What Microsoft should do in search

We can start by asking why Microsoft should be in search. Microsoft has two extremely successful businesses with Windows and Office that generate $13.1 billion and $12.4 billion in operating income in 2008. Granted search advertising is an attractive market. It is predicted to grow from $20 billion in 2007 to $40 billion in 2012. It’s also extremely profitable once you obtain scale and cover your fixed costs with 60-70% gross margins. However, market attractiveness is not enough: pharmaceuticals are also an attractive business with high margins and no one is suggesting Microsoft goes into pharmaceuticals.

In one word Microsoft’s problem is Google. The spread of broadband, Wi-Fi and mobile data network presages an always on connected future where applications can run in the cloud as effectively as they do today on your local desktop. In this world desktop operating systems and desktop applications become irrelevant. The company best positioned to take advantage of that future is Google. Google’s scale gives it an inherent cost advantage in providing services in the cloud. For instance, Google is rumored to have 15 data centers to Microsoft’s 4. Moreover, while Google may not be interested in Office like applications per se, it can offensively offer them for free to disrupt Microsoft’s core business without undermining its own given that it relies on advertising rather than software sales.

Looking at it from this perspective, Microsoft’s logic for being in search is clear: it is a strategic defensive move to protect its core Windows and Office franchises. Given this logic, Microsoft has to be in search. This leads to the fundamental question: can Microsoft succeed in search? Bing has around a 10% market share in the US and an irrelevant market share in the rest of the world. Even with Yahoo’s 18% of US search share, it will be a distant number 2 in search in the US and irrelevant in the rest of the world. Moreover, there is evidence that a fair amount of Bing searches are “passive searches” brought about by partnerships. These are much less valuable than active searches where the user intentionally goes to a search engine to search for something. Add to that the fact that Microsoft has 50% less search engineers than Google and a quarter of the data centers and it’s hard to see how they can succeed. Worse, even if Bing was a slightly better search engine than Google (as it is for some categories of searches), analysis of consumer behavior suggests it would not be enough for consumers to switch. People cannot tell the difference between the speed of two PCs if one is 25% faster than the other. The speed increase has to be above 50% for consumers to start to be able to tell the difference. Similar consumer behavior seems to affect search. As a result Bing would have to be significantly better than Google for a large number of users to switch which is unlikely to happen given the speed at which Microsoft and Google are copying each other’s new features in search.

Let’s ask the question a few different ways. Can Google sustain its leadership in search? It is not difficult to argue that search as we know it today is extremely primitive. We can easily envision a world where things like natural language search crossed with social and behavioral information will revolutionize the way we search. Unfortunately for Microsoft and wannabe competitors, Google is best positioned to dominate the “Search 2.0” market. The reason is simple: fixed costs. The amount of existing and new information on the web is incredibly large and growing exponentially. Google’s index for instance is rumored to have grown from 1 billion pages to 40 billion pages between 2000 and 2008. A company which developed a better algorithm would still have to crawl all the information on the web to apply its algorithm in order to offer relevant results. Doing so requires billions of dollars to build the type of data centers that Google and Microsoft have built. High fixed costs create huge barriers to entry. As a result, I don’t see anyone displacing Google in search in the next 10 years.

Given all this, what should Microsoft do in search? Historically the company has dealt very effectively with competitors which threatened its franchise by entering their business and giving away the product for free. When Netscape released its browser, Netscape Navigator, in December 1994, Navigator was priced below $50, with evaluation copies downloadable for free, while matching software for servers was priced above $1,000. Marc Andreeson, Netscape’s chief technologist, had a vision that the browser could grow to become a PC user’s primary interface to all of his or computing needs. The browser, he opined, would “reduce Windows to a set of poorly debugged device drivers.” By May 1995, Netscape held a 70% share of the browser market.

Recognizing the threat, Bill Gates sent his top team a nine-page memo title “The Internet Tidal Wave.” Promising to “embrace and extend” the Internet, Microsoft released its own browser Internet Explorer (IE) 1.0 in My 1995. With the August 1996 release of IE 3.0, Microsoft then largely matched the then-current version of Navigator. IE and its matching server software were offered for free and IE came bundled with Windows on new PCs. Versions of IE were also made available for Apple and Unix operating systems. Microsoft also struck deals with Internet service providers (ISPs) to use IE to provide Internet to consumers. To convince America Online, the largest ISP, to promote IE, Microsoft agreed to place AOL’s icon on the Windows desktop, even though AOL competed with Microsoft’s own ISP offering.

Netscape tried to survive by quickening its pace of innovation, by opening its source code so that others could improve its products, and by diversifying into new products. However, Netscape’s development costs rose rapidly, its browser share plunged, and webmasters increasingly optimized their websites from IE, not Navigator. By late 2002, Microsoft’s browser market share surpassed 95%.

Could Microsoft use a similar strategy with Google? At first glance, the answer would seem to be no. Google is offered for free so there is no way to undercut their price. However, Google is more vulnerable that appears at first glance: almost all of their revenues come from search advertising on Google’s own websites where it bears no traffic acquisition costs. Moreover, the vast majority of the revenues comes from less than 5% of the searches which are concentrated in the following categories: travel, product, real estate, vehicles, jobs, services, finance and legal. Long tail searches actually have very little economic value: what ad do you want to put next to “GDP of Zimbabwe”?

In other words Microsoft can disrupt Google by attacking its profit engine in the short tail of searches. This would prevent Google from spending billions of dollars on various new initiatives which often are aimed at Microsoft’s core business. Granted most of Google’s new initiatives fail, but the sheer number of initiatives combined with Google’s rapid iterative product strategy are clearly giving Microsoft cause for concern. To achieve this objective, Microsoft should do the following:

1. Attack the short tail of searches where all of the value lies:

  • Deploy most of your engineers to work on the aforementioned short tail.
  • Offer 85-100% of the revenues to sites that use Bing’s equivalent of Adsense (we now know that Google gives 51% for AFS and 68% for AFC).
  • Keep trying new initiatives like Cashback (which was aimed at product searches) to lure users to use Bing for those short tail searches (though discontinue them if they fail, as Cashback did).
  • Spend a few billion buying vertical search sites (Kayak, Trulia, Indeed, etc.) and deploying them globally to attract an ever greater share of those short tail searches.

2. Get scale in order to cover your fixed costs. The Yahoo deal is a step in the right direction, but it’s a pity that it has taken so long to be implemented. It should already be live! Continue doing such deals but please don’t acquire AOL or Ask. The integration would just distract you. Just power search for them.

3. Leverage your operating system leadership and installed base by putting a big search box in Windows 7 on the desktop. You can offer local searches, but the default should be Internet only searches using Bing (at least when there is connectivity). At the very least put a Bing search Box in the taskbar.

4. Use the market’s distrust of Google and their monopoly position in many markets to your advantage:

  • Present yourself as the friendly, open and responsive alternative to Google to publishers. Be fully transparent on revenue shares, etc. Giving 85-100% of the advertising revenues back to them will also help that cause 🙂
  • Give away Windows Phone 7 to mobile operators and handset manufacturers and tightly integrate mobile search into it. Many operators fear Google as much as they fear Apple and are open to partnerships. Given how irrelevant the revenues from Windows Phone are going to be just give it away but make sure you get Bing to be the default search engine.

5. Keep your research lab working on long term disruptive technologies in search such as natural search. Maybe you could use all the data on people’s hard drives to improve search results. Being the dominant OS you have the easiest access to that information. It’s likely any innovation would be rapidly copied by Google, but you never know, besides you have to stay product competitive.

Good luck! The world’s web publishers need a strong alternative to Google and you are our last remaining hope in search! You owe it to yourself and to us to succeed. We’re all rooting for you!

Newsletter

Select list(s):
  • Hi Fabrice,
    I completely agree with your point of view.
    2 things to provide some additionnal information :
    1) Have a look at http://il.youtube.com/watch?v=8TH8UbeNQdo&feature=related to get an idea of the disruptive technologies Microsoft Research is working on
    2) Another aspect of search, that will bring value to the user and revenues to the publisher, is the user and restitution interface. Google’s one is nearly the same than its first version. As you say, we are at the primitive age… A lot has to ne done on that topic. You can have a look at http://www.youtube.com/watch?v=BZuFUZpEZ-A to see what Microsoft prepares.
    Best regards,
    Juan (from Paris. PS : I do not work for Microsoft. As an entrepreneur I am very interested in following both fantastic companies, Microsoft and Google)

  • Thanks Fabrice this is an excellent piece. The subsequent acquisition of ITA (wihtout a doubt the best GDS out there) of course adds to this but seems to indicate the Google is again more agressive in populating the verticals. They have also been getting all the hype with their real estate initiatives (which eg recently launched in the UK with my portfolio company Zoopla). There are also issues on culture that MS needs to address. It’s amazing how well G does in attracting strong engineering talent (often in the form of startups)… and in keeping it. The lure of the G is hard to resist for these new upstarts.

  • The strategy of attacking the short tail is imminently sensible. But I believe Microsoft will be unable to execute on some of the points you list:Point – For Microsoft to use the market's distrust of Google would be a hard sell. Much as Google is distrusted, Microsoft is distrusted even more.Point – Windows Phone 7 is, by many acounts, a disaster – it may not get any traction at all. (http://www.infoworld.com/d/mobilize/windows-phone-7-dont-bother-disaster-211?page=0%2C0)Point – Microsoft may very well run afoul of anti-trust laws (again) if they attempt to integrate Bing into Windows 7.Although I agree on the need for a strong competitor to Google, I do not think Microsoft is in any position to become that competitor, unless they hit on a truly disruptive technology…and my bet would be some form of realtime search.

  • Microsoft has lost is critical user base, and in search it is just another "child". Remember wen Jimmy Wales said it will build a great search engine? (WikiSeek, in 2007), or that guys that leave google and formed Cuil? The thing is: they had tried to beat the unbeatable because the concept of "search" (as his today) exists in peoples mind because exists Google! In my opinion its Facebook that will be disrupting Google as leader in web advertising. Not by building a search engine, but something similar and with more value to the user. There is no need to crawl all the billions of pages out there, but only, those that matter, the relevant ones, those that are "alive". Now, in our time, most of websites become alive not by PageRank, but by Facebook friends suggestions. And thats the point of disruption. I think it is just a matter of time until appears a "FacebookSense". Meanwhile, tough times..It's just my gess 😉 Good luck!

  • I don’t know Fabrice, your logic breaks down for me a bit. I agree that cloud computing is the wave of the future (and actually Microsoft was already there philosophically when I interned there back in 2001), I just don’t know that taking some of Google’s revenue just for the sake of taking Google’s revenue will get you there. Strategy is all about being unique, not about taking a me-too position. When everyone else was doubling down on healthy food, KFC introduced its (ironically named) double down sandwich -bacon in a fried chicken “bun.” It’s been selling off the charts. Microsoft is not good at search, and they’re not good at complex algorithms – that’s not who they are. They would need a completely different hiring profile, completely different sales force, etc. What they are good at is business productivity products in the corporate environment (the home piece is really just an outgrowth of corporate). I would rather see them invest in servers to run their applications remotely and then really beef up the applications with useful features that they have been avoiding because they would take too long to run on low-memory machines (and yes, such machines still exist, my current work PC crashes ALL the time).

    I would also think about branching into business intelligence software – I think that that is going to be the next wave of the computing future, and most of the products out there don’t require algorithms more sophisticated than what Excel runs already. I used to be in charge of buying BI software for my old company, and most of the products were trying to sell themselves based on sort and sum functionalities – not rocket science. Yet, the ability to customize BI tools for specific corporate environments could be vastly useful in many contexts – for instance hardcoding the region -> district -> store hierarchy or automatic data lookup tags like OECD, ASEAN, etc. (could save me a ton of time in VLOOKUPS, and even better if the software somehow knows that Congo, Democratic Republic of; Democratic Republic of Congo; DRC; Congo – Democratic Republic of, etc. are all the same place). BI software takes a ton of computing power and storage and so would motivate Microsoft to expand capacity, but doesn’t take algorithmic complexity and it’s targeting to their prime audience.

    If they were to go into search, still I would want them to focus on corporate and business productivity. For instance if I did a search for “Zimbabwe GDP” (which by the way I did in Google just for fun and you’re right – no ads at all!) , I would love it if instead of linking me to wikipedia they knew my company had access to EIU and that that was our preferred datasource, and they took me there as the first choice. Or if I typed in just “Zimbabwe” I got an EIU country report instead of LonelyPlanet. Knowing company specifics and creating tailored searches would be really value-add. For instance, I really wish Gmail would stop showing me ads to hire my boss as a speaker.

    The point is, strategy is about sticking to what you know, and finding ways to do more of it and be better at it. If you’re not making tradeoffs, you probably have a very poor strategy. Look, you would never get OLX into social networking just because it’s going to be huge, it relies on word-of-mouth marketing which also will be useful for you, and just because they list classified ads, right? (essentially your Microsoft argument)

  • I disagree with the 10 year part. One reason is, as you pointed out Google as well as Bing is free. And one thing we all learned from the Friendster, MySpace and Facebook wars is that users are NOT loyal. They go where other users are. We can debate which is better… MySpace or FaceBook, I have to say MySpace. But I use FaceBook more because more people go there instead. However, when the debate comes to Google and Bing. Google may have more of an index, but what does that have to do with me when I’m searching for “Superman comics”. I use Bing more because it’s search narrows down what I’m looking for better. Google may index more, but who really looks past the first 2 pages of a Google search anyway. I have to admit my adoration of Google has waned tremendously.

    Five years ago I read an article in Forbes about publishers using Adsense and making a steady decent income. Today every single one of them no longer uses Adsense. Google has a bad reputation among good content providing publishers. Have you ever heard the term “Google Slap”? That’s where Google bans a person out of the blue for absolutely no reason at all. And when you try to contact their response is, “We can’t divulge the reason for the ban for security reasons”. Do I sound like a disgruntled publisher? well I am. Google in my opinion was great innovative company 5 years ago, however today they are nothing more than a acquirer.

    Back to my point, which is PERCEPTION IS REALITY. It just depends if the public rightly or wrongly thinks one option is better than the other. For some reason it reminds me of what Guy Kawasaki said once “I’d never go with the wisdom of the crowd, because the crowd has done done some pretty stupid things”