How Google fails at failing

Since the announcement that it was killing Google Wave, Google has turned on the spin by proclaiming how they “celebrate our failures.” There is a lot to admire about Google, and one of those things is its ability to experiment and, as CEO Eric Schmidt said, “try things.” It’s not just hard for many organizations to find the culture and capacity to do that, it’s hard for them to acknowledge when those things don’t work.

Danny Sullivan, at Search Engine Land, mapped out many of Google’s most notable recent failures, and wondered just what the company was really gaining in terms of knowledge:

“But in its statements to the world, Google rarely sounds like it’s celebrating these missteps. It doesn’t really document anything that was learned. It just seems to say as little as possible to move on.”

But the bigger problem I see at Google is its approach to developing those new things. Just because you enable it, or allow it, doesn’t mean your approach to you develop new products and services. And what strikes me about Google is that so many of these products seemed dead on arrival.

Om Malik had a great post yesterday that highlighted this core concern:

“Google, thanks to two brilliant engineer-founders, has become a great company seemingly able to solve the world’s most complicated engineering problems. That ability made it turn search into the great money machine that it is. It knows how to tweak machines and make them do unfathomable things. But what it can’t do is internalize empathy. It doesn’t know feelings. It doesn’t comprehend that relationships are more than a mere algorithm. You can see this in its many offerings; they’re efficient, but devoid of emotion, and emotions are what drive interaction.”

Jeff Jarvis, at Buzzmachine, had a post yesterday where he mused about Google’s failed product launches (skip past the Net Neutrality rant to the second half):

“Now Wave has had its detractors who are now cackling, but it’s not the specific platform that concerns me. It’s that Google can’t figure out how to launch new platforms. Wave was a bust. Buzz was a bust. Knol was a bust. Orkut was mostly a bust…

“I worry that Google isn’t an entrepreneurial company anymore. It didn’t start those platforms under the hard economics of entrepreneurship. And it hasn’t nurtured some outside entrepreneurs well. If it did, Dodgeball would be Foursquare today.”

So what’s going on here? There are four flaws I see in how Google approaches new product development:

  1. Failing to develop with users in mind. When many of these new products are launched, I find myself asking fundamental questions. What problems is this trying to solve. What need is this trying to fill? How will this fit into my life? Even if a product is new, the answers to those questions should be apparent. Often, I feel like Google products have been developed simply because someone was trying to solve an engineering problem, rather than trying to solve the problem of a user.
  2. Poor design. Flowing from No. 1, because the products seem to be engineer-driven, they seem overly complex and hard to navigate. If someone has to invest significant time in learning how to use your product, don’t expect them to do it. Apple and Facebook have been relentless in their willingness to simplify their user interfaces.
  3. Not iterating. When something doesn’t hit big, the approach should be to take the feedback from users, digest it quickly, and then adjust the product rapidly. Now, anyone reading this, please correct me if I’m wrong. But when a Google product doesn’t hit big, I don’t see rapid changes to it that suggest the team is learning lessons. This leads to the impression, as Jarvis notes, that they’re just throwing stuff at the wall to see what sticks.
  4. No follow up support. After the initial launch, it seems like a lot of new products fall right off the radar pretty quickly. It doesn’t seem like there’ s a widespread follow up from a Google evangelist pushing developers and users to work with new services that don’t catch fire right away.

There have been notable exceptions, Android being one. Though I would argue here that in terms of design and function, it largely mimicked an existing product, the iPhone, with some additional twists, such as being a more open platform.

As a user, my larger issue is that this all becomes an un-virtuous circle. With so many fizzled products, I’m wary of investing lots of time in new Google products, worried that the track record suggests the plug will be pulled sooner rather than later.

For now, that record of failure hasn’t yet wrecked Google’s popular image, its brand, or its reputation for being one of Silicon Valley’s most innovative companies. But it is time for the company to take a hard look at its processes before it gets defined by its failures, rather than its success.


Tags: , ,


Share this Post

  • Pleb

    I hold Google shares and… sadly have to agree!

    Can’t beat either Apple OR Microsoft at their respective games in my eyes and that’s two very different competitors for you. They have two things in common that GOOG just plain doesn’t share with them.

    a) A fearsome dedication to testing and finishing pre release

    b) A fearsome dedication to product support

    Get wise Google! These are not optional.

  • Dave

    This shouldn’t surprise anyone who’s ever gone through the Google interview process. Everything is centered around algorithms and being able to answer tricky mathematical questions. This makes for a great search engine, but it’s hard to see how building the whole company technical competence only will make many of their other initiatives successful.

  • I agree that the listed “four flaws” of product development should (and do) lead to product failures, for Google and others, but I think you haven’t considered the fact that Google’s practice of publicly experimenting with new product ideas in fact tests against these factors, possibly by design!

    Perhaps it is a bit unfair to talk about “Google’s record of failure” when what we are seeing is a set of early-stage product tests, visible at much earlier stages that most companies would show us. At least we see them; how many potentially great product ideas never see the light of day in more formal companies like HP or Microsoft? (Answer: many…)

  • Oldster

    But wait! You’ve missed the latest failure, changes to that were forced users at the end of June.

    User response has been overwhemlingly negative from day one. Have they reverted to the old version or at least given users the option to choose? No.

    Have they acknowledged that user rejection of the changes make the Classic Coke introduction fiasco look like a marketing sucess story? Nope.

    Gotta say, this establishes a new benchmark for corporate arrogance is the face of consumer rejection of change. What’s next? Google Bob?

  • Elgoog

    RE: Android
    “it largely mimicked an existing product, the iPhone” Mimicked. Now there’s a charitable way to describe what they did.

    RE: Google stuff
    The sole reason Google puts ANYTHING out for general use is to be a conduit to suck data into their cloud, where it feeds their ad machine. Wave didn’t attract the number of eyeballs and clicks they hoped, mainly because nobody understood it.
    Bottom line: The only reason Google makes a feeble attempt at customer support is to keep those clicks coming. They saw no future for Wave to generate lots of clicks, so they killed it. End of story.

  • bluevoter

    When creating new products, many tech companies and startups, including
    Google, assume that the smartest people are already in the room.
    Time-to-market and burn rate issues cause them to jump into development
    before they’ve validated the concept with potential users and kept them
    involved in the ongoing development process. Failure to do so often means
    that the resulting product doesn’t address a real problem in the world, and
    thus fails to create an enthusiastic user base that will not only contribute
    suggestions for future enhancements, but also evangelize the product to
    their network of friends, colleagues, and Twitter stream readers.

    Product definition and validation, as well as triage of user requests, is the
    responsibility of a product manager. So I would look first at the responsible
    product manager for a failed product, and then at the people who bring
    pressure on the product manager to skip steps in the development process
    and/or to release a product prematurely without a feedback mechanism and
    engineering resources to enhance the emerging product. Google’s
    Marissa Mayer talks a good game about prototyping, testing, and data
    gathering, but those steps are only effective if someone has first identified
    a real problem in need of a solution.

  • Paul

    What’s wrong with solving engineering problems? There are enough tech types out there that you might also be giving a sizeable group of users the product they want and will use. It’s tougher to add features than remove them, I’d suggest.

    The difficult part lies mostly in properly designing the user interface to present essentials. Leave the remainder as background settings – as idiot proof as possible – or reasonable, logical and conservative (if you’re going to err on a side) defaults.

    Expecting users to learn serious apps for ‘frivolous’ (for want of a better word) needs is, well, illogical. A feature-rich app is a waste of time these days. The instant gratification theme is almost better characterized as light speed gratification, it seems. That’s producing hordes of bored – and boring – people who haven’t the persistence or resources to think for themselves beyond their increasing self absorption. Pray that we never have to deal with something as serious as another Great Depression.

    While I’m about it, Apple and Microsoft as paragons of pre-release testing? You’re kidding, right? They may be BETTER at it by now, but they both have serious fiascos to their names. Google is far younger; it at least has a culture, I think, which seems to encourage experimentation. Can you imagine it being run by Steve Jobs? Didn’t think so. 🙂

    Thanks, nonetheless, for a thought provoking article and comment(s). Opinion pieces such as this are always enjoyable reading.

  • Paul

    P.S. Fix your comment counter. There are two labelled as comment 1. 🙂

  • I do not agree that knol has gone bust.
    Knol is thriving. There are many highly qualified and employed professionals writing actively on knol. There is very useful content which is accessed daily on it. As an author I know that my page views have doubled in the last one year. There are many authors whose page views have doubled in the last one year.

    Yes we thought a million knols would be accumulated by this year end but it did not happen. It will take some more time as Google has not supported the platform actively in promoting knol writing.

    We do accept that Google is not iterating enough to develop acceptable products and also it is not providing adequate followup support. Knol needs both of these and it has demonstrated its potential in attracting many professors and students. Knol is on its way to success. Unlike Wave, which many of us could not make sense of, Knol has number of groups working actively for the success of knol. Any other criticisms of knol are applicable all social media platforms and they are not special to knol.

    We as a knol author association, Knol Author Foundation are working everyday for the success of knol. We invite online authors as well as others to come and write on knol and welcome all to access articles on knol in a large variety of subjects.

  • There is no follow up PR/evangelism. Google doesn’t understand the value of marketing or PR because it found its initial successes organically and now that is rooted in its culture.

  • Whatever you say, but Google is going to rule the future!

  • One problem that immediately caused concern was Google’s decision to automatically give users a ready-made circle of friends based on the people they most frequently e-mailed. Unless users changed settings in their profile, this list could automatically be made public, allowing anyone to see who a user corresponded with most frequently. [P]rivacy experts immediately pointed out this could cause problems for journalists, businesses or even people having an illicit affair.