This is a guest post from Joca Torres. Joca is the director of product development and product management at Locaweb, Brazil’s leader in web hosting, cloud servers and SaaS applications like email marketing and online stores, serving more than 250,000 customers.
I met Joca at Business of Software 2011 where we had a great discussion on ways to monetize consumer SaaS apps, and in particular, ContaCal, the subject of this guest post. Joca recently wrote a book in Portuguese called “The Startup Guide: how startups and established companies can create and manage profitable web products”, and he was kind enough to translate a few sections of it for us. This is Joca’s second guest post in this series. If you haven’t read it yet, his first guest post was Why the Hurry to Launch an MVP?
In my previous post I explained the three reasons why we need to be fast in launching an MVP. Now I want to explain how I did that with my own MVP experiment.
Back in mid-2011 I had 3 product ideas that I tested market acceptance to define which one of them I would build an MVP. I’ll explain in a future post how I run this market acceptance test. The idea that presented the best market acceptance was a calorie counter system, which would not only allow the user to count the amount of calories but also the quality of those calories. It would do so using colors. Red calories should be limited to a maximum of 10%. Yellow calories up to 35%. Green calories as you wish.
Now I had to build this app, but there were two issues in the way:
- The last system I coded and put in production was in late 1990s, when ASP was just launched by Microsoft and even though PHP already existed, no one had ever heard of it. So my hands-on programming experience was very out-dated.
- My day-to-day job coordinating product development and product management at Locaweb already took almost all of my time. I only had 1 or 2 hours a day to invest in this new product and with my outdated programming skills that would take a long time.
For this reason I decided to outsource almost everything required to build my MVP except the two main aspects that cannot be outsourced:
- product management, the ability to look outside and understand people, their problems and figuring out a way to solve those problems.
- project management, the ability to look inside, coordinate all efforts required to do something and removing the obstacles.
Preparing to build the MVP
The first thing required to build an MVP is deciding what to build. So I put together a wireframe, using PowerPoint.
Sorry it is in Portuguese but here’s a quick explanation about the slides:
- slide 1: plain simple login screen.
- slide 2: also very simple sign up screen. I end up making it even more simple asking only for name, email and password, to minimize sign up attrition.
- slide 3: this is the main screen of the application. Here the user search for foods, inform amount eaten, and receive information on number and quality of calories.
- slide 4: a simple report on the evolution of calories intake.
- slide 5: another report on the weight evolution.
- slide 6: some additional features.
With the wireframe in hands I needed some visual design stuff. I needed a logo and a look and feel to be applied to the website and the product. For the logo I used a Brazilian design crowd sourcing web site called We do Logos which is quite similar to 99designs. I just provided a brief explanation of my product idea and received more than 50 logos from many different designers. I selected one and paid US$310.
For the look and feel of the website I used themeforest with tons of cool templates for websites. I picked one for WordPress and paid US$35.
Actually building the MVP
As I mentioned earlier, my hands-on programming experience is quite out-dated so I decided to hire someone to do the coding job. I found 3 guys in Rio de Janeiro – I live in São Paulo – that had an interesting development offer, very well suited for startups. They call it StartupDev. I presented them my wireframe, the logo and the WordPress template. They analyzed the material and we discussed to refine the feature specs. With that in hands they gave me an estimate of how much time it would take to do it but – here’s where they innovated – I could only have 2 days of their time. Their service is restricted to 2 days by design. Not because they had other things to do, but because they build MVPs in 2 days. This is their value proposition. For this reason, I had to get down to a minimum feature set that could be developed in 2 days in order to put my product in the market. At the time, I was one of their first customers of this new web app development model, so I paid US$3,000. I believe they are a bit more expensive now.
I’m also not a good designer so I needed some help to put the logo and the template in my WordPress installation. I hired a designer who did the job. Since it was just minor adjustments, it cost only US$80.
Last but not least I needed someone with nutrition knowledge to build the list of food with calories and to classify those calories in green, yellow and red. I hired a nutritionist who charged me US$350 for this work.
Here’s the result:
ContaCal website first version
ContaCal signup page first version
ContaCal application first version
Summing up the investment in building the MVP I spent a total of US$3,775
- logo: US$310
- WordPress template: US$35
- software development: US$3,000
- visual design adjustments: US$80
- nutritionist: US$350
If you are a software developer or a visual designer you certainly can cut some of these costs. However, cutting these costs may harm your MVP because you may be tempted to do more than what’s needed just because you can. Outsourcing can be a good way to put a constrain in your MVP development.
The 10 days
Below is a chart showing number of signups since the beginning of the software development:
- 23/08/2011: Software development started with StartupDev guys.
- 25/08/2011: Software done and deployed and beginning of logo and template adjustments.
- 01/09/2011: WordPress based site ready.
- 02/09/2011: AdWords campaign turned on. Until this day I had only test signups. From this day onwards I got real users!
- 04/09/2011: Email sent to people who previously showed interest in the pre-launch website.
- 07/09/2011: Facebook campaign created.
So as you can see, from the beginning of the software development until turning on the AdWords campaign it took 10 days! 🙂
In my next post I’ll explain the theory behind the fast software development with quality. What do we need in order to produce high quality software in a hurry? Good professionals? Good methodologies?