Make it Real: How to Make Your Application for Mozilla’s MVP Lab Stronger
(Shout out to Patrick Lu for putting this post together!)
After helping many people in office hours and sifting through some of the ideas on Slack, we’re seeing a common pattern here at the Mozilla MVP Lab:
Many ideas are too vague and aren’t yet a product idea. The first question we look on our program application is “One line summary of the idea.”, followed by “Description of product idea: What will the Minimum Viable Product be in 8 weeks?”
This isn’t necessarily the most important question for your application, but it is one that gives us context for the rest of the application. The successful applications will have something cohesive, to the point, and well thought through.
A VAGUE IDEA
An example of an idea that isn’t quite a product: “We’re going to build a product to combat misinformation in the news. We’ll use deep learning to build a model to classify that news and give it a score”.
FROM VAGUE TO SPECIFIC
Turning it around, here is something that would be a product: “We’re going to build a news aggregation app that ranks each news article from major publishers (NYT, WSJ, …) by how accurate the information is… The main user interface will be a feed of news articles which the user can sort by accuracy score and publish date. We’ll pull in all the news articles either through the website’s API (https://developer.nytimes.com/apis), or through the google news api (https://newsapi.org/s/google-news-api). To score each article, we’ll allow users to verify the accuracy with a thumbs up or thumbs down and combine that with Snopes’ api (https://www.snopes.com/faq/do-you-have-an-api/) which we’ve already contacted them about access. Overtime, we can figure out how to build in a machine learning portion as the users are classifying the accuracy for us, and hopefully we can have a deep learning solution.”
From the first example, we don’t truly have a good understanding of what you will be building, and from that we have to assume that you don’t either. In the second example, it’s clear that there was thought and action put into the product. The second example knew about The NY Times API, the google news api, understood how the user interface would function and how the users would interact to score the news.
STEPS TO SHARPEN YOUR IDEA
A small guide on how to get from the first example to the second:
1. Get on a Zoom call with your teammates.
2. Brainstorm the idea & go through a user walkthrough — what are the first few user screens going to look like? What will a user do on our app, and how will that help them?
3. Make it real…Outline a real “data” scenario — For example, outline the actions they will take or write out the information they will input into the system and how the system or community users will respond. In our experience, builders often like to experiment with their systems using “hello world” or whimpy data to test functionality. This rarely gives you real insight into the USER VALUE that your system will provide. So make it real with real scenarios!
3. Draw out some screens on a napkin or use some service like https://balsamiq.com/ or https://whimsical.com
4. Look deeper into some assumptions you have — can a deep learning model really solve a problem? Where will you get the data?
5. Talk to some potential users! Who would use your app / website? You probably have friends or family who would. Walk them through it, get feedback, and iterate!
If you have already submitted an application, feel free to go back and update anything given this advice. If you haven’t already, dig deep to create your best application.
