LeWeb – Path and Flipboard moving away from lean startup methodology

Dave Morin CEO of Path and Mike McCue CEO of Flipboard were both interviewed on the main stage yesterday at Le Web.  They are both high profile CEOs of high profile startups that have raised a lot of money for very design focused mobile apps.

They have also both recently launched new iPhone apps and we had demos of both.  Morin and McCue were both at pains to show how beautiful their apps are and how much time and effort they had put into making the interface intuitive and easy to use.  They both also explained that they had wanted to get the apps right before launching and had preferred to take their time and get it right rather than launch early with an imperfect product.

Morin explained how with the first version of Path they had launched early with a minimum viable product but they found that iteration cycles on mobile were too long to make the lean methodology work, which was why they switched methodology for the second release.

So we have two leading companies that are moving away from the lean startup methodology just at the point when the rest of the world is adopting it as an orthodoxy.

I haven’t figured out what to make of this yet.  Path and Flipboard have both raised huge amounts of capital so they have the luxury of not being lean if they don’t want to, and that might explain why they are adopting a different approach to everyone else.  Alternatively, it might be that as great design becomes more and more important there is less and less to be learned from minimum viable product.  Put differently, if great design is the essence of the product there is no way to quickly develop a rough version.

The point about iteration cycle times on mobile is also interesting.  I’m assuming that is because of app store approval processes, and if we move back towards a more web oriented world that will become less of a problem.  And if we don’t, it won’t.  I’m not sure which way that will go, but if Morin is right the outcome will have a big impact on the way startups approach product development.

Enhanced by Zemanta
  • John Connolly

    Thank you for highlighting this. I think it is worth startups drawing a distinction between a product’s functional scope and the level of design polish/ transaction integrity. The Minimum Viable Product mantra is, I believe, most valuable in terms of scope, but maybe less relevant in terms of ‘polish’ when it comes to either i) design led differentiation or ii) mission critical applications (e.g. in payments, a pet hobby of mine). Some applications just aren’t well suited to quick & dirty (artificial organs, security etc…).

  • Nic,
    This is not that surprising considering that both companies have now created a brand associated with great design and quality. it’s ok when you are building your product and still are in a beta phase to launch something that still has flaws. Once your product is known and used, customers are far less forgiving. So the lean startup methodology is still applicable, but applies for the first iteration of your products. When extending your product offering to new platforms (e.g. smartphones after an initial tablet launch) you can ill afford to get it wrong. To summarise once you have build some equity in a brand, you need to fulfil expectations.

  • That logic works for companies like Flipboard who have a product with a large user base, but maybe not for Path who have still to find their sweet spot. And even with Flip board I wonder if there isn’t a way they could have tested demand for their iPhone app more cheaply.
    Thanks for a great comment.

    Sent from my Android phone using TouchDown (www.nitrodesk.com)

  • Gordon Guthrie

    Are they moving away from lean? Or is it that they are not doing public MVP’s and iterating in private?

    It also depends what you mean by MVP. I am a bit Old Skool and think the P is a product that people should pay for, but the ‘industry standard’ is any artefact that can be used in an experiment, be it wirescreen, landing page, user survey, etc, etc.

    The other critical thing is ‘what is the company currently testing’. If version 1 clearly shows that there is a market for your product, then why would you retest that in version 2?

    I think there probably is, and should be, a revolt against ‘pop lean’ – the sort of lean that Eric Reis talks about when he says ‘launch it and see what happens is an experiment that always produces a positive result’. In other words, when you launch it, something indeed  happens.

    I also think there are systemic problems with the term ‘lean’. It really should be ‘leaner than X’ – in other words a measured lower cost base on a like-for-like basis.

    Just because they are funded doesn’t mean they aren’t or can’t be lean relative to the competition.

    For straightforward stuff like online travel search versus high street travel agent the ‘leaner than’ approach is pretty straightforward. But when I see a bazillion iPhone app companies doing apps with teams of 3 devs and a time to launch of 12 weeks, and identical spec Macs as their dev platform, and all proclaiming they are lean, I always think, hmmm…

  • JP

    Nic,

    Couldn’t it be just a case of knowing what users want and not needing to wait for their feedback? Although they might be “violating” principle #3 of lean (decide as late as possible), they might be improving principle #4 (deliver as fast as possible)!?

    JP

  • If you really know what the users want then I agree the best thing is to build it as quickly as possible. If you have a strong conviction which relies on unproven hypotheses the lean startup methodology has it that you should launch early to test those hyptheses as cheaply as possible.
    I think Path definitely falls into the latter camp and Flipboard on the iPhone (rather than iPad) probably does too.
    ________________________________

  • Is it possible to iterate this sort of app in private?

    I would have thought you need a large number of users to learn anything of real significance.
    And neither of these companies charge users…

    ________________________________

  • Gordon Guthrie

    > Morin explained how with the first version of Path they had launched early with a minimum viable product but they found that iteration cycles on mobile were too long to make the lean methodology work, which was why they switched methodology for the second release.
    That’s a fairly restricted view of lean – the fast release cycle. Both Stephen Blank and Eric Reis are systematizers – Stephen Blank took things that had worked from his 8 startups and common activities from companies he interviewed to define Customer Development. Eric took technical activities like Agile and integrated them.

    So lean includes a lot of stuff that people were already doing, but ‘named the parts’ and took things out of an ‘intuitive’ context and into an analytical context.

    The things that they systematized can be loosely divided into:
    * validation
    * execution excellence

    both of which are terms I know you are familiar with from long before the Toyota Production System hove into view.

    But when people talk about ‘lean’ they tend to mean doing the new stuff – in fact just doing the old stuff properly remains half the battle – I suspect both Path and Flipboard do the old stuff well.

    Lean is important because it makes it easy to discuss your quality and to communicate amongst your organisation. That will have the longest impact I think.

    In relational to the particular context of iPhone apps I think Morin is right. Releasing 50 times a day, using self-healing clusters and release-roll back, as well as other techniques like split testing don’t work because of the approval cycle of apps. You would be putting a saddle on a pig really.

    Its perfectly possible that Path do continuous releases of software to an ‘Always Be Working’ jailbreak source repository that they feed to their dev/testing machines and then eventually polish up for a formal release, who knows.

    But once you have a working application that has a measure of traction, some understanding of your successful demographic then you can change your testing regime. At the beginning your hypothesis are crude and broad-brush, but once you get patterns of use you start filling in your knowledge and get more precise in your testing and more focussed as well.

    Producing really nice handling, beautiful apps like Path probably involved a lot of closely observed usability testing – that, being heuristic, only needs a group of 5-8 suitably qualified users.

    Anyhoo, that would be my take on it.

  • Nic: sorry no, “the rest of the world is not adopting it as an orthodoxy”. Without proper data to work with, non-manufacturing Lean is just slideware, and in the eyes of nearly every angel and VC out there it remains practically unfundable. How do you fund a startup who knows nothing and doesn’t know how long it’ll take to find out something?

    In my opinion, the main people talking up lean are celebrity VC bloggers, none of whose funds would actually invest in Lean Startups (sure, they’d individually take a punt on an individual entrepreneur, but that’s quite a different thing), and all of whom should know better than to follow the herd of software engineers who’d like Lean to be true.

  • When Ries et al talk about minimum viable product, they do explicitly say that depending on your product and market, MVP doesn’t necessarily mean a low performance / bad design product – it could be that for what you’re doing, you need a very high level of design to really test the value. The lean methodology is all about doing the minimum needed to learn what really adds value to your customers. I wasn’t at the talk, but it sounds like Morin is essentially saying that because they couldn’t do really fast iterations (a technique) they’re just going to scrap the whole methodology – of learning what your customers want. Path is certainly beautifully designed, but I still can’t really see the core problem it’s solving or its core value proposition; and unless they get to the bottom of that I don’t think the design can really make much of a difference in the long run. On a related note, start-ups where ‘pure design’ is the reason for success tend to be in markets with a mature value proposition. For example, people rented rooms on gumtree, craigslist and random websites for ages. The value prop was clear. Then AirBnB came about with an awesome design and user experience and blew the others away. FlipBoard could fall into the same definition; there’s an established value prop around reading media. The history of Apple can even be viewed in this way, to an extent – they were relatively niche (big niche) until personal computers got to the pure commodity stage, and ‘power’ stopped really meaning anything – pretty much any computer you buy can do pretty much anything you want – and so their enhanced design and user experience is leading them to grow faster than ever before. The flip side is that I don’t believe design can do much for you before the basic value prop is proven; and in early stage markets just having the core functionality that people need is the most important. 

  • The correlation between design and maturity of value prop is a very interesting idea. One could even use that as a way to think about markets generally.
    Back to MVP, I can see that testing the hypothesis that design is key to a product might be difficult to do quickly/cheaply.

  • You are right that lean is over used as a term, and describing it as an orthodoxy was an overstatement. That said there is much more to it than you allow. I think investors everywhere are backing startups that follow the principles of lean, and and many of them have moves far beyond the slideware stage. Conversocial, a company that we invested in recently, is a good example.
    Sent from my Android phone using TouchDown (www.nitrodesk.com)

  • Thanks Gordon

  • I’m not sure how pretending to know your customer base, exact product, and complete financial picture for the next 3 years is more fundable than just admitting that you’re unsure. The old model was all about precision, not accuracy. The new model admits we don’t know. Yes, it’s uncertain. No, it’s not unfundable.

  • Pingback: Eric Ries: In The Founders' Studio at Leancamp London 2 | Lanyrd()

  • Pingback: Eric Ries answers questions on the application of lean startup principles in some edge case scenarios « « The Equity KickerThe Equity Kicker()