Unintentionally we did this part very well at BackupAgent. Our MVP landed some initial customers and a partnership with a large service provider. We learned the potential of the market and channel and for that reason it was easy to convince investors too. A ‘leap-of-faith’ event was in 2006, when a Swiss customers faxed us a 30k order without any form of sales. This happened one day before we pitched to our current investor.
I don't consider a landing page an MVP. This is an experiment to find a channel to customers. A landing page could be part of an MVP.
Dropbox released this video way before its product was ready to launch. The video went viral on Hacker News (blog of Y-Combinator) and the result was a list of 75k beta registrations. The video should capture as much value as possible and should cover both product as well as solution. My recommendation is not to confuse the MVP with the Minimum Viable Video, please do invest in good voice-overs and graphics. You need to do this to convince the customer you are a professional company, which is dead serious about delivering this.
The best example of a concierge MVP I know is Manuel Rosso’n Food-on-the-Table. Manuel initially delivered the service to a single customer using an excel sheet. It was an MVP combined with a very effective way of customer development. I recommend you watch his video presentation on last year’s Startup Lessons Learned conference.
Mockups are especially great in a business-to-business startup, because early adopters are more likely to pay for a product that is delivered to them in the near future. We use mock-ups in BackupAgent to convince our customers to try our next version of BackupAgent and gather feedback.
This MVP would be a perfect example of the definition I provided in the first slide. Essentially you strip down the service to the bare minimum to sell it to customers. An example is a Dutch startup, Wercker. They’re working on a cloud deployment platform for developers. The initially limited it to one programming language: node.js. Also, they sold the product to a single customer on the promise of adding the language this customer is using to the platform.
There’s a couple of things to take into account when you’re building an MVP:1: You’ll probably build more then one, because this is an iterative process. Don’t expect to get it either right or wrong after one iteration.2: You need to embed means to capture data at each stage – for internet consumer services there’s Dave McClure’s AARRR model (Acquisition, Activation, Retention, Referral, Revenue)3: You need some administration for learning. The best known practise is to use the Lean Stack by Ash MauryaThere’s a nice video on this by David Binetti from Votizen.Beware of calling evry little change you make a pivot. You only pivot when you change fundamental elements in your business model.
Part of the challenge which an MVP can overcome is to work around missing competences (no designer, no developer, etc). Example: Twitter Bootstrap'minimum' means that you start to produce value for customers faster and with less effort. Example: Aardvark.What if your users do not pay for the service? They're not your customer, so in that case your business model will depend on customers who would pay for the value those users represent. An MVP would simply validate this produced value. This is a leap-of-faith assumption.
Too minimum: can lead to false conclusions. This is because the product is not viable and promises or produces no value. Testing the wrong part of the business model: Entrepreneurs sometimes falsely assess their 'riskiest assumption' Procrastination: most common pitfall - spending months to produce this 'MVP' and not being able to deal with disappointing results which leads to more effort into the wrong direction because of 'loss aversion'
Building a Minimum Viable Product
Robbert van Geldrop| Founder, BackupAgent
Building a Minimum Viable Product
Part 1: Defining the MVP
Part 2: Tools, practices and pitfalls
Part 3: Forms and examples of MVPs
After presentation: Your MVP
Defining the MVP
• the least amount of effort required to offer some value to customers
• a full product or service which completes the value consumption or
convincingly demonstrates its availability
• a way to collect the maximum amount of validated learning
• a product is also a transaction, so a customers buys something or vows a
strong promise to buy something after which you already deliver value
• Quote from Otto Hilska, FlowDock:
‘Asking for money was one of the best decisions. Customers took us more
seriously, and we started getting better feedback’
Minimum Marketable Feature
MMF + MMF + MMF
Set of experiments that focus on a part
of the business model
Read all about Ash Maurya here
Defining the MVP
An MVP naturally comes in later stages of Customer Discovery:
• It’s a rudimentary solution to a problem worth solving
• An MVP must still trump any alternative solution which your
customers have used or considered
• Basically, it’s the next step after building some landing page and
running an AdWords campaign to validate some demand
MVPs will attract early adopters:
• Customers who can live with its limitations
• People who buy into the vision and the ‘why’
• Customers who buy NOW and are relevant for Validated Learning
Remember! Build, measure, learn
Not every change is a pivot
Practices – other considerations
‘Reduce waste’ = work around your missing competences
‘Minimum’ = deliver value to customers faster and with less effort
If you users do not pay for the service, it can still be an MVP only
if those users are part of your ‘leap of faith’ assumption
• Can lead to false conclusions.
• This is because the product is not viable and promises or produces
Testing the wrong part of the business model:
• Entrepreneurs sometimes falsely assess their 'riskiest assumption'
• Spending months to produce this 'MVP' and not being able to deal
with disappointing results
• Leads to more effort into the wrong direction because of 'loss
Twitter Bootstrap: UX and design covered
Windows Azure: integrates with Visual Studio 2010
Amazon MTurk: outsource questions, replaces algorithm
99Designs: cost-effective logo design
KISSmetrics: captures usage
Paypal API: easy payment, voluntarily after receiving
We worked with a team of two
Total effort was approximately 2 man weeks
The service was completely functional
Quer.io – tools used
+300 visitors in a week (promoted via personal twitter accounts)
+60 people used the service
Referrals were limited
We had to do small iterations to deliver real value
We got some press coverage by Sprout
Key learning: journalists were very enthusiastic and used it as a
tool to outsource and speed up desk research
Tel: +31 88 700 8000
Thank you for listening
Robbert van Geldrop