Five Vital Questions Every Data Leader Should Ask Before Every Project
Beginnings are always difficult, and Data Science projects are no exception to that rule.
There is nothing more frustrating for a leader or for a team, than a project that doesn’t see the light of day. The project starts strong. The team delivers a solution. But it ends up like an academic exercise, not impacting the business in any way.
Lack of commitment from relevant stakeholders, design mistakes, faulty execution. The reasons are many.
The truth is, many projects are doomed from the start.
Here are five questions to ask at the beginning of a project to insure it doesn’t lose steam and end up in a drawer.
Why is the client asking for this?
The first step.
A project should always arise from a business need. That much is widely accepted. The next part is not.
Your client doesn’t always know what he needs. Ask questions, be curious. Never follow blindly what a client asks for.
This way if you start working on something, at least you’ll know you’re working on the right thing.
What does success look like?
Images are powerful.
You have now clarity over what you are going to be working on. It’s the time to think of how you can make an impact.
The best way is to decide on a KPI. It can be anything from revenue to churn rate. The narrower the better, because it will be easier to attribute the changes to your work (e.g. there are a lot of things that impact revenue, so it will be difficult to argue that you really moved the needle)
If you have a KPI, you should talk to your customer to understand what sort of change they would like to see. Set the expectation from the very beginning. You don’t want to have a very successful project and a client that tells you they are unimpressed and expected something different.
Do I have the buy in from the relevant stakeholders?
Identifying the key players is everything.
You might get a well thought out request from a client, set expectations about KPIs and get to work, only to find later on that you needed some other player to do something for your project to have impact.
Think from the start about what areas inside a company are needed for your solution to be successful and ensure all the relevant parties are engaged with the project from the start. Everything becomes ten times more difficult afterwards, and you’ll be in a worse negotiating position.
Involve everyone from day one.
How can we make it smaller?
Machiavelli famously said:
Make mistakes of ambition and not mistakes of sloth.
Well, Machiavelli was not talking about Data Science Projects. Granted, sloth is not a quality you want in Data Scientists. But you want to temper your ambition as well.
Think of the minimum viable exercise you can do to show value to the business. Work on that and if it’s successful build on that towards something bigger.
How is the the solution going to integrate with the customer?
Data Leaders tend to think in a void.
Well, no solution is an island. Solutions should be integrated with the day to day operations of a business. It is of no use to have a wonderfully predictive model that can’t be put into production because of technological constraints.
Think about how your solution will integrate with the business before starting out.
No matter how big the project, it’s always the little things that make the difference.
Take your time at the beginning to ask these questions and avoid larger problems down the line.
