Wednesday, May 12, 2010

Minimum Viable Product - Automate only when needed (especially for Market Place Businesses)

These days, most entrepreneurs seem to like to start a marketplace business on the web. (A marketplace is something which connects "people who need" with "people who have" or in other words buyers with sellers).

Recently I have been talking to a lot of people from the entrepreneurial community who are bootstrapping some kind of a marketplace business and one interesting fact I observe is that most of these web entrepreneurs (90% of whom are also creative non technical entrepreneurs) seem to be in the mindset of building a web application of their ideas rightaway with all the features. People seem to have forgotten what a MVP (Minimum Viable Product) needs to be.

I agree that these days building a dynamic web application is so cheap, but still are we not forgetting something here? Isn't the fundamentals of bootstrap or lean startup is to demo, sell and build?

The idea I would like to convey here is "Automate only when needed...Do it first manually and understand the pain points. Your MVP (Minimum Viable Product) need not have automation".

To explain this, I would like to give you a simple example of a idea of building a web startup like Craigslist (again a marketplace business) where people post items for sale and connects buyers with sellers. This is how I would bootstrap this marketplace business:

(1) Create a simple static website which looks like the listing page of the marketplace where you list the items posted by the sellers. You can use drag and drop websites like weebly or snappages to create this listing page or even it could be as simple as a wordpress blog.

I know the question that pops up in your head: But I need a user login mechanism to allow my sellers to post their items. Well do we really need this? I mean, do we really need this now? Isn't that just automation.

First test the concept manually. See who are your potential customers and if they are willing to pay for your service.

(2) Just have a button next to every listing on the static listing web page created in step (1) and attach the button to a static html form which allows the seller to type information about their item for sale and upon hitting submit, the form contents are emailed to the entrepreneur. You can use form builders like wufoo to do this.

(3) Now at the backend, manually look through the emails and the entrepreneur connects the sellers with potential buyers.

The idea is to do all this connecting sellers-with-buyers manually behind the scenes so that neither of them know that this process is being done manually. All they know is that the website is damn slow in providing responses. But hey, if you connect the sellers with the buyers and if they are happy, they will come back to your website. Remember, your business is not the website but the service of connecting buyers-with-sellers.

At a point where you really start feeling the pain of the manual process, then start working on automating it.

Even better if applicable sometimes just carry out the whole process through email conversations (ignoring steps 1 to 3 altogether).

In the process of doing it manually, you build relationships with customers, understand the pain points which need to be automated and best of all you spend less money/time validating your idea. If you don't get traction of your idea when you did it manually, find whats the problem with the idea/execution. Don't focus on automating the concept.

Always remember to take the right action at the right time.

The ideas I have mentioned here could be extended or applicable to any kind of startup.

Related Posts


  1. I agree whole heartedly, partially because this is what I'm doing with my current venture. I setup a static website explaining what it does, sign up buttons that are currently going to a Chargify hosted page, and I act on the emails I get when people sign up. The biggest surprise here is, that people do sign up, the volume of sign ups is not high enough to automate any thing yet, it takes me 10 minutes to provide the service people signed up for. The process isn't automated, it's just a checklist that I can go through very quickly.

  2. Great article. I think many people are tempted to flesh out marketplace features on a site for a couple of reasons.

    You mentioned the first one - the cost for doing so is fairly cheap (or is it?). Getting a basic feature in place is one thing, but there is added cost to make the feature flow as part of a unified user experience.

    The second reason is a desire to want to appear as professional as the competition. Bootstrapping principles counter this idea, but I think it's sometimes hard to overcome this bias.

    It makes we wonder if, as an entire product *category* matures, are Minimum Viable Products as effective? If a person is recreating an existing marketplace, do they stick to an MVP? Or is it just for new services that are differentiated from the competition?

    Your post brought to mind Freecycle, which does a good job of using simple email technology and community self-management to conduct a marketplace.

  3. While I agree in principle to what you said, there may be certain circumstances when this may not apply.

    Taking the example of a marketplace, think about The differentiating factor with odesk is tracking the work done by service providers and following a pay for hours worked model. There were other online sites which functioned as marketplace for developers/freelancers. But odesk's hypothesis or leap of faith was that hourly tracking and payment model would work. To test this hypothesis odesk had to develop/automate the process.

    The point is in certain cases, you look at existing companies (analogs/antilogs) and get a fair idea of what works and what doesnt. If your leap of faith requires you to automate, develop some features then you should go ahead and do so. What do you think?

    PS: If you have read "Getting to Plan B" by John Mullins & Randy Komisar, you would better understand the point I am making.

    These were what could be called Analogs (read "Getting to Plan B" by Randy Komisar, John Mullins)

  4. Thank your for your comments Vid, JM and Shashi.

    JM - In my opinion, as the entire product category matures, the need for MVP becomes more important because the pie is shared by more people and you need to make sure whether you have a chance. You asked an interesting question: Are MVPs good for new products/services or for re-creation of existing products/services. For new products/services, the need for MVP is more pressing because one needs to first validate their assumption for the need. And MVP gets us that data with the least cost. For the second case of re-creation of existing products/services, one needs to make sure that the existing product/service is not widely known already and there isn't too much competition already. If that's the case, MVP could have all of the features in the existing product or it could be a copy of the existing product (remember good artists borrow and great artists steal =) ).

    Shashi - I agree to your comment on the need for automation if your leap of faith requires you to automate and develop features. However, you need to keep in mind that as an entrepreneur/business man, you should try to first sell it before you build it. Or atleast validate your assumptions for the need before you go ahead and spend your resources on automation and building completely.

  5. I agree with the concept in principle, but I am not quite sure I agree with extending this concept to the example cited. Yes, it makes no sense to perform more automation then is required or to digress from the MVP when setting up a new website but at the same time it is important that the boundaries framed when defining the MVP are not too narrow thereby reducing the USP of the product you are selling.
    I would think that transparency is one of the key requirements of a website of the nature being discussed. So if a customer tries to do a posting at a time when the 'entrepreneur' is not available via email, he would not get a response for his action at that point. Wouldn't that turn off the potential 'seller' from using the website in the future ? I would think that when I use a website of that nature I would want immediate response (especially if we are considering the example of a website like craigslist).
    I would think that an actively interacting interface would attract people more. For example, sellers would love to know how many buyers were interested in the product he had posted or did the product he post actually have any traction amongst potential buyers, etc.

  6. Srivathsan - I completely agree with your comment. But what you are suggesting is phase-2 of the product. Remember we are talking about the minimum viable product here - the least we can do to validate our customers, the better.