Given all the Google hype, I was expecting more. The speech was basically a product marketing tee-up presentation for Google OneBox for Enterprise. (Tee-up means you first motivate a product-specific agenda in thin disguise and then introduce the product with a subsequent knock-down of that agenda.)
Dave started with the (more than somewhat self-serving) statement that consumer technology is driving high-tech innovation, citing the iPod as an example.
Enterprise software bashing is getting popular these days in search circles. (Fast did the same thing in their "contextual BI" presentation.) I think people forget that perhaps SAP's UI would be as simple as Google's if all SAP did was to assist in finding documents on the Internet. Unfortunately, automating a company's operations (or letting end users analyze data) is a bit more complicated than that.
Does this mean that enterprise software vendors shouldn't focus more on design and usability? No way. But it does mean that people should be realistic in comparisons.
Dave then proceeded to ask one of my favorite questions: why is it that Internet search is a greater than $10B market and enterprise search is a less than $1B market? I didn't like his answer -- which was that enterprise search didn't provide enough value or impact -- so I'll tell you mine. Internet search is 10x (and by my estimation perhaps 50x) enterprise search because:
- Search is predicated on the existence of the thing you're seeking. When you are searching the entire Internet that assumption is usually true -- i.e., something probably exists that answers most any question you can pose. Dave cited a person who diagnosed his dog with lyme disease and thanked Google for it. That person can thank Google -- but, he should probably have greater thanks for the person who wrote the article that helped them do it. This existence assumption just isn't true inside the enterprise.
- Enterprise search is typically thrown in as a band-aid by IT departments. CIOs know they need a search engine in their corporate portal, so they buy one, index whatever they can find, and hope that solves the information-finding problem. It's a generic, haphazard approach that lets the CIO check a box, but doesn't really help people find content. (What's needed are content applications and custom publishing systems for content, combined with BI tools for quantitative analysis.)
- Consequently, enterprise search has low user satisfaction. I can't tell you the number of times, I've heard the words "enterprise search sucks" (which incidentally scores about 2.7M hits on Google). See this post entitled Just How Bad is Enterprise Search by Tony Byrne or this one by John Udell titled How Enterprise Search Could Suck Less. People don't write these kinds of articles about relational databases or BI tools.
- The office worker era. From the 40s or so thru the 60s.
- The knowledge worker era. From the 60s thru the 90s.
- The self-directed innovator (SDI) era. Now.
- Talent. You have to know how to find them. Dave shared several anecdotes about how Google does so with contests, quizzes, cryptic billboards, and such.
- Direction. They need to know what the company is trying to accomplish.
- Information. They need to find information to get their jobs done. (Tee-up in process.)
- Are impatient
- Work anywhere
- Are process averse
- Have intermingled work/personal lives
- Collaborate with a broad network
And then, tada, the big conclusion: search is the application for SDIs. Search is not middleware or infrastructure; it is an application. (I don't know why this was such a big deal, but it seemed important to him to reposition search in this way.)
He then did a demo that included both regular Google OneBox and OneBox for Enterprise. (Both, by the way, are based on recognizing regular expressions, parsing them, and then returning special information on top of the search results in response to a search.)
Here was the flow of the demo, modeled around a sales manager going to NYC.
- Search "UA 10" -- OneBox returns links to track the flight status. Flight's on-time.
- Search "Weather New York" -- OneBox returns a five-day forecast. Bring your raincoat.
- Search "East Coast Sales" -- Returns nothing. Anti-demo to show security and the need to login.
- Login as sales VP.
- Search "East Coast Sales" -- OneBox for Enterprise returns a Cognos Metric on top of the results list showing sales. The VP knows their is an opportunity at XYZ corp.
- Search "XYZ opportunity" -- Returns information from salesforce.com about XYZ corp. Tom Winkler is a key contact there.
- Search "contact Tom Winkler" -- Information about contact Tom Winkler is returned, presumably also from salesforce.com.
However, I worry that OneBox and other enhancements are starting to clutter Google's otherwise clean UI. (Will Google fall into the Microsoft bloatware trap?) I worry about picking which OneBox provider to invoke when I type "opportunity GE" -- in a global organization on 5 different SFAs, each holding scores of GE opportunities.
Finally, we should remember that this is cream-skimming and that cream-skimming is what Google does best. They skim the cream off content, seemingly getting more money for helping people find conent than the authors of it.
With OneBox for Enterprise, they will skim the cream off enterprise apps. Going forward, they will be able to answer the simple questions from enterprise applications just as they return "weather New York" today.
That's not bad. But it's not BI. And it's not ERP either.
1 comments:
Why do I think that the self-directed innovator is going to become the inspiration for a series of hilarious Dilbert cartoons?
Post a Comment