree more with J.G.’s comment that Google search hasn’t evolved at a rate equal to that of Microsoft Windows. Here are some of the highlights:

    1. Google Desktop Search… I use this at least 20x/day to find my emails, IMs, webcache, files. Indispensible!

    2. Images.google.com… if you’ve ever wanted a graphic/picture for an internal powerpoint this is your savior.

    3. Google Book Search… find any passage from over 1M books! instantly!

    4. Google Code Search… no this is not just google search against source code. You can even look for regex’s here!

    5. Checkout site search by typing “site:www.somecompany.com this that” and you have an automatic search index of all content on http://www.somecompany.com (goodbye inhouse search engine!)

    6. Search results have gotten SOOO much more accurate b/c of algorithmic refinements over the past 10 years. How quickly we forget.

    7. Google Alerts delivering customized subscriptions to searches on new content daily!

    And this is just the innovations in search… not to mention Google Maps, Gmail, Google Docs, Picasa, Google Analytics, Google Reader, and many others.

  5. andilinks says:

    Companies evolve and mature and the MS Google comparison is very interesting. Makes you wonder what it will be like to hear the phrase “aging Xer” or “elderly boomer.”

    I’ve lived to see AOL go from a struggling pioneer to neglected has-been and it sure didn’t take all that long.

    We live in exciting times.

  6. It was the same at newspaper companies, which is why they are dead. Replace search/adwords and windows/office with classified/display and it explains why talent eventually flees large companies.

  7. JG says:

    Where is this degree of innovation left coming from?

    @Bernardo: Look into all the people doing interactive search (Google is only single-shot), or doing clustered search (Google only gives you 10 links at a time), or doing transparent search (see this comment http://battellemedia.com/archives/004866.php#comment_137162). Lots of published research about all the possibilities, and all the advantages. These techniques make search orders of magnitude better.. Google could pick up on so much of it, implement it at scale, and run with it. They do not.

    Not to mention the fact that Google has been a closed system for a whole decade now, and has not allowed the open web community to meta-search its results. Meta-search is a known way of improving search results. They do not allow this sort of openness and community-wide building.

    @MatthewQ:
    (1) The actual quality of google desktop search, if you’re doing anything beyond known-item seeking, is terrible. I’ve seen it evaluated. It’s bad. Now, I know folks inside Google, and they say it’s because there are internal disagreements about how much technology to release outside of the ‘plex. Some ppl are fearful that if they put too much search intelligence into GDS, then someone will reverse engineer it and steal the secret sauce. Whether that’s true or not, what I am left with is an underperforming product.

    (2) Images: If you ever want any image, it’s fine. If you ever want *a specific* image, you’re toast. How do you describe the fact that what you’re really looking for is not just any picture of Madeline Albright, but that one where she is stepping out of the limo, and there is a bird flying over on the left-hand side, and the person next to her is wearing a red sash. I’m not saying that sort of image search is a solved problem. But there have been many steps taken toward solutions, including query refinement by drawing in color splotches on a blank canvas. That way, I could draw a red sash-like splotch where I remember it, and try to get to that image again. Has Google even bothered with this, improved its user experience around image seeking? No. So it feels like their innovation is slowing.

    (3) Google book search: Where’s the interesting search technology here? Book scanning and OCR, in order to find quotations. That’s nice, but it’s not a search innovation problem. It’s a search “hardware” problem, meaning that it’s data collection. Once you have the data, actually finding the passages, even with OCR-degraded text is a problem that was solved a decade ago, before Google even existed. So in this process, where is Google search “software” innovation? Google gets credit for putting together the data. They do not get credit for making search better.

    For example, instead of just looking for a book with a particular passage, what if I wanted to find all books that described social conditions in rural areas of industrialized nations during time periods before a large economic crash. What’s the query that I should use, to find all those books? How does Google help me formulate that query? How does Google help me refine it, for example, by telling me what time periods or nations I should restrict my query to? Now if it helped me do that, THAT would be an improvement, and a valuable application of book search.

    (4) Google Code search: Ok, you’ve got me on this one. Cool beans. But.. ultimately.. only useful for a very, very small segment of the world’s web population. Niche vertical. Good, but hardly world changing. It’s not to scale.

    (5) Site search: Yes, that’s useful. It’s a nice feature. But hardly earthshattering. It’s just a filter over the ranked lists that go into a search result. I’ll bet there was some pretty clever engineering that went into enabling that filter at scale. Google’s got some amazingly brilliant solutions along those lines. But from a user’s perspective, it would be much more interesting if Google did something like look at my past 4-5 queries, and then automatically suggest to me: “Hey, we see that you are looking for [...] type of information. Why don’t you do a site-restricted search to these 10 sites”. Then it could give me an interface for placing a check mark next to all the sites that I wanted to use. Think about it as me doing a search for what sites I want to use to do a bunch more searches.

    So rather than me having to somehow miraculously discover what those 10 sites should be, Google should be applying its massive search intelligence algorithms to uncovering and discovering those sites, for me. Because a site operator is totally useless if the user doesn’t even know that a site exists! So google should be helping me *find* those sites, to begin with.

    (6) Can you back this one up? Can you show that they’ve really gotten that much better?

    Personally, I think Google results have stayed about the same. They haven’t gotten worse, but they also haven’t gotten much more accurate. Not noticeably so. Rather, what I think has happened is that the web has gotten larger, and spam has gotten larger. So Google has done a fantastic job of keeping its results just as relevant as they were 10 years ago, despite the fact that there is so much more good information, and so much more bad information. But there is so much more to be done. Improved interactivity, clustering, visualizations, transparency, etc. It’s not enough to stay the same. You have to get better.

    (7) I’ll admit to not having used Google Alerts.. but as far as I understand it, it’s just an email-based push version of your standard search query? It’s the regular search, with a “filter out what I’ve seen before” wrapper on it? Ok, that’s nice, but from an information seeking perspective, not at all clear how innovative it is. Oh, I’m sure they had to do some engineering to make it work. But is the Alert itself somehow fundamentally different from, or better than, a regular search? Can I perform relevance feedback on the whole-document level on the alerts that I get, and tell Google whether or not an incoming alert was relevant, so that they can update my user/topic model? Are there more transparent, deeper-level relevant feedback mechanisms, so that I can tell Google exactly what it was about that document that was relevant or was not relevant? For example, if my alert is for [elephants in africa], and I get a email about information on an african elephant in a Detroit zoo, is there some way that I can tell Google that I liked the fact that there was an elephant, but that the location was wrong, so that Google’s backend alerting system can learn from that mistake? Has Google opened up and improved its system to the point where it allows me to do this? As I said, I haven’t used it, so you’ll have to tell me. But I suspect that they haven’t bothered with this.

    I’m not interested in all those other things.. Maps, email, Docs (which they bought), Picasa (which they bought), Analytics (which they bought), etc. Microsoft also has Flight Simulator, but so what? That’s not relevant to Windows. We’re talking about the Windows/Search analogy. Not side products.

  8. JG says:

    I’ve had several discussions with folks who’ve left Google lately. Here’s a direct quote from one of them, who is starting a new search related company: “It’s very hard to take risks at Google.”

    But this statement I believe. I believe it’s hard to take risks, because over the past 10 years I’ve rarely seen any risks taken. And that has, in large part, led to my outsider perception that Google does not innovate in search any more.

    It’s starting to sound like my perception, which I’ve been ranting about on your blog, John, for years now, might not be that inaccurate?

    I would be interested in you doing a followup interview with Louis Monier, John. Do you remember doing an interview with him a few years ago, as he was in the process of going to Google? I remember something about him saying how excited he was, because he was being given time to think and create and innovate new spaces and ways of doing search. Then wasn’t he gone again, like 1.5 years later? Maybe it was because, ultimately, it was such a hard place to take risks.

    I’d be curious to have that followup interview, to get the full story.

  9. Andrew says:

    JG, Google is not a search company, and has not been one since it became profitable. It’s an advertising company. The big innovations that google has made in the past six years have been nearly all in advertising; how to bring in more money for both Google and Google’s customers–the advertisers and affiliates.