Leave a comment
Get the GH Bookmarklet

AMAs

Sangeet Paul Choudary is the founder of Platform Thinking Labs, a C-level advisory firm on platform dynamics, an Entrepreneur in Residence at the INSEAD Business School, the Co-Chair of the MIT Platform Strategy Summit at the MIT Media Lab, an advisor at 500 Startups in Silicon Valley.

He is also best-known for his writing on the Platformed blog where he analyzes platform business models and network effects.

He advises Fortune 500 firms as well as high growth startups on designing platform architecture and strategies for creation and governance of network effects. 

He serves on the advisory board of OneVest – CoFounders Lab, the largest online community of entrepreneurs globally. He has mentored startups at MassChallenge (Boston), Bombacamp (Chile) and JFDI (Asia), and is a Fellow at the Nanyang Technopreneurship Center in Singapore. He is also an advisory board member of the SaltGrid Initiative which aims to create a reputation layer for the emerging platform economy.

Sangeet was formerly the head of innovation and new ventures for Intuit Asia-Pacific and was responsible for managing Yahoo’s largest M&A deal in South Asia, an investment in a two-sided marketplace connecting local businesses with consumers.

Also, download his book, "Platform Scale" for free on Amazon this week!

  • SE

    Sean Ellis

    about 2 years ago #

    Thanks for doing an AMA with us. What do you think is the best way to think about a funding strategy for most network effect businesses. Is it generally better to try to raise a lot up front to aggressively build the network or is it better to let it develop organically and raise smaller amounts over time? Part of the inspiration for this question is the subtitle on your book "How an emerging business model helps startups build large empires with minimum investment." I haven't had a chance to read the book yet, so I may be interpreting the subtitle incorrectly.

    • SP

      Sangeet Paul

      about 2 years ago #

      This is a really good question and I'd say this depends a lot on the network.

      1. If you can kick-start the network on standalone mode, you need very little investment. E.g. Instagram

      2. If you are building a connection-dominant network (Section 3.1 from the book) where users need to connect before they interact, the cost of buiilding these connections and experiencing user level network effects is incredibly high. E.g. Facebook, LinkedIn. If you were to build one of these today (arguably users are much more distracted today than a decade ago), I'd recommend raising some serious cash.

      3. Irrespective of what you build, your intiial investments to network effects depends on the critical mass you are gunnign for. Critical mass is the minimum size of the network at which interactions happen reliably. One definite way to minimize critical mass is to target a niche. Hence, figure out how you can go narrow in a way that interactions definitely take place.

      4. I talk about the Uber-Lyft battle in the book and why that is unique compared to all other network effect businesses. That is a classic demonstration of the role that multi-homing costs play. If one or either side of your network can exist on multiple networks simultaneously, you have a real problem. This is what happens with both sides on Uber and Lyft but it's a problem that LinkedIn, for example, doesn't face. If multi-homing costs (costs to exist simultaneously on multiple networks) are low, your investments to create exclusive user stickiness and hence defensible network effects is much higher. In all such cases, network effects don't choose a winner, burn rates do. As a result, you'll note that Uber has only one major competitor in every geography because it is inefficient to fund more than one competitor in the long run. It's fascinating but if you tease out this one concept, you realize why certain ideas are inherently more difficult to execute from a network effects perspective. Logistics may soon have a similar challenge. Social interactions have already seen this play out as Facebook experienced competition from unrelated competitors like Whatsapp and Instagram because multi-homing costs were low.

      5. One final thing I'd consider, especially on a two-sided network effect, like a marketplac, is whether you are building something where people discover each other and move out (If I find a good woodwork guy on TaskRabbit, I want to stick with him) vs. those where people return to the network for repeat transactions (since I likely travel to a new location every time, I still need Airbnb irrespective of how good my last experience was). In the first kind of model, a success for the user works against success fo the network. Hence, user LTV is low and such networks often need not just high investment but also further sources of value for users to stick around. Someone who has done it well is Upwork. If I find a great designer in the Philippines, I don't take my work off-netowrk but use Upwork's work monitoring tools (additional value) to transact on the platform.

      All the above ideas are discussed in detail in Sections 3 and 4 of the book. You should check it out for more detail there. But great question to get started!

      5 Share
      • AL

        Arsene Lavaux

        about 2 years ago #

        Merci @sanguit, this is inspiring.

        Regarding 2., isn't it a catch-22?

        Don't you need traction, that is, proof that you are building a network with 2-digit month-over-month growth to be able to convince early investors to invest serious cash in your venture?

        Looking forward to your thoughts on this point.

        • SP

          Sangeet Paul

          about 2 years ago #

          That's why #2 is the most difficult to do. If you have some form of access to capital, go for it. Don't go for the slow burn. But if you don't then you have a higher likelihood of being one of the many companies out there that tried to build the next FB or LI but failed. Connection-first networks are incredibly difficult to execute, especially when the connection requires 2-sided opt-in.

  • AL

    Arsene Lavaux

    about 2 years ago #

    Hi Sangeet: Thank you for leading this AMA.

    Imagine you are looking to create a two-sided eCommerce marketplace on a C2B model (yes, I mean C2B and not B2C, imagine consumers aggregate their demand in real-time to attract supply).

    What would you focus on to crack early stage growth, which side of the marketplace? Or anything else as a lever of scalable growth?

    • SP

      Sangeet Paul

      about 2 years ago #

      1.Can I aggregate demand long enough for a sizable mass to build up? If so, I'd aggregate demand.
      2. Can I get the first marquee supplier on-board without any commitments? If so, I'd target supply.

      3 Share
      • AL

        Arsene Lavaux

        about 2 years ago #

        Thank you @sanguit !

        Have you experienced a C2B business with both?
        What do you focus on then for such business?

        Fascinating subject. Thanks for your expertise.

        • SP

          Sangeet Paul

          about 2 years ago #

          I've seen the latter (model 2) happen quite often if you have access to a marquee supplier. Model 1 is more difficult but the way a lot of marketplaces do it today is hook users around something and then bring in the supply side. Where you're going slightly differently, I believe, is how do you aggregate demand for a short period without having to hook users over a longer lifecycle. My only concern with that is that I am a big believer of LTV economics and if you believe that users will come in only transactionally and lave, your LTV may not be too high.

          • AL

            Arsene Lavaux

            about 2 years ago #

            For Model 1, if aggregation delivers something truly unique (because consumer targeted demand aggregation is a supplier's dream and can cause them to provide that unique value), something the target user couldn't get otherwise, then I believe you are much more likely to create stickiness to your platform. And LTV as a result.

            Actually, new forms of product or service innovation can emerge in these networks. The network of the future may actually redefine our current vision of product-market fit.

            The alignment between the two should become more agile. And much more cost effective (lower R&D cost, lower marketing cost, lower overhead cost etc...).

            Am I drinking too much French wine?

            • SP

              Sangeet Paul

              about 2 years ago #

              French wine can never go wrong. :) All that you're saying above makes sense. I view networks as layered value transfer. Once you create a network, you can build all sorts of new value interacitons on top of it. So, largely in agreement.

  • BW

    Brand Winnie

    about 2 years ago #

    Hi Sangeet!

    How would you validate product/market fit in a network effect business? How is this different than for a non-network effect business?

    • SP

      Sangeet Paul

      about 2 years ago #

      This is tricky because you don't often have just one market to evaluate the fit with but two. To the extent possible, you need to stage this and see ifyou can provide a standalone service to one, validate it, hook them in and then test the other side.

      However, there are some cases where this becomes easier:

      1. IF fit with one market is more important (e.g. the other side is a contract workforce when you start), then P/M fit remains largely similar.

      2. If you can focus only on one market and allow them the ability to bring in the other side, then you need PMF for only one side. E.g. Kickstarter

      3. If you can focus on a market segment that has both your roles, then you can validate both roles while focusing n one market. E.g. Etsy's initial buyers and sellers - both sides - were craftspeople. Easier to validate.

      4 Share
  • XZ

    xikai zhao

    about 2 years ago #

    Hi Sangeet,

    Thank you for sacrificing your time to do this AMA. My company has this mentality to AB test everything. However, some of us think AB test everything is going to slow us down. How do we balance this two? Any wisdoms? Thank you

    • SP

      Sangeet Paul

      about 2 years ago #

      I have a rule of thumb to A/B test things that work a bit. A/B testing is good for optimizing a currently working experience, not good for finding what users want when you don't know what they want.
      Nothing substitutes real world user feedback. If you feel you've built out a working user experience, go ahead and improve it by A/B testing. But before it starts working, spare yourself that effort.

      Of course, if yhou're building something that is standard (like booking airline tickets), you knwo what already works because those experiences have become standardized over time so go ahead and A/B test. But I'd reckon we have less and less of those kinds of products coming up today.

      2 Share
  • DL

    Dylan La Com

    about 2 years ago #

    Awesome to have you here Sangeet!

    Borrowing this question from Peter Thiel–

    What important truth do very few people agree with you on?

    • SP

      Sangeet Paul

      about 2 years ago #

      The fact that not everything will become a pure play platform or a marketplace but that every business (no matter what) and every individual will need to employ elements of platforms and networks to do what they do.

      In tech, we still see this as a small category in a corner. I believe that networks are fundamentally going to change every world system that we've built and will create many new ones.

      That's what I'm excited about, and about the fact that we are at the very, very early stages of this transformation.

      4 Share
  • JM

    Jason Meresman

    about 2 years ago #

    Hi @sanguit - thanks for today's AMA!

    What in your view are the best examples of B2B startups that leveraged network effects to grow and why?

    • SP

      Sangeet Paul

      about 2 years ago #

      B2B application platforms: Salesforce and SAP Hana
      B2B commerce: Alibaba, Mercateo
      B2B supplier ecosystems: Tradeshift, SPS Commerce (These companies are moving towards building network effects)

      In general, B2B network effects is more about signalling and reliability by curating quality than about providing abundance and choice by collecting quantity.

      2 Share
  • AL

    Arsene Lavaux

    about 2 years ago #

    Une autre petite question @sanguit ?
    :)

    Are there any examples out there of two-sided eCommerce marketplaces, let's say B2C, where picking one vertical on the supply side substantially reduced customer acquisition cost?

    If so, was this impact somewhat quantified?

  • KA

    karim Abd El Kader

    about 2 years ago #

    Thanks a lot @sanguit for taking time to do this AMA.

    I really looking forward to read your beautiful book; i find it very interesting to do a transactional marketplace in Egypt whereby women can buy and sell their used brand clothes using my platform and i get a commission from each transaction and i will work delivering the items to buyers:

    - Please tell me what do u think about this business model and what are the advantages, flaws of it as well as what are the key factors and dynamics that i should put emphasis to when it comes to building such business.

    - Also, it will be great tell me other rising and innovative digital business models related to fashion and clothing.

    - What is the best acquisition, retention and virality strategies for such platform

    Many thanks,
    K.

    • SP

      Sangeet Paul

      about 2 years ago #

      I can't answer some of your questions in a lot of detail without understanding your business model better first.

      But as with all used goods, curation is going to be the most important thing. Your job isn't just delivery, it's also centralized vetting and curation. It's also helpful to create this around use cases. Some high end wear is bought just for speciic occassions. If you can identify some of these occassions, you can beenfit from lot of external liquidity.

  • AA

    Anuj Adhiya

    about 2 years ago #

    Hey @sanguit SaltGrid sounds like a really cool idea. Can you talk more about how a reputation layer for the platform economy would work and its importance?

    • SP

      Sangeet Paul

      about 2 years ago #

      Open platform economies create transparencies. One form of transparency is aggregating and making options available. But a bigger transparency is indicating quality of what's available. This is where a reputation layer becomes very imprtant. Think of it as TripAdvisor or Yelp but based not on social inputs but on all kinds of signals that you can track based on what users do.

      2 Share
  • AP

    Anthony Panepinto

    about 2 years ago #

    Hi @sanguit, thanks for doing this AMA.

    I've created a double-sided marketplace app that helps event planners discover top quality venues daily and streamline communication with venue representatives. (www.venuevortex.com).

    I've successfully on-boarded the supply-side (venues) but I'm having trouble with the demand-side (users). What lever would recommend I pull to un-leash growth prospects?

    We're targeting event planning professionals, and non-professionals alike. Essentially anyone who is planning an event (conference, birthday, business dinner & banquet, product launch, wedding etc...)

    Cheers!
    Anthony

    • SP

      Sangeet Paul

      about 2 years ago #

      It's a bit difficult to answer this question based on marketplace mechanics without understanding what kind of feedback you've been getting from the demand side. Why are they not interested?

      • AP

        Anthony Panepinto

        about 2 years ago #

        @sanguit It's not that they're not interested, it's that I haven't gotten the product in front of enough of them to generate a consensus. So the question is really eluding to how can I get in front of more eye balls!?

        • RT

          Raul Tiru

          about 2 years ago #

          Hi Anthony, I hope you appreciate my feedback too :)
          If you're looking for more traffic, I would start with SEO. The are many things you could optimize on your website.

          In my opinion, you have a very likable product; utilize social media could get you in front of the right people.

          Work together with the venues and do some co-promotion.

          Hope I could be of help :)

  • MS

    Michael Sadler

    about 2 years ago #

    Hi @sanguit,

    I built a platform that matches buyers and sellers of bulk off market real estate and am wondering whether it is going to scale.

    Right now brokers are saving portfolios of property for sale and buyers' criteria and they are being matched.

    They want a widget for their website that will save information Renturly.net for matching, as well as a way to upload properties from a spreadsheet.

    Is this going to scale?

  • HQ

    Hila Qu

    about 2 years ago #

    Hi Sangeet, thanks for doing this AMA.

    From your experience, do you see any different dynamics between the two sided marketplace/platform business in Asia and in US?

    • SP

      Sangeet Paul

      about 2 years ago #

      One of the biggest difference that I see is when you're targeting the SMB markets in South Asia as one side of your marketplace (retailers, local service providers etc.), you often have to invest more in concierge and hand-holding than in the US. The Do-it-for-me attitude is more dominant than the self-serve attitude. A classic example: When Google launched AdWords in Asia, a whole new set of agencies sprouted up that didn't have the media connections that traditional agencies did but defined their services around running Adwords campaigns for SMBs and evolved from there. This demonstrates the higher need for concierge in South Asia.

      2 Share
      • AA

        Anuj Adhiya

        about 2 years ago #

        This "do it for me" attitude is prevalent across interactions in SE Asia - ie a cultural trait.
        In some arenas though, I see this changing where a small audience is willing to DIY it. Do you think that observation is correct?
        And if yes, do you think there is value/merit in constructing interactions to accelerate (at whatever pace) that trend in a 2 sided marketplace? If yes, what strategies might one employ to make that happen?

        • SP

          Sangeet Paul

          about 2 years ago #

          That observation is correct. I'd say that moving people from DIFM to DIY is a matter of demonstrating proof of concept. If you can demonstrate proof of concept with a small DIY base, then the DIFM base will also come on board. That's what happened with AdWords in the US.

          2 Share
  • HS

    Himanshu Singhal

    about 2 years ago #

    Hi Sangeet, Thanks for hosting this AMA
    What do you think is the best strategy to scale/grow in a C2C marketplace- buying and selling used cars- with a high ticket size, perishable supply and very low customer repetitive behavior?
    Also, it would be great if you can give examples of startups who have successfully scaled their business model with the above characteristics.

    • SP

      Sangeet Paul

      about 2 years ago #

      Used car is one of those really unique markets where curation trumps absolutely everything. I would focus on figuring the curation model out and figure out how to scale curation because that is the bottleneck. If that doesn't scale, the market won't.
      An exact parallel doesn't come to mind. The market for babysitters (SitterCity) is high curation but also reasonably high repetitive consumption so it's not a perfect parallel. One option is to try asking the sell-side to pay to curate it in exchange for getting it liquidated faster. Something like this has been tried successfully in B2B application markets where enterprises have lesser time and willingness to experiment with unknown apps so app creators pay the platform to curate and if it's good, feature the app.

      2 Share
  • KS

    Kamil Szybalski

    about 2 years ago #

    Sangeet, what books are you currently reading and what's on your bookshelf? Could you take a Shelfie for us - http://Shelfie.com

    • SP

      Sangeet Paul

      about 2 years ago #

      I like to read wide (I skim 5-8 books a month) and take from every book the slice that I can internalize and make my own.

      I am currently reading Who gets what and why which is a great book on real world marketplace and clearinghouse design (financial, education, labor and healthcare markets) and is fascinating background to how markets work in general without technology facilitating anything.

      I am in a bit of a behavioral economics phase at the moment to understand why users act the way they do so I've also got Phishing for Phools and Misbehaving on my list to go to next. I want to re-read Addiction by Design after I get through these to see what else I can internalize from that book.

      2 Share
  • SP

    Steven Pesavento

    about 2 years ago #

    Thanks for joining us for this AMA!

    In your book you talk about Interaction first design over technology focused design. Would you mind elaborating on what that means for the GH Community and how they should apply that when building their own platform business?

    • SP

      Sangeet Paul

      about 2 years ago #

      I'm glad you asked this. This has a few implications.

      1. Whenever growth is part of the product experience, stop hacking it and start designing it. My concern with too much growth hacking is that in the long run, all well-intentioned advice reduces to tactical step-wise ideas which miss the intent but follow the word of the idea. I specifically talk about it in Section 5 where I talk about the need for designing virality.

      2. This has bigger implications in the case of platforms because unlike SAAS, platforms and networks and marketplaces aren't software businesses. They are businesses that rely on users using software to interact with each other. Much like television is a content device but telephone is an interaction device. In these case, you need to think interaction before you think technology.

      3. We talk a lot about solving a pain point and doing customer development but I'd say we take that a bit further with platforms and ask ourselves, how can we simultanelously solve 2 pain points of different user groups so that the interaction happens. THis is what I devote all of Section 2 in the book to. This is where building a network effects business is very different from buiding a product. You can create a great user experience for one particular user role and still fail at the interaction if you did not focus on all the roles participating on the platform.

      4 Share
  • AS

    Alex Sherstinsky

    about 2 years ago #

    Hi Sangeet,

    Thanks for doing an AMA with us.

    What team composition do you recommend for a company to turn their product into a marketplace/platform?

    • SP

      Sangeet Paul

      about 2 years ago #

      The most important thing to look a when making this transition is what does marketplace governance mean for you? It might be curation mechanisms, trust creation, insurance etc. There's a whole chapter in the book on the 7 Cs of Trust. I'd like to figure how I get moving on those. This is the single biggest transition when you move from product to platform. Tech challenges are manageable. Data challenges scale so you may need more data resources but the big challenge and new skillset involved has to do with marketplace governance.

      3 Share
    • SN

      Sin Nalisa

      about 2 years ago #

      I read your book, I was actually one of the first to receive it via PDF because of the pre-order. Great investment!!!

      Question 1: Is it possible to have a 3 Sided Marketplace?
      currently: our platform helps Musicians book shows with Performance Venues. But Musicians need to activate their Fans when pre-selling tickets. Are we technically correct in assuming we are 3 sided? Can you name examples of other 3 sided marketplaces?

      Question 2: Bill Gurley at Benchmark is an advocate for marketplace startups. He insists that the best way for startups to optimize pricing is to take the most minimal percentage possible. That way it fends off competition and allows faster onboarding of users. What are you thoughts on this subject?

      • SP

        Sangeet Paul

        about 2 years ago #

        Great to hear that! On your questions:

        1. You can have more than 2 sides BUT you should know the dominant interaction that makes your market work. If you can't choose one dominant interaction, execution is going to be a nightmare. In your case, it sounds like the dominant interaction is between musicians and venues. Without that succeeding, the secondary interaction between musicians and fans won't take off.
        e.g. many real estate markets have distinct roles for sellers, buyers and brokers but the dominant interaction is between brokers and buyers.

        2. Yes and no, pricing should depend on your ability to retain the transaction and that depends on multiple factors:
        a) are you a one-time or a repeat transaction platform?
        b) does every transaction occur between new people or are repeat transactions between the same people the norm?
        c) Do you provide enough value during the transaction once the two sides have met so that they find it more expensive to leave the platform than to stay on-board?

        Answer these questions to determine how low you should go. Moreover, be open to non-transaction models. E.g. Sittercity works on subscription because they figured that works better for a curated marketplace.

        2 Share
        • AA

          Anuj Adhiya

          about 2 years ago #

          To your point about "are you a one-time or a repeat transaction platform?" - are there any additional pricing considerations if the repeat transaction frequency is very low, maybe 2-3 times a year/user?

          • SP

            Sangeet Paul

            about 2 years ago #

            The bigger problem with 2-3 times a year is that users will forget you unless you hook them through some other mechanism like a very lucrative loyalty program (expensive) or a more organic form of ongoing engagement.

            • AA

              Anuj Adhiya

              about 2 years ago #

              Is the a parallel to be found or lesson to be learnt from say the real estate example you provided?
              Most people will not need a real estate broker more than a few times in their life, let alone in a year.
              How do they retain users and transactions? Is it simply a function of ad budget in their case to achieve the stated goal of staying top of mind?

  • AA

    Anuj Adhiya

    about 2 years ago #

    I know you've said that in some communities frictionless participation is not a good thing. When is it a good thing?

    • SP

      Sangeet Paul

      about 2 years ago #

      When you can regulate participation in some other way. E.g. Twitter (140 characters, opt-in and follow me). The other extreme is the ChatRoulette example that I give where participation is anonymous and interactions are randomly assigned. Worst scenario for frictionless.

      2 Share
      • AA

        Anuj Adhiya

        about 2 years ago #

        So is it correct to say that it may be a good thing when you can regulate participation AND when it's a "general purpose" community vs a niche one?

  • RT

    Raul Tiru

    about 2 years ago #

    Hi Sangeet, thank you for your time :) I have a few questions, if you don't mind.

    1. Hypothetical, If you were to start a platform, what kind of platform would that be?

    2. How come you don't allow comments on your blog?

    3. If it's not too much to ask, which of the following are the same and which are different: Platform, social network, market network, multi-sided market.

    (referring to: http://techcrunch.com/2015/06/27/from-social-to-market-networks/#.opkrvz:LQeR)

    Thank you so much :) I'm happy for you that you've finished your book. Great read.

  • AC

    Abhijith C

    about 2 years ago #

    What is your view in terms of an Open Platform and a Controlled one? Open Platform being any user can post anything (ex. Facebook, Twitter, Instagram etc) and controlled being where only few "members" can post content and the general user can only interact with it (ex. Product Hunt).

Join over 70,000 growth pros from companies like Uber, Pinterest & Twitter

Get Weekly Top Posts
High five! You’re in.
SHARE
31
31