Thread by @joshuamschultz

Data can be thought of as water.

It is flowing through your company.

Many companies can't harness it and use it because of their culture, not their tech literacy.

Some things to consider...

Data is flowing and moving.

The more you capture, the more energy you have to use.

However, this is only potential energy.

Keeping all of it creates a lake in your organization.

This ends up just being a stationary body of potential.

You don't need a lake to use water to power things, you just need a constant and focused stream.

The key is to work backward.

Here are some basic questions to walk you through using data better in your organization.

1. What problems do you feel data can improve or solve?

Instead of looking at the data, excel files, and apis available, assume its all worthless and a distraction.

Instead of focus on what needs help or solving in the business.

2. What data points do you need?

Once you know the problem, what are the numbers, pieces of information, or knowledge that would solve it?

Make a list of these points.

3. How can you capture them through activity instead of entry?

Think through your processes and workflows at the business. How can you grab this information from people doing their normal jobs? No extra entry / forms. Is this generated from current activity that you can capture?

4. Which ones aren't capturable yet?

Once you have grabbed the ones you can through activity, what is left?

Where is this created?

Can you build a useful tool that will create it, and keep workers from having to enter it?

5. Which aren't capturable but exist elsewhere?

Finally, you'll arrive at a few that you just need to record at various intervals.

By pairing these with nearby, data-entry heavy processes, the impact will be minimal.

Ex Entering invoices, orders, and emails are all entry heavy

By pairing, you are lowering the "task switching" occurring, and adding through a marginal adjustment (rather than through a new process)

An extra field here, a scan there - you can create the data with minimal work intrusion if carefully planned.

6. How will you save these data points?

Will you keep these in various databases and pull when you need?

Will you create a data lake and store there, and then automate the compiling of this data?

What methods and calculations will do on them?
Having the raw data is one thing, but you'll always need to perform "transforms"

Days between, GM as percent of Rev, number of shipping days.

You can use excel macros, python scripts, or other data pipeline tools.

In what way will you make available or distribute the information?

You can export to excel and email.
You can create a weekly newsletter that auto emails metrics and key charts.
You can create an export to Power BI or other dashboard tool.

What is the feedback loop and monitoring method to make sure the data is helping or solving?

Too much data and it won't get used. Not in the right place and it won't get looked at. Not put together right and it won't be helpful

Meet with your team on a regular basis to update and iterate your data policies.

What is too intrusive?
What is too error-filled?
What is not helpful?
What else can we add?