loading...
Cover image for What is YOUR MVP

What is YOUR MVP

waylonwalker profile image Waylon Walker ใƒป1 min read

Minimum Viable Product

What does an mvp look like for you.

Is it a csv file?

Does it need a Model?

Does it have a dashboard?

Some screenshots thrown into powerpoint?

Is it a full blown application?

Discussion

markdown guide
 

If I'm asked for data I generally ask:

What is the problem that has led to this request?
Having some kind of background is much more helpful than just 'give me a dashboard that does xx'. Taking the business problem, turning it into a data problem, coming up with a data solution then presenting back a business solution is an art form in itself.

What will solve the problem? one grid? five pages of grids? a chart? A number?
No point making a beautifully responsive, interactive dashboard if they are just going to download the raw data into a spreadsheet or pick a number out and put it in a slide deck.

What external data do you have that we need to align to?
Nothing worse than finding out there is logic you should have used or another dataset they are checking yours against. The words no one wants to hear ... 'er ... your numbers look a bit wrong'

 

โœจ Fabulous as always @helenanders26 ! All three are stunning responses with so much wisdom behind them.

I have slowly pulled back to start with some sort of compiled dataset for MANY requests for reasons highlighted in #2. Most folks seem to enjoy being able to touch the data a little bit as long as they don't have to spend days recompiling reports. There is always room to riff off of their ideas when jumping to the next level with responsive interactive graphs.

 

You're absolutely right, folks like being able to tinker around and 'drill down' at certain points. It's important to take an iterative approach so you don't aggregate away everything only to find they need more detail than you thought.

 

I'll start it off with a big IT DEPENDS. In some cases I have an existing application that it makes sense to easily slap in another route. Sometimes I have a technical audience, but the data is so big that it would not be feasible for them to sift through the data. Here we might start with a csv that they can start running some hypothesis testing in their own tools.

My biggest thing is that it's got to be repeatable. Anything I have ever made difficult to repeat has bitten me. It can be quickly scaffolded from a template, lean on existing applications, but must be able to be reproduced.

 

When setting up the requirements of a product I use the MoSCoW (Must, Should, Could & Won't haves) method. The MVP includes all must haves.

 

This sounds really interesting. I probably don't list out the won't haves as well as I could rather than just trimming off must-haves.